Origins of the Apple Human Interface

苹果人机交互界面的起源 拉里·泰斯勒（Larry Tesler）与克里斯·埃斯皮诺萨（Chris Espinosa）借助档案文件，详细解析了苹果公司在1979年至1980年代中期如何确立其基础性的图形用户界面标准。泰斯勒详述了Lisa项目背后激烈的技术争论与快速可用性测试，正是这些测试催生了单键鼠标、屏幕顶部菜单栏、标准化键盘快捷键以及桌面图标隐喻。埃斯皮诺萨则追溯了Apple II团队并行的可用性研究工作，展示了零售教程与文档交付期限如何促成了后来被Macintosh所继承的规范化开发者指南。两人共同打破了苹果仅是抄袭施乐帕拓高（Xerox PARC）的传言，强调了针对完全初学者开展的严格实证测试如何从根本上重塑了个人计算。 拉里·泰斯勒与克里斯·埃斯皮诺萨回顾了苹果人机界面的诞生历程，依托1979至1980年的内部规范文件，证明了苹果是通过实证可用性测试而非直接模仿，从而在很大程度上背离了施乐帕拓高的模式。 (6:34) 苹果引入了针对毫无经验的受试者（包括苹果入职第一天的员工及外部非电脑用户）的快速用户测试，以解决围绕单键鼠标、屏幕顶部菜单栏以及剪切和粘贴范式展开的设计争论。 (35:19) Macintosh的用户界面规范融合了Lisa的图形鼠标机制与Apple II部门由文档驱动的一致性标准，克服了第三方开发者对基于ROM的统一界面的激烈抵制。 (87:08) 早期的办公套件策略假定用户只需要一个捆绑的核心套件，这一前提由于高昂的成本和封闭的软件生态导致Lisa在商业上受挫，直到数十年后才被微软Office证实其合理性。 (108:41)

[102:51] Chris Espinosa: The constant tension is that the people who are designing are almost always by definition expert users and will often overlook that what they think is obvious is not at all obvious to a novice user. [78:14] Bruce Tognazzini: When a user hits a snag, he will assume it is his fault; he will not report it, he will hide it. [102:37] Chris Espinosa: Having to test new designs and test your assumptions of what your audience knows is pretty much eternal.

[0:00] [Music] [1:01] eventually opening a public museum [1:03] focused on the history of computers [1:06] hardware software documentation and so [1:08] forth designed for adults not a kid's [1:10] hands-on museum but it's an adults [1:12] oriented museum it's a long-term project [1:14] we're probably going to take four or [1:15] five years to get there to build a [1:16] building to find a location in sight and [1:18] so forth the meantime we're focusing on [1:21] continuing to build a collection which [1:22] is the finest in the world it's got [1:24] thousands of artifacts and thousands of [1:25] photographs and documents and books and [1:27] so forth we're inviting people to help [1:30] us with this project by doing lots of [1:32] things we have weekend volunteer days [1:35] when people come and help us work on the [1:36] collection we'd like as many people as [1:39] can to become members of the museum and [1:42] you can do that in several ways one [1:44] program if we just started is that you [1:47] can become a member for a hundred [1:49] dollars and get a card that gives you [1:50] access to the History Center and also [1:53] gets you access to Moffett Field so on [1:56] Wednesdays and Saturdays you can without [1:58] going through the process you went [1:59] through today get on to the field and to [2:01] have access to our visible search [2:03] location in addition for the next I [2:08] guess about six months you have an [2:09] opportunity one time only to become a [2:11] founding member of the Computer History [2:12] Center by donating a thousand dollars or [2:14] more and be permanently enshrined as one [2:16] of the founders of the Computer History [2:18] Center it's the same kind of program [2:19] that was done by the computer museum ten [2:22] or twelve years ago when it was in its [2:23] founding stages what else are we doing [2:28] we're we're having a we are mounting a [2:31] bunch of expeditions in various places [2:34] we've done several at trade shows and [2:36] conferences we've got one that's going [2:38] to be opening next week at the new gates [2:39] computer science building it's Stanford [2:41] where we're taking part of our [2:42] collection and and showing it off and [2:45] until we have our own permanent space [2:47] that's a good way to show off some of [2:49] our wares we've got more stuff coming as [2:52] we speak there are four tractors loaded [2:54] being loaded in Boston and heading [2:56] across the country with the next [2:58] installment of of the machines and and [3:02] the part of the artifacts from Boston [3:04] and [3:04] rooted in there is what the johniac and [3:08] cm5 thinking machine that we just got [3:11] from MIT and a whole bunch of other very [3:13] valuable artifacts if you want to get [3:16] more information about the Computer [3:20] History Center talk to me or Gwynn bell [3:21] or Zoe or dag or any of the other people [3:23] who are on staff we'd like to thank the [3:27] people at NASA Ames here for their [3:29] hospitality both in the use of this [3:31] conference center but also in the use of [3:33] the warehouse space for which we're very [3:35] much in debt to them there are some [3:38] flyers in the back of that you can pick [3:39] up that describe what's going on here of [3:42] both NASA and generally in their center [3:44] for excellence and information [3:45] technology this is one of a series of [3:48] talks that we've been giving in [3:50] cooperation with a Bay Area Computer [3:52] History perspectives P Turner Spee runs [3:54] that and he will introduce tonight's [3:56] speakers Thank You Lynn yeah I don't [4:00] really run Bay Area Computer History [4:02] perspectives it's more a joint venture [4:04] between myself I work at Sun [4:06] Microsystems and Jeannie trico also [4:08] works at Sun Microsystems and Jeannie [4:11] was at Xerox PARC [4:12] many years ago I was part of history [4:15] herself [4:17] we started this computer talk series [4:19] over four years ago like 1973 and two of [4:24] our speakers at the first program are [4:26] here in the front row [4:27] no Hardy and George Michael from [4:29] Livermore and so a lot of people have [4:31] been coming back over the years from one [4:33] program to another and Larry Tesler [4:36] who's one of our two speakers tonight [4:38] himself did a program along with bob [4:41] Harriet Larry breed Don wise over three [4:44] years ago on the Stanford card stunts [4:48] which I wrote up in the description was [4:51] the largest bitmap display ever you know [4:53] it was like several thousand students [4:56] holding up colored cards and sequencing [4:58] that was a feat itself to do the [5:00] programming of the instructions back [5:02] then at that time and Larry has also [5:06] been coming to many of the programs [5:08] before and since then so it's like this [5:11] history talk series has developed the [5:13] history of its own so tonight we have [5:17] and Chris Espinosa from Apple talking [5:21] about the early Apple user interfaces [5:24] and I think as you may have heard Larry [5:28] came to Apple from Xerox PARC back 1980 [5:32] and he so figured he was bringing a lot [5:34] of user interface knowledge with him [5:36] only he's found out recently that user [5:39] interface testing and development [5:41] existed at Apple before 1980 and Chris [5:44] had a part in that Chris I think that's [5:47] correct was probably the youngest Apple [5:49] employee ever did demos at Apple while a [5:53] high school student [5:54] so Larry yeah pass the mic where's Barry [5:59] there you are okay yeah I already have [6:04] one [6:10] it's your tape in there Chris okay fine [6:16] good in here [6:34] okay well glad to be here this is a talk [6:38] actually that's based on one that [6:40] Kristen I gave it Apple back in July and [6:42] we got clearance thankfully from the [6:45] Apple lawyers which came about two or [6:46] three weeks ago so we could give it here [6:49] just in time to announce it and we're [6:51] grateful to Apple to release this for [6:55] public disclosure because we think it's [6:57] a general interest asking about what was [7:02] early user interface design at Apple is [7:04] kind of like asking about early engine [7:05] design at mercedes-benz or something [7:08] everybody was doing it so paying on who [7:11] you you could have gotten a lot of [7:12] people up here to give a talk and you [7:14] get different perspectives on the same [7:16] things that happened so today you're [7:18] going to get the perspective of a couple [7:19] of us to get it to be a little more [7:21] grounded in something more closer to [7:23] fact than the opinion or perspective [7:26] we're trying to use as much as possible [7:28] actual original documents that that were [7:31] created at the time so you see actual [7:33] contemporaneous thinking as it was going [7:37] on instead of just our memories which [7:39] fade over time and and tend to [7:42] aggrandize over time and so on so here's [7:46] something which has which is not from [7:48] Apple this is the cover of bite from was [7:52] it August 81 that was the small talk [7:57] issue colorful this one doesn't have [7:59] color of us very colorful balloon and [8:02] the reason they show it is that there [8:05] was a lot of influence from Xerox [8:06] obviously on the user interfaces that [8:10] were done on the Lisa and the Macintosh [8:11] but maybe not as much as some people [8:14] think media tends to just say oh they [8:16] just copied it so I'm gonna show you [8:20] what it looked like to be working on [8:23] from that article which I wrote to be [8:27] working on a small talk at that time [8:30] 1981 and you see what the windows look [8:33] like kind of a little rectangular title [8:36] scrollbar without any arrows and and [8:40] it'll pop up menu there are two pop-up [8:42] menus that you could use for each window [8:45] so it was a pretty different looking in [8:47] her face and one of the principles [8:56] though that was in the small talk [8:57] environment was to have no modes and [9:00] there was actually yeah a photograph of [9:03] a t-shirt that was made for me by a [9:05] friend that said don't load me in that [9:09] represented service this mission I was [9:11] on to try to eliminate all modes from [9:13] user interfaces which is necessarily a [9:15] good idea but it was definitely the [9:17] mission I was on and at the end of that [9:22] article had mentioned that Apple [9:23] Computer had gotten a license to small [9:25] talk and was maybe going to do something [9:27] with it but maybe not gonna do anything [9:29] with it and in fact Apple and Europe [9:33] Eckerd Digital Equipment & Tektronix all [9:36] did do various amounts of small talk [9:38] development after that but it was really [9:40] independent companies that that managed [9:43] to make small talk somewhat of a success [9:46] I'm jumping to the end here before I go [9:48] to the beginning just to let people know [9:50] what the Lisa was the idea of the Lisa [9:53] was to have something that was very very [9:55] easy to use people working people loved [9:57] the Apple 2 which was what we had at the [9:58] time at Apple but it took a long time to [10:00] learn how to use it and what we wanted [10:03] to end up with was something that was [10:04] easy to learn and just to show you that [10:06] we did do that this is a study done at [10:08] the end where there were 18 subjects 6 [10:10] using Lisa graph 6 using Lisa right Lisa [10:12] list [10:13] Lisa graph was a business graphics [10:15] program at least list was a list manager [10:16] kind of flat file database [10:18] Lisa Wright was a word processor and as [10:20] you see for all but one user using Lisa [10:23] list it was under 20 minutes to get [10:26] through the core competency to use the [10:29] application and if you ever had a chance [10:31] to use a Lisa and I hope hope to be [10:33] one-upper reading here some day then [10:35] you'll see that it was extremely easy in [10:38] many ways more bulletproof I'd say than [10:41] the Mac that's simpler than the Mac [10:43] actually but but we loaded a lot of [10:46] things into it to try to make it so [10:48] people couldn't get themselves in [10:50] trouble ok so let's start at the [10:52] beginning I started at Apple on July 17 [10:55] 1980 and I just left there a few months [10:58] I was there for 17 years and this memo [11:01] was written on July 18 1980 so one thing [11:03] I discovered when I left big Xerox [11:05] Corporation and went to this little [11:07] Apple start-up was that you just really [11:09] get engaged immediately and they put you [11:11] right to work and one thing I was told [11:14] the day I got there was that the user [11:17] interface design was basically done and [11:20] we were they were just going to you know [11:23] write it up and finalize it and [11:25] implement it and that was the end of it [11:26] and it was too bad I took so long to [11:29] show up at the office because they were [11:31] hoping I could have contributed to it [11:32] but now there was no time left to do [11:33] that and this was the 18th and they told [11:37] me that all the decisions had to be made [11:38] by the 23rd and by one week later the [11:41] external research external reference [11:42] specification had to be done which was [11:44] the complete spec but we couldn't [11:46] compromise quality at all and and so [11:49] what could I do between the 18th and the [11:51] 23rd to make sure we made all the right [11:52] decisions so I started explaining how it [11:55] was really necessary to do months of [11:56] user testing and careful iteration and I [12:00] gave my whole spiel about the right way [12:01] to design software and they said well [12:04] sorry we only have 5 days into the [12:06] weekend in there so we got together and [12:11] we can't with a compromise which was to [12:13] lay out a schedule that was very [12:14] aggressive and people promised that they [12:15] would meet their part so let's go [12:17] through the schedule this was the 18th [12:24] turn this a little easier to get out [12:35] this No [12:37] it is pull oh that kind of pointed [12:44] you're the one with the laser oh okay [12:51] so here with the schedule Monday we're [12:56] going to run user tests so we're gonna [13:00] well we're all gonna try it ourselves [13:02] everybody's going to try bill atkinson's [13:03] prototype Bill Atkinson was the person [13:05] building the prototypes at that time and [13:06] thinking up a lot of the ideas and we're [13:09] going to evaluate several issues that [13:11] were going on then we're going to bring [13:13] in after two hours of that we're gonna [13:15] take a break for lunch then at 1 o'clock [13:17] we were going to bring in sue Espinosa [13:19] who was in charge of training and also [13:20] coincidentally is Chris husband knows [13:22] his mother and then she was going to [13:25] give her opinions about that and maybe [13:28] help us decide on what would be easiest [13:31] to train in her opinion and then at 2 [13:33] o'clock we were gonna discuss for an [13:35] hour and all of more issues and now we [13:39] would be done we will have discussed all [13:41] the issues and then between 3 and 5 [13:43] o'clock we would present the solvus [13:44] Steve Jobs and Steve would make the [13:46] decision which which we knew would be [13:48] definitely forthcoming and then on [13:53] Tuesday we were going to take the things [13:57] that Steve liked and we were going to [13:58] test them on subjects from 9 to 10 two [14:02] subjects two more subjects from 10 to 12 [14:05] we must've figured we'd be slowing down [14:06] by then 1 2 3 we evaluate the test [14:09] results [14:12] so that was it okay so remember that was [14:15] supposed to be July 22nd okay so now the [14:19] rest of the for the next 15 minutes or [14:22] so keep in mind that that was the [14:23] scheduled for July 21st and 22nd okay [14:26] well we actually did that we did all [14:29] those tests we did talk to Steve Jobs he [14:32] gave his opinions we did some more [14:33] testing and then bill went to write up a [14:37] user interface specification external [14:39] reference specification I was a little [14:42] worried because we hadn't really done [14:44] extensive testing but we tested a couple [14:46] things so this is data August 6 which [14:49] was as you know as you may notice a few [14:52] days later than the deadline date that [14:53] we had provided but it took a while and [14:55] it said that you had one week to review [14:59] and return your comments so we were [15:01] already starting to slow down a little [15:02] bit here and let's look at some of the [15:07] names of the people who got this memo [15:08] because it'll turn up to be important [15:11] later in the story or just interesting [15:12] right now Ken Victor who is gonna [15:15] respond to it later and jef raskin also [15:18] barry smith in addition to that notice [15:21] we sent it to Mike Markkula who was one [15:23] of the founders of the company and Steve [15:24] Jobs trip Hawkins who was the head of [15:26] product marketing we sent it to [15:28] everybody in Lisa software group we only [15:30] send one copy to Lisa Hardware they get [15:32] to share we send it to every vice [15:36] president we could think of and to our [15:39] two people who became Apple fellow [15:40] somewhat later rod Holt and Steve [15:42] Wozniak who's here in the back today I [15:45] don't know if he remembers receiving [15:47] this memo [15:49] but he probably does it was pretty [15:51] dramatic okay so we did some I think we [15:55] did some things right they´ll started [15:56] out by talking about who is the customer [15:59] so it wasn't just some easy-to-use [16:01] computer it was an easy to use computer [16:02] for specific people so what was it first [16:05] of all it's a single user workstation [16:07] that was an important point in those [16:08] days was still an important point if you [16:10] were its son or something [16:11] and designed to enrich the productivity [16:14] of the office worker so our market was [16:16] office workers and it went on and talked [16:18] about the hardware and so on it's more [16:21] than a piece of hardware though it [16:22] includes major software and we had [16:24] specific applications in mind word [16:26] processing graphics data management and [16:28] communications it wasn't intended to run [16:30] thousands of applications not the Lisa [16:32] the concept that trip Hawkins had was [16:35] they had done studies and they found [16:37] that most people on the Apple too ended [16:38] up using only two or three applications [16:40] therefore this was the great leap all we [16:44] need to do is provide two or three [16:44] applications that's all anybody will [16:46] ever want it was an interesting theory [16:48] and Apple funded a whole project based [16:52] on that assumption and then it said the [16:55] purse upwards integrated through a [16:57] powerful and simple user interface and [16:59] it doesn't say it here but John couch [17:00] who was in charge of the Lisa project [17:02] and did a sort of Kennedy we were gonna [17:05] send a man to the moon and bring him [17:06] back thing he said all I want is I want [17:09] to be able to go into a spreadsheet make [17:11] some put some numbers in go and graph [17:14] them take the graph paste it into my [17:16] word processor document and included in [17:18] a report you didn't say paste you know [17:20] put it in my word processor document he [17:22] said oh I can do all those things an [17:23] integrated way then will be way ahead of [17:25] anything available on the Apple to where [17:27] there are all these separate programs [17:28] that don't interact in our interact with [17:30] each other that's all he wanted whenever [17:32] we brought him any issue he said [17:33] remember what I want and he went through [17:34] the whole mantra and that was a very [17:36] healthy thing it kept us focused and [17:39] then we also knew when we were done [17:42] well who are the users non-technical non [17:45] analytical people none analytical it's [17:47] very important point first time users [17:51] well 1980 there were a lot of first-time [17:54] users almost everybody and here are [17:57] examples administrative assistants and [17:59] secretaries also secondarily managers [18:02] accountants and executives wasn't for [18:04] programmers wasn't for kids it wasn't [18:06] for people at home doing recipes this [18:08] was an office system and so when you [18:11] start looking at it and comparing it [18:13] maybe with the math you'll see some [18:14] differences based on that and you may [18:16] see some similarities to Windows design [18:20] philosophy so there was a whole section [18:22] on the design philosophy and what were [18:23] the most important features to put in [18:25] graphic images well I was pretty vague [18:28] menus a pointing device that you could [18:32] move around freely as opposed to the [18:35] cursor keys we were using up to then and [18:37] we were going to let you invoke commands [18:39] from the keyboard also because there [18:40] were times when you would do things over [18:42] and over and people felt that sometimes [18:44] the mouse slowed you down a bit so let's [18:47] look at this keyboard this was the [18:50] keyboard that existed on the hardware [18:52] when I arrived it had a code key in the [18:56] upper left corner and some of the keys [18:59] on their fronts had some characters in [19:02] other words in order to make the [19:04] keyboard the width they wanted they [19:05] couldn't fit all the characters on the [19:06] keys so they squeezed a few extra [19:09] characters on the Front's of the keys [19:10] and then have a special key to invoke [19:11] those and later we decided we could use [19:14] that key together with other keys to do [19:16] another function that you'll see later [19:18] now I didn't like the hardware designed [19:21] too much and I wanted to change it but [19:22] they said you can't change that you know [19:23] you came from a research place we just [19:25] have to ship the hardware that we [19:26] already built otherwise we won't be able [19:28] to ship this product when it's gonna go [19:29] to market I said well when is that and [19:30] they said six months that was a 1982 [19:35] jump ahead and ship till middle of 83 [19:37] probably my fault okay [19:40] so here was the pointing device kind of [19:43] a brick that's a mouse and you know I [19:47] the color is me I added color last night [19:49] just to highlight things you should look [19:50] at but it had two buttons at that time [19:52] and I hand wrote on here something about [19:54] the shape I wasn't too happy with the [19:55] shape and I wasn't sure what [19:57] distinguished these two keys and whether [19:59] there was a little space between them [20:00] where they were flush or what so you'll [20:03] see some of my handwritten scribbles on [20:04] these memos and here's how the mouse was [20:07] used well actually they didn't know a [20:08] lot of how the mouse was used yet they [20:09] mostly were focused on the window [20:11] manager how the user would manage the [20:13] windows on what we now call the desktop [20:15] and so at that time the idea was this [20:18] there would be a click and a drag and [20:22] those are two things you could do you [20:24] could click and let go or you could [20:25] click hold drag the mouse and let go [20:27] and sometimes we called it drag [20:30] sometimes we called it draw various [20:31] nicknames and then there was the buttons [20:33] got names the point button and the grow [20:36] button and depending on where you [20:37] pointed in whether you clicked or [20:38] dragged and which but you know which [20:40] button you used this page here shows [20:43] what happens if you went into the title [20:45] tab of a window that little thing that [20:46] had the name and it either would display [20:48] the window menu move the window open or [20:51] close the window or change the size of [20:53] the window all of it done by clicking in [20:55] the title so there was no grow box in [20:57] the lower right there were no menus that [20:59] were it was all done with just one of [21:01] two mouse buttons and a click or drag [21:03] operation now there was a menu at the [21:06] bottom of the window but had nothing to [21:07] do with window operations it had to do [21:08] with operating on the content inside the [21:10] window we'll get to that now I should [21:13] stop saying when I got there because by [21:15] then I'd already been here for two weeks [21:16] so I had some influence on this I can't [21:21] remember what so here's a later two Jeff [21:23] now Jeff isn't misspelled that's jef [21:25] raskin who would does spell his name [21:26] that way and he was for some reason [21:28] became the person who we were always [21:30] sending letters to on the lisa project [21:31] and as you see this scrollbar was on the [21:34] left side it had the horizontal [21:36] scrolling arrows at the top the vertical [21:38] scrolling hours at the bottom now those [21:41] of you who know the the next machine [21:42] know that that's the features that you [21:44] have the two arrows close to each other [21:46] so you don't have to move very far to go [21:48] between the two arrows that was actually [21:50] designed in 1980 by Bill and then at the [21:53] bottom in the Edit menu had [21:54] cut paste doop to copy things and undo [21:57] and then on the side there were the [22:00] equivalent of today's icons this was [22:03] taken from small talk where you took the [22:04] tab of a window threw away the window [22:06] and just kept the tab called it [22:08] collapsing the window and these were [22:09] windows that were available but closed [22:11] right now and you could open them up [22:18] okay then we had this idea of billion [22:21] menus there what just to jump ahead [22:23] they're what we call dialog boxes today [22:25] and there were two kinds there were [22:27] kinds where you somehow open a submenu [22:33] and you get to type into that submenu [22:35] don't ask me exactly how and the other [22:37] was where it actually opens up a big box [22:39] like we know today that was this thing [22:41] called forms and so it was a kind of a [22:44] vague notion at that time no one had [22:45] actually drawn any pictures even of what [22:47] those looked like I think and again we [22:49] said that menu commands can be a vote [22:50] from the keyboard and I put a little [22:51] note saying disagree somewhat and I wish [22:54] I could remember what I disagreed with I [22:57] think I was afraid they would do too [22:59] much keyboard stuff they were scroll [23:03] bars and you would hold a button down to [23:06] scroll continuously now there was an [23:08] interesting idea here that you could [23:09] double click to flip to the next page [23:10] you know anybody that knows about user [23:13] interface design with with single and [23:15] double clicks usually you make it so [23:17] that when you click once you can start [23:19] doing something and then when the second [23:21] click comes it doesn't sort of negate [23:23] what you already started to do this is [23:25] kind of interesting because you clicked [23:26] once it would sort of start to scroll a [23:27] little and when you click the second [23:28] time we would realize you wanted to jump [23:30] a lot so you'd get a little motion and [23:32] then a jump should I actually might have [23:34] been kind of cute but I don't think we [23:35] ever implemented it then there was [23:38] something in the called the elevator [23:40] shaft with the elevator in it and that [23:42] was what is still today that little box [23:45] on the scroll bar that you see there was [23:49] a abbreviations window that was that [23:50] code key we had to find a use for it so [23:52] you could type code and then hit a [23:54] letter on the keyboard it would look up [23:56] in a table and an expand your [23:57] abbreviation this was kind of taking the [23:58] kind of thing you'd find on a weighing [23:59] word processors so we put that in [24:01] because this was for secretaries [24:05] and there was something called the [24:07] system error window and that's what you [24:10] now know as alert boxes [24:11] in Mac and Windows and these were things [24:13] that would come up when there was a [24:14] problem and it would make some kind of [24:16] beef and it would tell you what the [24:18] problem was so I put a little question [24:20] mark here is the system frozen when that [24:22] comes up can you do other things and we [24:25] eventually decided it was modal but [24:27] actually that's a quite a while to [24:28] decide now we talked about testing we [24:33] said that the designs already based on [24:35] studies of user reactions to various [24:36] models as you saw we did three or four [24:38] hours of testing and we're gonna do more [24:42] testing and avoid any glaring design [24:45] flaws I was happy that that was put in [24:49] because that meant we could go and fix [24:50] some more problems than I was sure was [24:52] there because we hadn't done enough [24:53] testing yet so why did it come down to [24:55] this what it looked like the basic model [24:57] was overlapping windows which is right [24:59] from small talk [24:59] there was that box at the top just like [25:02] in Smalltalk that little title bar the [25:04] scroll bar was on the left just like in [25:05] small talk only instead of just that [25:07] little box they were the scroll the [25:09] scroll arrows there we thought it would [25:10] be nice to have something to point to [25:11] and small talk you push the button and [25:13] then the arrows appeared and we thought [25:15] that was a little weird and there was [25:17] this menu at the bottom and now if any [25:19] of you who's ever used a system called [25:20] UCSD Pascal if you remember they were [25:25] menus in UCSD Pascal and you would type [25:27] the first letter of the menu name and it [25:29] would bring up a submenu and you typed [25:31] the first letter of menu name and you go [25:32] through a hierarchy of menus now also [25:35] there was a machine Who am I gonna [25:36] forget the pig gonna forget the number [25:38] heel at Packard computer mm now it was a [25:44] later model in that HP something [25:45] thousand that had a rope the row of [25:48] function keys on the top of the keyboard [25:49] and at the bottom of the screen under [25:52] each key still have it they were in use [25:57] you hit that key that would activate [25:58] that menu item these were the two [26:00] products that that the authors of this [26:02] interface knew the best they used UCSD [26:04] Pascal dealer software development half [26:06] of them came for you love Packard and so [26:08] this was kind of a blend and so you [26:10] could they had function keys at the top [26:12] of the keyboard that you could hit that [26:13] would make invoke the corresponding [26:15] things and they also had these [26:17] hierarchal menus that were straight out [26:18] of UCSD Pascal not the little pop-ups [26:23] that were in small talk so that's where [26:25] it was in August cut and paste were two [26:29] of the commands in the menu that came [26:31] from small talk and as you see there was [26:33] a wastebasket that's what you called the [26:34] clipboard today we used to call the [26:36] wastebasket different from the trash can [26:38] which was a confusion we had later and [26:40] that's why we changed the name but when [26:42] you cut something and when the [26:43] wastebasket like in the office you cut [26:44] something and throw in the wastebasket [26:45] but then you could take it and paste it [26:47] somewhere else and it was serve a weak [26:49] metaphor but and so you'd select it with [26:51] the mouse notice that when you select it [26:53] there's this little arrow here that [26:55] turns out to be a significant point [26:56] later I'm moving fast because there are [26:59] a lot of slides here now Ken Viktor [27:02] remember was Ken here today yeah Ken was [27:05] one of the reviewers of the document and [27:06] he provided some input very very quickly [27:09] and he said why can you only select it [27:14] with the mouse [27:15] now this he must be someone who used [27:16] Emacs or something he wanted you to be [27:18] able to select from the keyboard [27:20] you know fourth line third word seventh [27:23] character you know she'd be able to [27:24] specify what you want and fourth [27:27] occurrence of the word Lisa you should [27:29] be able to say things like that and he [27:31] was maybe years either you know he sort [27:34] of didn't get it or he was years ahead [27:35] of his time whichever way you want to [27:38] look at it but it kind of this is what [27:41] he would like to use and we had to sort [27:43] of explain to him that our target market [27:45] didn't want to find the fourth [27:46] occurrence of Lisa in the third [27:47] paragraph they could see it with their [27:49] own eyes and select it with the mouse [27:50] and that was just fine now here was [27:53] something I wish we'd listened to at [27:55] this point it said that you were growing [27:59] windows from the lower right now I [28:01] showed you something before showing it [28:02] you grow it at the grew at the top I [28:04] think in this document it was sort of [28:05] ambiguous whether you grew it from the [28:06] top or the lower right and he didn't [28:09] like that idea he wanted to be able to [28:11] grow from any side and I do think that's [28:15] better and we didn't pay any attention [28:17] to him I think we were worried about the [28:19] screen space being too tight or [28:20] something moving a document within a [28:23] window he wanted to know whether we [28:25] could do what we now call live scrolling [28:27] as you move that elevator up and down [28:28] the shaft [28:29] the text to move continuously problem [28:32] was that the Lisa processor ran at what [28:35] 1/4 5 megahertz it didn't even seem that [28:39] fast to me I couldn't do it so some [28:44] things were just pragmatic he didn't [28:48] like the idea of cut and paste he wanted [28:49] us to use move copy and delete and [28:51] transpose excuse me just getting over a [28:56] cold in other words that you have two [28:58] things you would select move this to [29:00] there copy this to there transpose these [29:02] two things well this wasn't the first [29:04] time I heard that argument and in fact [29:06] this one of those religious arguments [29:08] like one button or one versus two [29:09] buttons do you want cut copy paste you [29:11] want you want move copy delete the [29:15] reason we went for well you can argue [29:17] either one it is two steps to do cut and [29:20] paste but there are a lot of advantages [29:21] to it one is that you know moving [29:24] deletes or start with the same thing [29:26] some people think that's good some bad [29:28] another reason is that you don't have to [29:31] be able to see the destination when you [29:33] are copying or cutting the source that's [29:35] the most important thing and on the [29:37] screen of limited size when you have [29:38] windows overlapping it's sometimes very [29:40] hard to get things all lined up so you [29:41] can specify two targets or you have [29:43] windows popping up and down you get very [29:45] confused the other thing is that I had a [29:48] secret agenda which is I thought that [29:49] the machine should be used not for what [29:51] they talked about office systems I [29:53] thought well that was good but I don't [29:54] want to be used just for that I thought [29:56] it was it would be a great machine for [29:58] publishing and that would be able to do [29:59] cut and paste into page layout which was [30:01] my own personal interest and so I was [30:04] advocating that because that was [30:05] definitely the way you'd want to do [30:06] paste page makeup but we did user [30:09] testing and the users slightly prefer to [30:12] cut and paste mam the wastebasket window [30:15] was able to not only have one thing you [30:17] cut out but multiple things you cut out [30:19] and he had some comments about how they [30:21] deal with the multiple items some of us [30:24] thought it should only hold one he [30:26] thought it was okay to have multiple [30:27] ones he lost in the end okay here's some [30:31] comments for me I was complaining [30:33] because nowhere in the spec did it say [30:35] what the response time should be it said [30:37] all about the user what they do with it [30:38] but it never said that this needs to [30:40] respond within half a second this knee [30:42] to respond within two seconds that needs [30:44] to respond within 10 seconds never said [30:46] it anywhere and I was worried if it [30:47] didn't say it [30:48] the engineers wouldn't do it if any of [30:50] you ever used to Lisa well it was really [30:54] slow I lost that argument they said [30:57] it'll be as fast as we can make it and [30:59] that was too slow then they'd have [31:03] talked about a really interesting [31:05] question which was well we want [31:06] hierarchical menus and we want keyboard [31:09] accessible commands so we seem to be [31:10] fixated on the way they do it and UCSD [31:12] Pascal you type a key get another menu [31:14] type of key get another menu and that [31:16] just bothered me and I thought there's a [31:19] lot of mistakes could be made people are [31:20] gonna forget which level of the [31:21] hierarchy they're in how they get back [31:23] out of it again it's just gonna be too [31:24] complicated and I didn't like it [31:27] so the will come to how that was solved [31:31] later the other thing I was complaining [31:33] about was the two button mouse we didn't [31:35] use the second button for much else for [31:37] a lot of different things we didn't use [31:39] it for much and when we did there was no [31:41] way to remember which button did what [31:43] and so I just wanted to eject the second [31:45] button and just have one and that's [31:48] August 13th operations on Windows I [31:54] proposed here that that we have this box [31:58] in the lower right hand corner that we [32:00] use for growing the window instead of [32:01] doing it from the top and also I thought [32:04] there would be ought to be a way to [32:05] split a window into panes and then I had [32:12] a whole section here about moving a [32:14] document within a window I go and talk [32:15] about scrolling and why I don't like the [32:18] way it is and then I said why don't we [32:20] put you know one arrow at the top of the [32:22] scroll bar one at the bottom of the [32:23] scroll bar and I actually prototyped it [32:29] those are you know that who know the [32:31] Xerox star know that the Xerox star had [32:34] that feature those of us who were from [32:37] Xerox were under a confidentiality [32:38] agreement that we weren't allowed to [32:39] tell the other Apple people about the [32:42] star and so as you go through this [32:45] you'll sometimes find that things seem [32:48] like isn't it an obvious solution you [32:49] could have got it from the star but none [32:51] of us are able to say why this how the [32:52] star was we were only live to comment on [32:55] other people's proposals and make new [32:57] ones and it was an interesting kind of [33:00] dynamic but gradually over times abuse a [33:03] few star like things crept in but not [33:05] many until the star actually shipped or [33:07] was announced in June of 81 but here we [33:10] are in 1980 and I suggested we try and I [33:13] shaped cursor for text selection instead [33:16] of an arrow and actually is something I [33:18] had done in my early years at Xerox in a [33:20] particular word processor I'd done I [33:22] also said I wish the cursor would be [33:23] bigger than 16 by 16 to go to 32 by 32 [33:27] which which we eventually did menus this [33:32] was what August 18th I think we should [33:36] experiment with menus at the top or [33:37] right of the window or even totally [33:39] detached at the top of the screen full [33:41] width so up to this time we were doing [33:43] sort of the way Windows does and every [33:45] window had its own menu and I was [33:48] thinking you know maybe we got to put it [33:50] at the top and you'll see why in a [33:51] minute when you see the pictures and [33:53] then I said hierarchical menus well [33:55] maybe we have to have and we can't get [33:57] rid of them but could we make a mouse [33:58] access instead of having to type all [34:00] these keystrokes so these were my two [34:03] causes of the moment Barry Smith who was [34:06] in product marketing responded about the [34:08] same time and he said [34:16] at this point bill had moved the menu [34:19] bar that edit cut copy pasting from the [34:21] bottom of the window to the top of the [34:23] window that was the change and Barry [34:26] liked that he likes him a lot better [34:27] near the top call him near the top and [34:29] then he wanted to know though what would [34:31] happen if there were a lot of menu items [34:33] and could you could the menu scroll and [34:36] those of you know the history of the Mac [34:37] user interface know that at some point I [34:40] don't know 1990 or so or we start adding [34:42] scrolling to Mac menus but we didn't [34:45] have them in the early Macker and the [34:46] Lisa and I was against it because that [34:49] would have encouraged people to have a [34:50] lot of menu items and I didn't think [34:52] they should so I kept that from [34:53] happening [34:55] you can't constrain people though he [35:00] also did not like the fact that you had [35:01] to type multiple keystrokes through a [35:03] hierarchical menu and so he was thinking [35:06] maybe we could get to something where it [35:07] was a single keystroke to invoke a [35:08] command no matter where it was in the [35:10] hierarchy and I think Bill was was [35:12] experimenting with that at that time and [35:14] he also didn't like the two button mouse [35:15] he thought it was confusing so August [35:19] 18th I guess that was August 13th before [35:21] us a few days later after intense [35:25] political polite discussions heated [35:29] sometimes bill and I decided hey let's [35:32] get rid of the second button the shake [35:33] things up one button Mouse so we wrote [35:35] this memo together so that people knew [35:38] we were realest realest erious wrote [35:41] Santa to the usual suspects and we say [35:45] we recommend a change to a one-button [35:46] Mouse and whoever wrote this comment and [35:49] I glad I don't know who was said well [35:51] maybe not sure that's a good idea but [35:57] turned out trip Hawkins who later went [35:59] on to found Electronic Arts in 3do was [36:01] the product marketing manager he loved [36:03] it he said that's what we got to do we [36:05] got to think more and a [36:06] consumer-oriented really simple get rid [36:09] of all this you know computer science [36:12] concepts and give give people something [36:14] real simple he liked the one button so [36:16] after that we had a period sort of [36:18] goodwill where we were able to to do [36:21] user interface design without a lot of [36:22] complaining which ended after a while as [36:24] you'll see [36:26] well if you only have one button some [36:29] things are hard to do you can drag a [36:31] short selection but what if you want to [36:32] select something three pages long it's a [36:35] long time to drag from one end to the [36:36] other so we thought okay well we'll use [36:40] a shift key on the keyboard since this [36:42] doesn't happen very often you'll click [36:44] the beginning of the selection you'll [36:45] scroll to where you want to go hold down [36:46] the shift key and click the end of the [36:47] selection so that's where we got this [36:49] shift button thing adjusting selection [36:51] and then we said the one button Mouse is [36:54] superior look at the reasons it's [36:56] appearing human factors you don't have [36:58] to think of a name for each button [37:00] that's important when you're trying to [37:02] ship product in six months the mouse [37:04] will be cheaper to make because one less [37:06] button I mean just think of what the [37:07] cost of that little a little switches [37:11] it'll be and something a little more [37:14] profound someday when we start having [37:17] other things in mice on our computers [37:18] like joysticks for games tablets [37:21] touchscreens we'll be glad we only had [37:24] one button because that that one finger [37:26] or whatever or pen that the user using [37:28] is the button and that turned out to be [37:31] sort of prophetic although very few such [37:35] pointing devices take advantage of that [37:37] in fact they usually have buttons anyway [37:41] okay so August 22nd there's a revised [37:43] version of the user interface and it [37:46] says this is it now we're done so now [37:49] instead of a week you only have three [37:50] days to return comments because we must [37:52] be done there couldn't be more than one [37:54] or two comments they're probably just [37:55] typos or something at this point we were [37:58] working on this interface for a month [38:00] already [38:03] okay so what did it happen this version [38:07] had a concept called dim highlighting [38:09] this was introduced for the first time [38:11] that I know of here that if some of the [38:15] menu items instead of eliminating menu [38:17] items that weren't relevant which we [38:19] would dim them we would we've grade them [38:22] out instead of being black type they'd [38:23] be kind of in gray type and then we said [38:26] they can be invoked anyway if the user [38:27] wishes although they wouldn't do [38:29] anything and and we also were going to [38:32] use dim highlighting for another purpose [38:33] which was to show toggles and as you [38:35] know that was later changed to have [38:37] check marks instead of just using this [38:39] for two purposes because that was very [38:40] confusing but at that point we thought [38:43] we'd get away with one mechanism for [38:44] both things and it was Occam's razor one [38:46] mechanism will do they don't have to [38:48] well turned out that one didn't do so we [38:51] had to apologize to Ockham now as you [38:56] know you can do command keys on the Mac [38:58] or at least a keyboard and you can [38:59] invoke commands from the keyboard and we [39:01] knew it was important to reserve some [39:03] for the most common commands because [39:04] might be okay for an application to use [39:07] you know command shift J for something [39:10] but we wanted to be sure cut copy paste [39:12] undo were the same for everybody [39:13] bold italic underline and normal we [39:16] thought these would be used the most and [39:17] so they're all sequence I'm not going to [39:19] go through them all today but we had [39:20] various theories as you see here it's [39:22] rather different from what we ended up [39:23] with ZX CV was sort of the order on the [39:27] keyboard cut paste copy undo was the [39:28] order in the menu so that was it over [39:30] time we refined this list led to many [39:33] very heated arguments people would [39:36] rather argue about trivial things like [39:37] this than important things but we did [39:39] have fun arguing does anybody remember [39:43] what they really are today [39:47] that's right controlling trilby and then [39:50] for a cut X word stock oh I didn't know [39:57] that well we did X because it was a [40:04] cross out and we did V because it [40:06] pointed down like this and you were [40:07] inserting like an upside down carrot [40:08] that was paste and Z was the closest one [40:13] because we figured you'd undo a lot and [40:15] C for copy that was easy so that's how [40:18] we did that okay so this was the anatomy [40:21] of a window we had title tab now notice [40:23] next to the title tab we now moved the [40:26] menu to the top as was commented upon [40:28] someone obviously had got a sneak [40:30] preview draft and then earlier memo and [40:32] look there are menus that are popping [40:34] down from the menu bar first time when [40:37] it was the bottom that didn't make a [40:38] whole lot of sense but now that it was [40:39] the top bill kid instead of having the [40:41] menu come here and replace the previous [40:43] menu like it did in UCSD Pascal he could [40:46] use the second dimension and have a menu [40:48] pop out to the bottom don't ask us how [40:50] we were gonna do third level menus but [40:52] at this point this was good enough we [40:53] had two levels of menus and we were [40:55] happy over here there was the elevator [40:58] shaft now we called it the thumb instead [40:59] of the elevator because we wanted to [41:01] have horizontal scrolling too and [41:02] elevators in those days didn't go [41:03] horizontally although Scientific [41:05] American says they're about two and then [41:07] we had continuous scrolling arrows and [41:10] noticed the arrows point this way but [41:12] when you clicked on these arrows [41:14] the text was going to move that way as [41:17] opposed to the other way which is what [41:18] it does now and that was an interesting [41:21] argument that I think I have covered [41:22] here too okay we had active folders so [41:27] we had to have passive folders the [41:28] folder that wasn't active was passive [41:30] now look at the bottom who has never [41:32] seen Mac OS 8 and when you drag a window [41:36] to the bottom of the screen it just kind [41:38] of pops and becomes just a tab and then [41:40] you can click it and it'll end this [41:41] it'll slide open guess what that wasn't [41:43] the original design in August 1980 that [41:47] what we were going to do on the Lisa and [41:48] it only took 17 years to implement [41:54] Microsoft did something sort of similar [41:56] we're not really the same in between [41:58] there's some there are some differences [42:00] but I won't comment right now now so [42:03] what happened if the window was narrow [42:04] well if the neighborhood's window was [42:06] narrow this shows that the menu wouldn't [42:08] fit it would stick out and if it was [42:11] over in the right and it couldn't stick [42:12] out through the right he would stick out [42:13] to the left and it was too close to the [42:16] bottom of the screen then it would do [42:17] this and so bill showed all the cases [42:20] and I you know a lot of us weren't too [42:23] happy with this so I went back to Bill [42:25] again I said bill the top of the screen [42:27] the top of the screen none of us liked [42:31] the idea of a 2-lane menu which is the [42:32] way Microsoft solved it some version of [42:35] Windows and here's an example of a menu [42:38] item being selected notice he's now got [42:39] highlighting of the menu title and he's [42:40] got highlighting of the menu item bill [42:42] had a long night one night and came back [42:44] with that and we had the help menu in [42:46] here the way we operated was that bill [42:51] spend every night programming and the [42:52] next day we would do testing and arguing [42:53] I don't know when he slept actually did [42:56] he sleep and then he would do it again [42:58] every night now here's a memo from Gail [43:01] Pilkington and publications the people [43:03] writing manuals for the Lisa were key in [43:05] commenting on the user interface design [43:07] had a lot of suggestions and a lot of [43:09] good criticisms and Gail actually had [43:12] been a Xerox before so she said I don't [43:18] like pull down menus I think they're [43:20] ugly and they cover things up so you [43:23] know I'm trying to cut something and the [43:24] first thing it happens as menu come down [43:26] it covers up the thing I'm trying to cut [43:27] no good and then bill had this feature [43:30] that you could push she still has you [43:32] could drag the mouse along the menu bar [43:35] and the menus would pop up one at a time [43:37] and we thought this is great the user [43:39] can see all the available commands in [43:41] one swoop well she thought it was [43:43] terrible that you that you had to do [43:45] that to see all the commands I didn't [43:47] understand because in UCSD Pascal you [43:49] had a you know hit keys for ten minutes [43:51] to see all the commands so I didn't [43:53] quite get that she also didn't like the [43:56] idea of dim highlighting she thought we [43:58] should just suppress the items [43:59] altogether so we had dissent in the [44:01] group [44:02] and we always have to sit in the group I [44:04] just chose that because there's a memo [44:06] that I had an illustrative of dozens of [44:09] people who had to send over every issue [44:10] nothing specific about Gail [44:12] in particular she said you know last [44:14] version of the spec I thought we only [44:16] had one issue which was when the menu [44:18] when that little menu bar at the bottom [44:20] was full we didn't know what to do about [44:23] it [44:23] and you should have just solved that [44:26] instead of moving it to another place [44:28] adding dropdowns and all these other [44:29] things you over designed to solve the [44:32] problem and besides they don't like X [44:34] you know Z for cut next for a paste she [44:37] had her own proposal that was Nimonic [44:41] September 22nd was now a month after I [44:44] first said bill why try to top at the [44:47] screen they all had a very long night [44:48] and came back with a menu at the top of [44:51] the screen with command keys check marks [44:57] every day and he had everything in one [44:59] night he invented the entire mechanism [45:01] of the menu bar topped the screen to me [45:03] it was just a place to put what he [45:04] already had but he once he realized it [45:06] was there he could do a lot of other [45:07] things and he implemented amazing amount [45:10] in one night so he wrote a memo to [45:14] justify it and said there are some [45:17] problems with it you have to give to go [45:19] way further to reach the menu that's a [45:21] problem but we decided since there were [45:23] a command keys and he made it very easy [45:25] to have a single stroke command key in [45:26] this version this all-night version that [45:29] okay if you really wanted to go to that [45:31] menu item a lot you can hit the command [45:33] key on the keyboard instead and you [45:34] wouldn't move the mouse at all [45:35] so that solved that problem the other [45:37] problem was that if there were a lot of [45:38] windows open you wouldn't really be sure [45:40] which menu applies to which window and [45:43] sure enough that's still true today but [45:46] trade-offs you know why was it good it [45:49] was good because you got rid of all [45:51] those ugly cases of sticking out menus [45:54] and and also the dialog box that came up [46:02] when you had a command with parameters [46:03] could always go in the same place and we [46:05] had a problem before with the dialog box [46:07] coming up and covering up the menu half [46:08] the window and so on and so we could put [46:10] it in a standard place under the menu [46:11] bar [46:14] so this was all the result of his long [46:16] night of invention and then a couple [46:20] more examples we could put some global [46:22] commands in the menu before we didn't we [46:24] hadn't figured out yet we were to put [46:25] the global commands they didn't really [46:26] go anywhere because every menu was in a [46:29] window but now having a menu at the top [46:30] we could have global commands like you [46:33] know shutting down or whatever and then [46:35] we could widen the title tabs because [46:37] the menu didn't take up room at the top [46:40] of the window okay now we're in February [46:44] 81 so we've leaped ahead now several [46:46] months we're all programmers are busy [46:49] implementing stuff they're designing an [46:51] operating system telling us that we're [46:53] holding up the project when they're [46:54] still designing the operating system and [46:56] where the yep I was in charge of the [46:57] applications group and we were the [46:59] reason the schedule is behind schedule [47:00] we the project was behind schedule even [47:03] though they still were designed the [47:04] operating system so I wrote a memo for [47:10] other people who were going to help out [47:12] with user testing and I read men made [47:15] guidelines for user interviews so never [47:17] argue with each other in front of the [47:19] interviewee after they leave you can [47:20] shoot each other doesn't matter but now [47:22] well the not well the subjects there [47:24] don't patronize a subject don't show off [47:27] your knowledge I wrote this because I [47:28] would be bringing engineers into the [47:30] room and I'd sit him in the back and the [47:33] user would struggle to do something and [47:35] the engineer would run forward jump out [47:37] of the chair run forward grab the hand [47:38] of the person and say it's easy you just [47:40] do this and even though they still were [47:45] designed the operating system so I wrote [47:50] a memo for other people who were going [47:53] to help out with user testing and I read [47:56] men made guidelines for user interviews [47:59] so never argue with each other in front [48:00] of the interviewee after they leave you [48:02] can shoot each other doesn't matter but [48:03] now while the not while the subjects [48:06] there don't patronize a subject don't [48:08] show off your knowledge I wrote this [48:10] because I would be bringing engineers [48:11] into the room and I'd sit him in the [48:14] back and the user would struggle to do [48:17] something and the engineer would run [48:18] forward jump out the chair run forward [48:20] grab the hand of the person and say it's [48:22] easy you just do this [48:23] and I go no you don't get it so we tried [48:28] gags we tried all kinds of things but [48:30] finally I said look I'll let you in the [48:32] room if you if you you know observe [48:35] these rules don't plant the answer take [48:37] turns asking questions if there's [48:39] several of us there so we get different [48:41] we may each see different things and [48:43] mainly try to get them to do the talking [48:45] this is not a lecture by you this is a [48:47] test of them and you want them to [48:49] verbalize all they possibly can so [48:53] difficult thing that would happen is we [48:55] bring somebody in to prove that the [48:58] user-interface thing they put in was [48:59] absolutely right and we were totally [49:00] wrong and then the end user would [49:03] usually prove we were both wrong it was [49:05] hard to both four things were hard to [49:06] use and then we'd have to go back and [49:08] redesign it or half the time the user [49:10] would say well why doesn't it just work [49:12] this way and we look at each other [49:14] sheepishly and fix it here was an [49:17] interesting one for those of you who do [49:18] user testing you don't if you're not [49:21] named things you don't just say what [49:23] would you call this people paralyzed [49:24] like oh you're asking me to invent a [49:26] name they just get totally paralyzed [49:28] they can't think of a name but do you [49:30] say something like if someone were to [49:32] bring you this if you wanted someone to [49:34] bring you this what would you say bring [49:35] me the and then people would come up [49:38] with something it actually worked pretty [49:39] well so people have similar techniques [49:42] I've heard for other things [49:46] April 3rd 81 well at this point they're [49:49] starting to get very nervous about the [49:50] schedule the operating system design was [49:52] done I think and so we were that we were [49:54] holding things up and they decided the [49:57] problem was that we're these engineers [49:59] instead of just writing code like like [50:00] good sheep we're going in and you know [50:02] worrying about whether it was usable or [50:04] not so they decided to take all power [50:08] away from us and put all the [50:10] decision-making in to a council so who [50:12] was on the council first person was the [50:15] division controller in finance second [50:18] one was the VP of Sales who was always [50:20] on the airplanes third one was the head [50:23] of publications for Lisa which was a [50:24] good idea fourth one was kind of the [50:28] architect of the group that was a good [50:29] idea and the fifth one was somebody in [50:31] product marketing who you saw had made [50:33] comments before that were pretty good [50:34] ones so I was looking at this going okay [50:37] maybe we can live with this but the main [50:39] thing they did that was very good as we [50:40] instituted a process for what to do if [50:42] you had a user interface issue don't [50:44] spend eight hours arguing with each [50:46] other [50:46] write it up make your points get back to [50:48] work and then we have a process for [50:50] dealing with it but the council will [50:52] rule by majority vote so I think what a [50:56] way to design a user interface by [50:58] majority vote of the controller the [51:00] sales VP okay [51:01] [Music] [51:04] but we started this motivated those of [51:08] us who decided we had to do this the [51:10] right way to actually run more user [51:12] tests as we thought well maybe we can [51:14] overwhelm them with facts and that'll [51:16] overcome this reliance on on votes of [51:19] committees so Chris Dewar I can't [51:23] remember Wallis Judd I think Wallis [51:25] maybe you've worked at Xerox before this [51:26] I can't remember these are two people [51:28] that got involved with user testing the [51:30] filer is what you think of them as the [51:31] finder today and they ran some tests and [51:35] came up with all sorts of comments I've [51:37] underlined a couple here like there [51:38] needs to be a way to select multiple [51:40] icons at once while they weren't icons [51:42] yet even but ways to manipulate multiple [51:45] things at once [51:48] but basically after making some comments [51:51] they said you know this is a good idea [51:53] they're their own personal experience [51:54] with testing was he hired a testing firm [51:56] you spent a lot of money [51:58] three months later you came back with [51:59] test results and then you did some [52:01] things and he did it again you know it [52:03] took forever and they said you know or [52:05] they used or they or they would take a [52:08] book and give it to a subject have the [52:09] user read the manual try out the manual [52:11] videotaped the whole thing and and she [52:14] said you know this wasn't bad just [52:15] sitting down next to the user and having [52:17] a dialogue with the user trying to get [52:19] the user talk and express what they're [52:20] doing is not a bad way to do it so that [52:24] was good that we we got the the people [52:27] doing the user testing to adopt this [52:29] much simpler and cheaper paradigm we [52:33] tape recorded the sessions and they [52:34] found that useful they went back and [52:36] listened to the tapes and however the [52:39] tape for hard to hear because the people [52:41] in the room were making noise so they [52:42] asked the observers to be quieter I'm [52:46] gonna have to remove these from the box [52:47] now [52:49] it starts going a little quicker now but [52:55] one problem she said was that we're only [52:56] testing Apple employees we would get [52:58] Apple employees the day they took the [53:00] job the day they started work we'd get [53:02] them in orientation the person running [53:04] orientation would say well after [53:05] orientation you'd go straight to work or [53:07] you go you know be a user test subjects [53:10] and hey you get out of work for a few [53:12] more hours so we got a lot of people to [53:14] come over and the user test subjects and [53:19] almost all of them had very little [53:21] computer experience because although [53:22] Apple would have liked to hire people [53:23] with a computer experience in those days [53:25] there weren't very many so and here's a [53:28] typical schedule one it was the we [53:31] weren't trying to - an hour any more one [53:33] person at a time one each day nine [53:36] o'clock one o'clock 3 o'clock much more [53:38] civilized and and we had a checklist [53:44] that the tester went through before each [53:47] session have your mind of the subject of [53:51] whatever I have no idea whatever they're [53:54] supposed to remember there's some list [53:56] is there coffee available do you have a [54:00] blank questionnaire ready for the person [54:01] because we haven't fill out a [54:02] questionnaire at the end but you have [54:04] the checklist ready have you rewind the [54:06] tape have we really have we rewound the [54:09] tape for this talk today have you [54:11] rebooted because it crashed if you [54:13] didn't are the cue cards in the testing [54:16] room etc after each session did you [54:17] cover up ELISA again because the room [54:19] had a window and at night you could see [54:21] in the window and see the secret Lisa [54:22] machine you'd close the blinds for [54:25] double measure etc etc so we had a lot [54:29] of things to remind the tester of al [54:31] that's a good idea if you ever run user [54:32] tests have a checklist now Wallace the [54:35] guy who received the last memo wrote [54:37] this memo about some testing results and [54:42] he said he spent about an hour subject [54:45] which was pretty typical that's why I [54:46] like to do testing myself and he said [54:49] well some functions were still unclear [54:50] in particular seven subjects out of [54:53] how many doesn't say but not a lot well [54:59] not very many most of the subjects had [55:00] trouble with undo last command they've [55:02] never seen such a thing in a program [55:04] before and they just didn't quite [55:07] understand what it meant then there are [55:10] other ones they understood but they [55:12] didn't feel comfortable with somehow one [55:13] was printed copy one was mail a copy [55:15] well I believe the reason was that we [55:17] didn't have printers yet on the machine [55:19] and we didn't have email in fact we [55:22] never had email so we're probably just [55:25] shown these and not really couldn't use [55:26] them and so of course they didn't [55:28] understand them graphics editor user [55:33] testing this was Lisa draw which became [55:36] Mac draw later implemented by the same [55:38] guy actually mark cutter and we wanted [55:41] to test various things as it was as [55:43] confusing as that confusing can you use [55:44] what happens if you try to use rulers at [55:46] the same times at crosshairs is that too [55:48] many things on the screen and [55:50] considering all the widgets that are [55:51] there today in a photoshop or something [55:53] why were we worried but there was even [55:57] worried about whether people could use a [55:58] graphics program because well you know [56:00] people most people can't draw so they [56:02] wouldn't use a graphics program and we [56:06] also want to know the error rates and so [56:08] a lot of counting was being done and [56:10] looking at the tapes to see how many [56:13] times people made errors and try to [56:14] reduce the error rate so we're trying to [56:17] be sort of scientific even though we [56:18] were on this headlong schedule you may [56:20] have noticed as six months have already [56:21] gone by and we haven't shipped effect [56:23] it's about a year at this point ok so [56:26] June 81 Pat Marriot writes this memo [56:31] about validating the prototypes and trip [56:33] Hawkins at this point is that his wit's [56:34] end he says aren't we done with all [56:36] these decisions yet let's get on with it [56:38] why are we still talking okay here are [56:43] some of the issues should the arrows [56:45] point apart away from each other or [56:47] towards each other because after all the [56:49] text moves in the direction that's [56:51] opposite of the our direction we could [56:52] flip the arrows and make a point to each [56:53] other and I wrote up a comment here yeah [56:56] we can wait on September to test this no [56:58] hurry I wasn't in sync with product [57:00] marketing who wanted all the decisions [57:02] to be made because I knew that changing [57:04] this was a matter of [57:05] you know about three minutes of [57:06] programming so what was the hurry I'd [57:07] rather work on harder issues like you [57:09] know getting the software to fit in [57:10] memory or something and there were some [57:13] questions about clicking and folder [57:15] click well by the way they worked folder [57:17] when you see folder here we used to call [57:19] Windows folders I forgot to mention that [57:21] earlier might been confusing here every [57:23] window is a folder even a document when [57:24] there was a folder not just the ones on [57:26] the desktop and we wanted to test we [57:30] wanted to do some tests on that the [57:38] whole thing [57:41] the mirror thanks okay what was that [57:47] [Music] [57:49] okay [57:50] October 81 October 80 when we took the [57:53] Chris's advice to heart Chris DeRose [57:55] advice and we did outside testing 24 [57:57] people who did not work for Apple we're [58:00] done in in a one-way mirror room in Daly [58:03] City they didn't know what company it [58:04] was that was showing it you know blind [58:06] test the psychologist had their day we [58:09] got to do this you know right and this [58:12] was the thing that there was no way [58:14] we're ever gonna do this too expensive [58:15] takes too much time but at this point [58:17] product marketing was going well maybe [58:19] we better test this on real people and [58:21] we wanted to know whether they could [58:23] read the manual and install the hardware [58:25] and get the thing running by themselves [58:27] and it turned out that it was it was ten [58:31] minutes on average that a person could [58:33] from they trying to open the box it'll [58:35] time they get set up and running [58:36] something we thought that was pretty [58:38] good [58:38] those of you who are at Apple later know [58:41] that on the Mac it was even easier at [58:43] the beginning but as the Mac got more [58:45] and more developed that number going up [58:47] and when that numbers started getting [58:49] about an hour Jon ligase got extremely [58:53] upset and he ordered new user tests and [58:55] they ran user tests about that and got [59:00] the time back down to ten minutes again [59:01] but there were some bad years in there [59:04] it was very hard to open a Mac box and [59:06] figure out how to set it up two-thirds [59:08] of the users say they felt comfortable [59:09] using the product as a result of the [59:10] first experience and about half of the [59:13] people who were a half of the third that [59:15] were uncomfortable were people who [59:17] happen to run into crashes and so on [59:19] which probably affected their trying to [59:22] beat the clock on setting up the systems [59:23] kind of frustrating and we had a [59:25] question there they had to fill out [59:26] these this is the result of the [59:28] questionnaire so those numbers are how [59:29] many people had that answer and they [59:31] didn't think it would be that difficult [59:32] to set up before they started and when [59:36] after they did it they found that it was [59:38] very easy [59:41] all right now it's still at 1981 now I [59:45] should mention that the Xerox star was [59:47] announced that NCC or something National [59:50] Computer Conference I think in June of [59:52] 81 and the Sun Lisa people flew out [59:57] there got to look at it Bill Atkinson [59:58] went I believe and and Steve Jobs went [60:00] or maybe he sent bill and didn't go [60:02] himself I can't remember at this point [60:05] we had a guy named Greg spike letter who [60:07] had joined the group working for Ellen [60:09] old in the training lease of training [60:11] some of you may know Greg because he's [60:12] he's recently more recently been an [60:14] entrepreneur sold a couple of companies [60:16] and Ellen is sending a memo to various [60:20] people copy and Greg and [60:22] responsibilities charting for user tests [60:24] now that user interface counsel we [60:26] talked about before with the majority [60:28] vote I don't know if they ever even have [60:29] one meeting now we kind of overwhelmed [60:32] them somehow and by this time a few [60:34] months later we had have a new process [60:36] so the new process was that based on [60:39] some lecture we've all all went to bytes [60:40] a management consultant we were going to [60:42] have people with authority people with [60:44] responsibilities people who are [60:45] consulted people who were conformed and [60:46] basically product management hell they [60:48] had all the authority since since Ellen [60:52] worked for them and engineering and [60:55] training which is Ellen's group had all [60:58] the responsibility so we did all the [61:00] work and then product management would [61:02] bless it and say that hey February 1982 [61:06] year and a half is gone by engineering [61:10] user test this is a sample test we had [61:12] there's a write-up up a test we had a [61:13] standard form to write him up in now [61:15] we're getting real formal here not just [61:17] you know run into Steve Jobs after the [61:19] test and tell him what happened we were [61:21] we were writing you know formal reports [61:23] what the person did how they did I'll [61:27] skip over a little bit of this oh well [61:29] the person had a john org chart and here [61:32] not having a graphics program myself [61:33] that had printing working yet I was [61:35] doing it in the word processor and [61:37] drawing little org charts here but they [61:38] were really doing it with a with a [61:40] drawing program and by the way the thing [61:43] that went over the product marketing [61:45] people was when we gave them the drawing [61:47] program it was the first one to get [61:49] working [61:49] I had mark cutter just try to get it [61:51] really reliable and don't worry about [61:53] too many features and then we gave all [61:55] the [61:55] product marketing people elisa some [61:57] temporary operating system in the [61:59] drawing program that's it [62:00] and single tasking operating system at [62:02] that time and and they all started using [62:05] it to make foils and you could do tax [62:08] you could do boxes you know and they all [62:09] did foils and they fell in love with [62:10] ELISA and they even start to understand [62:12] the Lisa and gradually the complaints [62:14] started dying down and they start [62:16] trusting like maybe we knew we were [62:17] talking about so that was a very good [62:18] political move and I always made sure [62:20] mark gave him you know good stable [62:22] versions of the draw program they draw [62:25] floorplan and then at the end it [62:27] comments to make and you could see that [62:29] the comments were summarized and people [62:31] were basically really positive about [62:33] this program sometimes they were [62:35] complaints and we address them now this [62:38] particular test was I guess run by Greg [62:39] stack leather and me he probably ran the [62:41] test and I was in the room or sometimes [62:42] we switched off and he made our comment [62:46] which was that i-beam which was the [62:47] cursor that you used to select text when [62:49] the person would start typing that [62:50] little cursor was covering what they [62:51] were typing so he suggested that we get [62:53] rid of it well there was no mechanism to [62:55] get rid of it in the software but he [62:57] didn't know that and so it was good that [62:59] he was there and he suggested it and we [63:01] then implemented that and I can I [63:04] pointed out that the stretch handles [63:06] were a problem initially what Mark had [63:08] done was he studded the outside of your [63:11] object like every fourth pixel with a [63:13] little handle so big objects had you [63:15] know beads all the way around the [63:17] outside it was like a necklace and this [63:21] really confused the people and I [63:22] suggested we just had the four corners [63:24] and the four sides and that fall and [63:25] that is the way it ended up first I was [63:28] showing you all the good comments we had [63:31] a lot of stupid suggestions too and now [63:34] remember we had user test guidelines [63:35] before Greg wrote some new ones and here [63:39] he was trying to address this to people [63:43] who might have been used to formal user [63:47] testing and had never experienced what [63:49] he was calling development testing and [63:52] you know things you test after the fact [63:54] and things you test while you're still [63:55] in development so he made some [63:57] guidelines and they were interestingly a [63:59] little different from mine [64:00] a different level and I think I didn't [64:05] bring them just the fact that he did [64:09] okay all right good near the end here [64:14] excuse me may 18 2010 so Greg said we [64:25] should run development tests so we did [64:26] we ran development tests and I taught 30 [64:29] Apple employees over a period of several [64:31] weeks several different applications and [64:34] to see whether you know we were on the [64:36] right track and by this time people were [64:39] accepting this is a good approach [64:41] because it was a lot better to be [64:42] informed by a user test than to just see [64:44] who talked the loudest in an argument [64:46] who had the most political clout and [64:48] what I found was that the way we taught [64:51] it made a lot of difference you could [64:52] take the same user interface and teach [64:54] it a different way and people get real [64:55] confused or understand it and or make [64:59] more mistakes or fewer mistakes and [65:00] terminology made a difference also so we [65:03] then started a terminology project that [65:04] Ellen old ran which ended up and I many [65:07] slides I didn't bring on that and that [65:09] ended up with the file menu the Edit [65:10] menu etc and as you know today and the [65:14] various commands that were in them you [65:16] know choosing all the words for [65:17] everything that had pretty much survived [65:19] into the Mac let's say 95% of them [65:21] survived into the Mac and most than when [65:23] at the windows also so that was a very [65:25] healthy terminology exercise is it a lot [65:28] of testing and a lot of debating lettle [65:30] linguistic analysis and one thing I say [65:34] here is that after all these tests I [65:35] discovered a lot of things that could be [65:37] better but we do have to ship a product [65:39] and it's been a long time since I've [65:41] been a park so okay we'll ship a product [65:44] but couldn't we fix these in the next [65:45] release well little did I know there [65:51] would only be one more release and then [65:54] we had a process again where we were [65:57] going to be very careful about [65:58] prioritizing all the user interface [66:00] changes and measuring their scheduled [66:02] impact and being sure they were worth it [66:04] didn't have to be zero necessarily but [66:06] it had to be worth the worth the [66:08] scheduled impact [66:09] I think by this time people were really [66:11] realizing that our only differentiation [66:13] was ease abuse and it better be really [66:14] easy and then I went through some [66:17] problems this thing went on by the way [66:18] for about 30 pages so I'm just going to [66:20] show you one page cursor changes were [66:23] not consistent and they were cursors [66:26] changing shape everywhere the i-beam [66:28] wasn't the only one like no matter where [66:29] you went in the window you got a [66:30] different cursor shape and I thought you [66:31] know that's too many let's simplify that [66:33] we had double and triple clicks at that [66:35] time and some people it was the amount [66:39] of time between the clicks mattered and [66:41] you couldn't control it so I suggested [66:44] we had a preference item that people [66:46] could control that time which of course [66:47] is what you do today in the Apple's [66:49] mouse control panel and I also had some [66:52] comments about how you teach it and in [66:54] fact when I looked at that whole memo [66:56] I'd say there were at least as many [66:57] comments I made about how to teach it as [66:59] to what it is and then months earlier I [67:03] think it was really one month earlier [67:05] Gregg cycle Heather had pointed out that [67:07] the i-beam should disappear when you [67:08] type I don't know if I forgot or [67:10] whatever but I suggested the same thing [67:11] again or observed it and I propose we [67:14] hide the cursor and here in the margin [67:17] as I'm something I wrote later saying it [67:19] has been implemented if you worry about [67:25] the where we're wondering about the [67:28] cursor changing to a clock or an [67:30] hourglass various other things that was [67:33] something that was actually at Xerox [67:34] Xerox had that at Parc and the thing was [67:37] the way we did it was we would wait a [67:39] certain number of seconds and if the [67:41] operation wasn't done we would put up [67:42] the hourglass trouble was by that time [67:44] the operation was probably done so we or [67:48] some of the applications forgot to put [67:50] it up at all so we came up with some [67:52] clever algorithm for having it come up [67:54] automatically and and disappear at the [67:57] right time and all that sort of stuff [67:59] now here's the big change the star had [68:02] come out a year earlier than this and we [68:04] were all set we were all told by [68:07] management and ourselves don't pay any [68:09] attention to that we've got a product to [68:10] ship but in everybody's heart they kind [68:13] of felt like you know most of the lease [68:15] is easier to use than the star but [68:17] there's one thing about the star that is [68:19] really really nice which is those the [68:22] little icons on the desktop there was I [68:24] con envy that was very serious and we [68:27] weren't allowed to work on it but around [68:30] March of that year some of the guys got [68:33] so troubled by it that they started [68:36] working at home on weekends on an iconic [68:40] filer we called it and a person was a [68:43] big secret and they started letting [68:44] people in on a little bit at a time and [68:45] finally everybody found out about it and [68:47] there was a major uproar that people [68:49] would dare to make changes at this late [68:50] stage but they'd already implemented it [68:53] and so you couldn't complain about the [68:55] schedule impact and the person writing [68:58] the manual thought it was kind of a good [68:59] idea so she kept saying it wouldn't [69:01] affect her schedule and so nobody knew [69:07] what to do so I said so I made a deal [69:08] with trip Hawkins and Wayne rosing who [69:11] was the engineering manager at that time [69:12] we'll run user tests and let the users [69:14] decide so I ran tests and you know what [69:17] it turned out that the interface we have [69:18] which I won't bother to show you but it [69:20] was kind of a more like find file it's [69:22] the closest thing I can think of was [69:25] actually equally good in terms of ease [69:27] of learning equally easy to use same [69:30] error rate very low people understood [69:33] both there was no difference and we [69:35] already had one so why did we change and [69:37] then the last question I asked everybody [69:39] is which do you like better and they all [69:41] like the picture one better and so we [69:44] made a decision that nothing to be [69:46] nothing to do with ease-of-use nothing [69:48] to do with these are learning nothing to [69:49] do with error rate it wasn't human [69:51] factors decision at all in the [69:52] traditional sense it was decision based [69:54] on what customers liked and there's a [69:58] whole history I won't go into here about [69:59] the icons they didn't actually come from [70:00] the star they came from IBM did I'm not [70:03] gonna get into that long story you can [70:05] read the Proceedings of Apple versus [70:08] Microsoft lawsuit or something to find [70:09] that out [70:11] okay so this is my last slide here so [70:15] and this was a memo that included on it [70:22] Chris Espinosa who was in the Macintosh [70:25] group at that time and the Macintosh [70:27] people who had been doing a totally [70:28] different user interface at some point [70:29] decided they really ought to do some [70:31] subset of lease user interface instead [70:33] and so they wanted to get a lot of ideas [70:36] from us as they design their user [70:38] interface so we started copying the Mac [70:40] people on it and with that September 82 [70:44] I'll turn it over to Chris yes I go show [70:54] the tape team now you use one for both [70:59] why don't we if you want we can talk [71:00] about that after the talk I forgot that [71:02] I had a videotape here is that monitor [71:04] turned on what I've got here is just 30 [71:12] seconds because the tape is really bad [71:14] of one of the very first user tests in [71:16] July of 1980 and the one I have here [71:20] bill and I took turns that this what I'm [71:22] going to be showing you is Bill with a [71:24] subject you can't see the subject most [71:27] of the tape all I can see is a [71:28] out-of-focus back of Bill's head but I [71:30] found half a minute where you can [71:31] actually see the screen of the Lisa so [71:33] it's really quick you better get a good [71:34] look and I think it's set up let's try [71:41] and is there audio on [71:45] I tied you up late have you type playing [71:49] with that [71:51] [Music] [71:52] option so here's that menu at the bottom [72:01] of the window this is really early and [72:04] there's that highlighted selection the [72:05] person made with the mouse and he just [72:08] explains something very simple and she [72:10] didn't understand it so he's starting [72:11] all over now we get to see his hair so I [72:15] think we'll stop okay [72:20] so there's nothing really much to get [72:21] out of that but I did promise to show [72:23] some early Lisi user tests now I've done [72:25] it [72:26] given the time I think we're gonna save [72:28] questions to the end and let Chris take [72:31] off on one his great thanks very much [72:36] [Applause] [72:55] as Larry said the the work that we did [73:00] on the Macintosh human interface tough [73:01] tailed with the end of the the very [73:04] fruitful period of Lisa human interface [73:07] design I can't I cannot believe how much [73:09] work they went through in a short period [73:12] months in the summer of 1982 put [73:16] together what we still see today in the [73:18] human interface of the Macintosh and the [73:21] human interface and Windows about what [73:24] was really interesting to me was that [73:27] when Larry first proposed this talk to [73:29] educate some people new people at Apple [73:31] from some companies we'd acquired on how [73:35] the user interface came to be that that [73:39] a lot of the people he was referring to [73:41] a lot of the people who were Co copied [73:43] on some of the documents a lot of the [73:44] work that they drew on really had [73:47] nothing to do with the graphic user [73:50] interface but had to do with a [73:52] long-standing user interface community [73:55] at Apple that had evolved in a totally [73:58] separate organization for totally [73:59] separate strain and that was the Apple 2 [74:02] in the Apple 3 human interface effort [74:04] that was going on in the personal [74:06] computer systems division which was by [74:08] that time physically remote down in the [74:10] big black triangle shaped building at [74:12] the intersection of 280 and Lawrence [74:14] Expressway called that triangle building [74:16] or by then the Bermuda Triangle because [74:18] it's where the Apple 2 group had been [74:20] moved to die the the thing that you [74:24] really have to understand about that the [74:26] crucial difference between the Lisa [74:30] Macintosh human interface effort and [74:32] what was going on in the Apple 2 group [74:34] was not the difference between the [74:36] graphic human interface and text-based [74:37] human interface though that was an [74:39] important part of it the crucial [74:41] difference was that in the Lisa and the [74:42] Macintosh the engineering groups were [74:45] completely integrated in and the driving [74:48] forces behind the making the computer [74:51] easier to easier to use whereas in the [74:53] Apple 2 the evolution of that had been [74:56] that the the impetus for making the [75:00] computer easier to use [75:01] which programs were selected to make [75:03] that had come from the publications [75:06] Department the training Department and [75:08] the department doing demonstrations for [75:11] in-store retail sales our effort and I [75:15] was a member of that group in our effort [75:16] was primarily to mask the complexity of [75:19] the real underlying software that we [75:21] were already shipping by putting sugared [75:24] user interfaces on top of the real [75:26] interfaces and by putting training [75:29] programs on top of the real computer [75:31] programs and this was remember the Apple [75:34] Computer by 1980 boy we've been around [75:36] for three years we had a long history [75:39] behind us in the personal computer [75:40] industry and we had a lot of accumulated [75:43] legacy software that we couldn't just [75:45] haphazardly change we really had to work [75:47] on it so one of the documents that I [75:51] have here is a document I've got it on a [75:57] video here although I don't think it'll [75:59] be that easier easy to read manuals [76:08] consistent and provide a consistent and [76:11] familiar and productive experience to [76:13] readers of manuals Bruce [76:22] this is the this is that document and [76:26] this is the change history of the Apple [76:27] to human interface guidelines that show [76:30] the first draft Bruce tag Nazim he [76:32] claims to have written in in September [76:39] of 1978 second draft in March of 79 [76:45] third draft in mid 1980 fourth draft [76:48] with Steve Smith in early 1981 and then [76:51] it got into the Macintosh group in 1982 [76:54] I started working on it in late 82 and [76:57] that's when I incorporated many of the [76:59] things from the Lisa efforts and then it [77:02] moved back over into the Apple 2 group [77:04] to cover the Apple 2 graphic user [77:06] interface but it originally started as [77:08] an apple to text human interface [77:10] document and it came originally from [77:14] Bruce's efforts with JD Eisenberg [77:17] writing an application called Apple [77:19] presents Apple which was the main [77:21] in-store or first user demonstration [77:25] program for the Apple 2 Bruce started [77:28] that well I don't know where Bruce [77:29] started out but when we when we [77:31] encountered Bruce Takenaga noesy me in [77:33] his long and checkered career he was the [77:35] proprietor of a computer store in San [77:37] Francisco which is where the apple core [77:38] of San Francisco used to meet and that's [77:40] how he got into the Apple community by [77:42] selling Apple computers so he had [77:44] frontline experience of what happened [77:46] when a person encountered a personal [77:48] computer for the first time when he came [77:51] to Apple he started writing demo [77:52] programs with JD Eisenberg and one of [77:55] the things that they did was they tested [77:57] on real users in real situations where [78:00] you bring somebody who's never seen the [78:02] computer before much like a the Lisa [78:04] group was testing people off the street [78:06] or new employees put them in front of [78:08] this program and gave them no advice [78:10] whatsoever I'm going to read Bruce tends [78:14] to philosophize a lot in this so it's [78:16] not going to be as useful to put up [78:18] slides but I'm going to read to you some [78:20] of his experiences from the [78:30] here we go from user testing our testing [78:36] method is as follows we set up a room [78:38] with five to six computer systems we [78:40] invite groups of five to six users at a [78:42] time to try out the systems often [78:44] without knowing that it is the software [78:46] rather than the system we are testing we [78:49] have two of the designers in the room [78:52] any less and they miss a lot of what's [78:54] going on anymore the users feel as [78:56] though there's always always someone [78:57] breathing down their neck the initial [78:59] ground rules are that no questions will [79:01] be answered as by the time the formal [79:03] testing begins we can supply a draft of [79:05] the manual usually by the second group [79:07] some glaring defects in the interface [79:09] have shown up and we have to give them [79:10] help getting past the stumbling blocks [79:13] 95% of the stumbling block blocks found [79:15] are found by watching the body language [79:17] of the users watch for squinting eyes [79:20] hunched shoulders shaking heads and deep [79:22] heartfelt sighs when a user hits a snag [79:27] he will assume it is his fault he will [79:29] not report it he will hide it make notes [79:31] of each problem and where it occurred [79:33] question the users at the end of the [79:35] session to explore why the problem [79:37] occurred you will often be surprised [79:39] what the user thought the program was [79:41] doing at the time he got lost here in [79:45] follows a true anecdote which [79:47] illustrates how difficult the most [79:48] simple human interface issue can be and [79:50] why thorough testing on real people is [79:52] so important as we tune in it's so easy [79:56] to entertain a crowd just by reading [79:57] something Bruce wrote as we tuned in the [80:00] authors of the software both of whom [80:02] pride themselves on clever interface [80:03] design have anguished for hours over [80:05] difficult passages in their program it [80:07] was to turn out there it was to turn out [80:10] their guesses were quite accurate [80:12] instead difficult passages it was the [80:13] simplest question of all the cause of [80:15] the problem an apple presents the apple [80:18] 2e the training program for teaching [80:20] fundamentals of using the new Apple 2e [80:22] find out if the user is working with a [80:24] color monitor User Profile new owner [80:28] customer and computer store or member of [80:30] a class learning how to use an Apple [80:32] computer test User Profile customers and [80:34] computer store non computers in the [80:36] classroom environment friends and [80:37] relatives first design a color graphic [80:40] is displayed on the screen [80:42] are you using a color TV on the Apple [80:45] anticipated problem those who are using [80:48] a monochrome monitor in a classroom or a [80:50] computer store situation wouldn't know [80:52] whether the monitor was black and white [80:53] or was colored with the color turned off [80:55] first attempt to color graphics [80:57] displayed prompt is the picture of of in [81:00] color 25 percent failure rate reason as [81:06] anticipated but incorrectly overcome [81:08] those seeing black and white thought [81:10] their color might be turned down they [81:11] didn't answer the question wrong they [81:13] turned around and asked one of the [81:14] authors whether the monitor in question [81:16] was color or not a decision was made the [81:19] authors could not be supplied with the [81:20] disk second attempt a smaller graphic [81:24] with large letter words and their own [81:26] vivid colors was substituted green blue [81:28] orange magenta which were the only [81:30] colors the Apple 2 can display prompt [81:33] are the words above in color color [81:34] failure 8 color TV users none [81:37] black-and-white monitor users none green [81:39] screen monitor users 100% this is why it [81:46] pays to have a variety of configurations [81:48] in your test suite third attempt the [81:51] graphic remain in the same failure rate [81:53] prompt are the words above in more than [81:56] one color failure eight color TV users [82:01] none black and white monitor users 16% [82:04] green screen monitor users 50% reason [82:08] the black and white monitor users who [82:09] answered incorrectly admitted they did [82:11] so on purpose [82:16] 50% of the green-screen folk considered [82:19] that they were looking at both black and [82:21] green two colors and answered the [82:23] question accordingly fourth attempt same [82:29] display of graphic and color text prompt [82:31] are the words above in several different [82:32] colors are the words above in several [82:35] different colors failure rate color TV [82:38] users none black and white monitor users [82:40] 20% green screen monitor users 23% by [82:44] this time the authors were prepared to [82:46] supply everyone who bought an apple with [82:47] a free color monitor just so we would [82:49] not have to ask this question it turns [82:53] out about 20% of the people we're not [82:55] really reading the question and we're [82:57] responding to our the words above [82:58] several different colors oh yes of [83:00] course green violet red there's several [83:01] different colors let's move on fifth [83:04] attempt same display of graphic and [83:06] color text prompt do the words above [83:08] appear in several different colors [83:11] failure eight none in it in case it [83:17] appears the authors were simply dull [83:19] fellows be it known that this was a [83:20] fully interactive training program in [83:22] excess of a hundred K of code remember [83:25] it was 128 K machine so and this was the [83:29] only interface issue that required more [83:31] than one correction it clearly [83:32] exemplifies how even the most careful [83:34] designers can totally miss when guessing [83:36] at how users are going to respond so a [83:40] reading from Bruce cognizant the Apple 2 [83:47] project had started out well the Apple 2 [83:50] wasn't really a project it was a [83:51] collection of projects we started out [83:53] with a number of pieces of software that [83:56] had our cherished angle bracket or [83:58] square bracket prompts by the time we [84:01] got around to the Apple 2e in the Apple [84:03] 3 we developed a more or less text [84:06] windowing interface called the file card [84:09] metaphor where we used special [84:11] characters in the character generator [84:12] ROM to do something like the file card [84:15] interface in the Lisa with pick one two [84:19] three four five six menus and you use [84:22] the Escape key to back out through a [84:23] hierarchy and press [84:26] numbers 1 2 3 4 5 6 to advance to the [84:28] next level it was a hierarchical spatial [84:32] text driven keyboard ribbon user [84:34] interface it didn't use the mouse it was [84:36] very successful for what it did we [84:39] ensured consents consistency across many [84:41] applications all the applications that [84:44] Apple implemented as well as third-party [84:47] applications for example if anyone was [84:49] familiar with quark catalyst quark [84:52] they're the people who make pork Express [84:54] now they made the original hard disk [84:58] manager for the Apple 2 in the Apple 3 [85:00] and it used the same interface it had [85:03] consistent use of the keyboard keys the [85:05] open apple always meant the same thing [85:07] in in in different applications and so [85:09] even though it was a text interface it [85:12] had many of the underpinnings of the [85:14] consistency of the graphic user [85:16] interface of the Macintosh and it was [85:18] driven by the exigencies of user [85:22] training which was my mother's group [85:26] because the less time we had to spend on [85:28] user training the cheaper and faster it [85:30] was and if the software were consistent [85:32] the training went more easily it was [85:35] driven by the [85:35] the exigencies of publications which was [85:38] the group I was running and one of the [85:40] problems we faced in publications was [85:42] that having not invented desktop [85:44] publishing yet we had six weeks minimum [85:48] from film to printers which was usually [85:50] the time that was reserved to write or [85:53] finish the software and we often had to [85:57] be finished with the manuals several [85:59] months before the product shipped which [86:02] in the software wasn't finished at that [86:04] point so in order to have manuals that [86:08] accurately reflected what the soccer was [86:10] going to do we had to have the [86:11] programmers tell us what the user [86:13] interface was going to be and then in [86:15] going through that with the programmers [86:17] having them do it the same way that [86:19] other applications did it gave us a more [86:21] full idea for okay it's going to do this [86:23] it's going to do this it's going to do [86:25] that if they followed some standard or [86:27] some guidelines in fact what Bruce did [86:30] before writing the Apple to human [86:32] interface guidelines and what the origin [86:34] of this document was was in fact a [86:39] publications style guide which was how [86:43] we wrote manuals and how we explained [86:45] portions of the human interface and [86:47] explained portions of how you use the [86:50] computer and by using consistent [86:53] explanations that drove consistent [86:56] design and in some of the applications [86:58] so that was some the origins of user [87:01] testing and of of consistency of new [87:05] user interface design in the Apple 2 [87:08] group as I said it moved into the the [87:11] Lisa group started with a separate [87:13] separate angle from which Larry talked [87:16] about coming from the Xerox PARC and the [87:19] to basically fused in the Macintosh [87:21] group where we had the the Lisa just [87:25] benefits of the years of Lisa user [87:27] testing experience in the massive amount [87:29] of design that went into their graphic [87:31] user interface it user interface plus a [87:34] lot of the Democratic grassroots issues [87:39] that had come up through the Apple to [87:41] remember Lisa was only dealing with [87:42] seven applications that was all you've [87:44] got on ELISA the Macintosh was designed [87:47] like the Apple Q is something where you [87:49] had many many applications from [87:51] third-party developers and so we took [87:53] the Lisa user interface grafted it on to [87:56] the guidelines that had come out of the [87:57] Apple two style guide and Apple to [87:59] guidelines efforts and published that [88:01] for third-party software developers to [88:03] use as style guides for their own [88:05] applications I remember very very very [88:08] clearly that one of the massive [88:12] controversies around the development for [88:14] the Macintosh circa 1990 1980 to 1983 [88:18] was developers would come up to us and [88:21] say you know if you make the user [88:23] interface consistent and if you put all [88:26] that software and ROM that makes it you [88:28] know if you make it hard to write to the [88:31] screen directly so that we have to use [88:33] your user interface software to talk to [88:35] the user how are we ever going to make [88:38] our applications unique and stand out [88:40] and be different from each other in the [88:41] marketplace they found a way I'm very [88:43] happy but the the human interface and [88:46] the idea of a consistent human interface [88:48] much less and graphic one was very very [88:51] controversial [88:52] software developers and only by having [88:54] the user testing under our belt to show [88:57] them that hey this is going to add [88:59] significant value to your product by [89:02] making your product easier to learn and [89:04] easier to use especially for people who [89:06] already know another application where [89:08] we able to really convince them that a [89:11] consistent graphic user interface was a [89:12] good thing to have I want to fast [89:15] forward a little skip through the early [89:17] years in the Macintosh and go to a tape [89:20] that I dug up from the Apple library of [89:22] the of some user interface testing that [89:25] that shows how amusing user interface [89:28] testing can be in the kinds of [89:29] situations and far-flung places that [89:31] members of our our glorious human [89:33] interface group went to find the truth [89:36] about what real users would do and this [89:39] is a tape of a couple of our user [89:42] interface designers Laurie ver tell me [89:44] and one other person I may come up with [89:48] her name oh I don't think it was Gita [89:51] traveled to France to parts dealership [89:55] for Renault automobiles who were [89:57] deploying a new automotive parts catalog [90:00] based on HyperCard and HyperCard being a [90:03] free-form [90:05] toolkit for designing new software [90:08] interfaces presented an entire raft of [90:11] new problems in human interface design [90:12] and then they actually went to France to [90:15] test auto mechanics in their own [90:18] environment and I'm going to play that [90:19] tape for you now yeah I'll turn it on [90:43] it might have been cake tamal yeah okay [90:47] Thank You Annette was that Annette oh [90:50] it's joy hi joy [91:14] you know okay this is hard [91:30] the first part of the user test was to [91:33] test users attempting to fill out an [91:34] order form using declarant even though [91:38] [Music] [91:39] these are the order and estimate forms [91:41] that the mechanics currently use and the [91:50] repair Mendel has instructions on how to [91:52] read the car [91:54] Katie's daleney teske procedure and the [91:57] cows to be recovered in an attempt all [92:01] the researchers were interviewed as a [92:02] mechanical educator in French [92:08] [Music] [92:10] if you have trouble with some of the tax [92:12] system to test Katie clips that we help [92:32] us evaluate results later I wouldn't a [92:35] test session Cleveland attended work [92:42] with the paper helped us understand how [92:43] it referred to using information for [92:45] example exact the fourth on top of the [92:47] catalog with their pages while [92:49] researching information research [92:51] techniques helped us to make design [92:53] recommendations for searching that the [92:54] electronic version default on process [92:59] once we finished testing the curt [93:01] affordability seizure we asked intended [93:03] to solve the same problem in the [93:04] documentation electronic system at this [93:09] point in our test easy no strike almost [93:11] 10 minutes simply respect the system [93:13] hate fighting scripts the easy to do is [93:16] click on the [93:18] system interpret horns no you should [93:24] explain the pop-up menu that lies who [93:26] lives there iconic kate says no visa [93:29] quickly I kind of like nothing about how [93:31] these pop-up menus in the initiative [93:32] dialog box which explains to him and [93:34] musical down on the icon question [93:40] McKenna didn't receive a lock box before [93:41] must refuse not only about the message [93:43] to the side of it but that has do with [93:45] them finally analyzing this click on OK [93:48] to get rid of it [93:53] [Music] [93:59] the ocean transportation that puts on [94:01] many Tibetans and the many to the left [94:03] of the screen unfortunately one of the [94:05] menu items worked at this point because [94:07] he's not so about the appropriate [94:09] vehicle identification to actually start [94:12] work session tries to smile economists [94:16] anniversary because once again he brings [94:18] up the ballot box itself and hold down [94:20] on the other side this time you didn't [94:23] see that he must become your [94:24] preprocessors clicking around the [94:26] strategy real assistant commentators [95:12] [Music] [95:31] so that just uh I think is an [95:34] illustration that of two things one that [95:36] bad design knows no bounds and and two [95:39] that that you never get a user interface [95:43] right and that user testing from the [95:47] earliest times when you have a a totally [95:50] unsaturated market people who've never [95:52] seen personal computers before much less [95:54] mice keyboards screens things like that [95:57] to today when we assume that anybody [96:00] who's ever you know been in Western [96:04] society has had to have encountered a [96:06] personal computer at some point and [96:08] knows the basics of pointing clicking [96:11] dragging pressing what a file is what a [96:13] document is how to print why to print [96:16] what a floppy disk is I mean II I I [96:19] still find it surprising that we don't [96:22] talk about sectors and cylinders anymore [96:25] on discs which was one of the main [96:26] things we had to explain the 1978-1979 [96:30] because there was some reason you needed [96:33] to know that I forget why but yes but [96:38] that some things never go away and that [96:42] having to test new designs and test your [96:46] assumptions of what your audience knows [96:48] is pretty much eternal so at that point [96:51] having run pretty late I think Larry and [96:53] I will take a few questions and I think [96:55] that we are going to take free liberty [96:57] to defer to several people in the [97:00] audience who are noted in Apple hue [97:02] interface history I see joy Mountford [97:05] and Annette Wagner back there who've [97:07] been instrumental in human interface [97:09] groups in the past and if there are any [97:11] others and if you have strong opinions [97:13] disagree with us have more information [97:16] than we're able to supply supply please [97:19] chime in and Larry do you want to pass [97:22] the mic around so people can ask [97:25] questions in the mic okay oh hello that [97:32] works well your clips nope here Larry I [97:37] wonder if you could comment on the story [97:38] that's often told that Apple ripped off [97:41] the Xerox PARC work I know Jeff Raskin [97:44] has refuted or denied that but it's such [97:46] a popular story could you go on record [97:48] as giving your perspective yeah well I [97:51] actually showed at the beginning of the [97:53] talk picture of the small talk interface [97:55] to see that in fact the early user [97:57] interface designs we're very close to [97:59] that the look of the windows in [98:01] particular and the icons in the desktop [98:02] but it evolved over time and you know [98:07] that menu move to the top of the screen [98:08] and so on if you actually look at either [98:10] a Xerox star or a small talk screen and [98:12] then look at Alisa or Mac I think you'll [98:14] see a lot of differences what was taken [98:17] was the concepts the concepts that there [98:20] ought to be overlapping windows that [98:22] there ought to be direct manipulation [98:24] and so on but the rock star for example [98:27] didn't have the concept of dragging [98:29] things and it had two button mouse you'd [98:31] click one place and then click another [98:32] place it didn't have a menu bar in fact [98:35] it was only one menu that was added late [98:36] in the development and it you know but [98:41] some things were similar they were there [98:43] was a horizontal and a vertical [98:44] scrollbar so certainly the concepts came [98:47] largely from there but as you see some [98:49] of the concepts came from other things [98:51] from HP systems UCSD from Apple too from [98:54] some work that was done at IBM that I [98:56] didn't talk about today that was [98:57] published around the late 70s so there [99:02] were a lot of origins for it I think the [99:04] Xerox work had a lot to do with it but [99:06] usually when the if you sat down to try [99:11] to use a star or small talk you would [99:13] find that there were I think more [99:14] differences than some [99:15] clarity's in what you would actually see [99:17] and do [99:31] question is the user testing we showed [99:33] for the Lisa had to do with novice users [99:35] and we the question was whether we [99:36] tested for how people our culture ate it [99:38] over time [99:40] well the system remember from the very [99:42] first page of the user interface [99:43] guidelines Bill Road was targeted at [99:45] novice users that was the target which [99:47] was about everybody in those days and we [99:50] did have some theories about how people [99:54] would you know use command keys as they [99:56] got more expert and so on but we didn't [99:59] run user tests one reason we didn't was [100:01] because it would have taken a long time [100:02] to get people up to speed so what we did [100:06] was we just used kind of ourselves I [100:08] don't mean the engineers but people who [100:09] worked in the Lisa division who had [100:10] leases like people in product marketing [100:12] they became expert after a while and [100:15] they would start telling us that things [100:17] were too slow or too hard or whatever [100:19] and that's how we got that kind of [100:20] feedback just from people inside and I [100:24] think it's more relevant today to check [100:27] for those sort of things than it was [100:28] back then when everybody was you know [100:30] 95% of the people were novices one of [100:32] the constant tensions and this was true [100:34] on the Macintosh as well especially [100:36] because we had the benefit of the Lisa [100:38] work before us and some of the the Lisa [100:40] machines was the constant tension is [100:42] that the people who are designing are [100:44] almost always by definition expert users [100:47] and will often overlook that what they [100:49] think is obvious is not at all obvious [100:51] to a novice user and so there there's [100:55] constantly the the dilemma which you've [100:58] seen historically in Mac system software [101:00] that the expert users want to put in the [101:03] features they want to use but the people [101:05] who want to keep this system pure for [101:07] the novices want to resist those and and [101:09] if you're lucky you get a system that is [101:11] easy to approach for the novice and [101:13] gradually unfolds itself for the expert [101:15] and if you're unlucky you get a lukewarm [101:18] mediocrity between the two where it's a [101:20] little too complex for the for the [101:24] beginning user to understand but still [101:26] not nearly powerful enough for the [101:27] expert user and and some of our designs [101:29] and some of the industry designs have [101:31] been one gradually unfolding and others [101:33] have been just plain mediocre [101:38] have a survey ssin and then a question [101:40] Larry and I have known each other for [101:42] years and it's interesting to me that [101:48] for one thing you folks weren't away [101:51] aware of the similarity to the word [101:52] start commands had already existed and [101:54] the word star editor for you know copy [101:57] cut paste kind of things but also I [102:00] think the the word assumption is is key [102:04] and what you just said is where [102:05] incorrect people assume that as they're [102:07] developing something that they know the [102:09] best way to develop it rather than go [102:10] outside and get other information you [102:13] know one of the things that we've I've [102:15] seen as a software person in the past [102:18] having to do work for people for money [102:20] so that they expect us to have something [102:22] done by a certain time I mean not for a [102:24] company that I'm working for but for as [102:26] a consultant is that they really do [102:28] expect the thing to work the way their [102:30] people want it to work and there you [102:34] have to really do a design review and [102:36] and and and all sorts of things with [102:38] those people first before you you get to [102:40] the right set up before you can actually [102:43] start actually writing code and and when [102:47] I was a survey ssin about the Lisa that [102:49] was very interesting when I first saw it [102:50] I first thought I think at the West [102:53] Coast computer Fair in San Francisco [102:55] Steve Jobs was there I don't know who [102:58] else from Apple what was their marketing [103:01] staff but I'd gone there with a friend [103:03] from Xerox and we looked at the star [103:05] system and we'd actually looked at the [103:08] fortune which was a system which was the [103:10] first 16-bit personal computer that was [103:13] available and they had a very fancy you [103:16] know this way and so forth and everybody [103:18] was looking at it and was beautifully [103:19] designed and everything everything else [103:21] that ran CPM or there and UNIX a version [103:23] of sort of here UNIX there was one test [103:27] in the software work that we had done as [103:30] for for real customers you know who [103:32] really wanted reliable software that [103:35] we'd accidentally come upon which was [103:36] called the keyboard mush you take your [103:38] hand and you run it across the keyboard [103:40] a few times and you see if your software [103:42] is still running [103:45] and I told this friend of mine from [103:47] Xerox about it so he wants over the [103:49] fortune system on the big podium up [103:50] there with a pretty girl sinning by it [103:52] and he tries it and the thing dies so we [103:57] kind of sneak away and we go over to the [103:59] Apple booth and there's Steve and the [104:02] other folks there and we decided we [104:04] weren't going to try it on and Lisa [104:06] so in reaching for a brochure though I [104:08] accidentally touched the upper right [104:10] portion of Lisa keyboard [104:12] they were networked at the time at the [104:14] show for demonstration and networking it [104:16] brought down that Lisa the guy helped me [104:20] try to reboot it it wouldn't come back [104:22] it also brought down all the other [104:24] machines at the booth on the network and [104:28] the only thing he finally could do was [104:30] to power cycle the whole system so I [104:33] just wanted to bring that out well now I [104:35] just want to bring out the looking for [104:43] the unforeseen okay right good point [104:45] just in the interest of time we're gonna [104:47] try it becoming shorter though yeah yes [104:59] the question is whether there's some [105:01] correspondence between the object [105:03] oriented programming metaphor and the [105:05] user interface metaphor and yes there is [105:07] in fact in the small talk user interface [105:10] which was one of the first of this type [105:13] and the small talk language which was [105:15] one of the first of the object-oriented [105:17] type the connection was very explicit [105:19] that you had objects that had generic [105:23] commands that could operate upon them [105:25] and that was something that existed both [105:27] in the language and in the user [105:29] interface so which was done even more so [105:32] in the star but we definitely the [105:35] programmers who were doing these user [105:37] interface definitely had that in mind as [105:38] a correspondence here John [105:46] oh yeah good user interface don't have [105:57] to I'm just saying it the user [105:58] interfaces that we were developing we [106:02] had that in mind but that doesn't mean [106:04] that a good user interface has to be [106:06] object-oriented no what did I use to [106:17] write the memos probably Apple writer [106:22] you let's see early on we were using [106:25] maybe Apple writer on the Apple - or [106:28] maybe UCSD Pascal probably using UCSD [106:31] Pascal program editor on an app used on [106:34] an apple - I didn't have I never had an [106:36] Apple three you're right you were [106:37] probably using either Apple writer on an [106:40] apple - or sophisticated people use the [106:42] UCSD Pascal editor and output it using [106:45] an inner off like program called Apple [106:47] script right we have put it to cue [106:49] printers when we could afford the time [106:53] and noise and to dot matrix printers [106:58] 80-column dot matrix printers and you [107:00] had a mix of those two before we got the [107:02] apple - an apple threes up jef raskin [107:05] had imbued us with polymorphic poly [107:08] eighty-eights using their editing system [107:11] and we did all of our editing on poly [107:13] eighty-eights Ciroc terminals back there [107:39] well I don't know about successfully [107:41] competing that part but yeah definitely [107:42] it was it was a extremely close to the [107:47] Mac user interface the gem user [107:49] interface by digital research and Apple [107:53] sent some kind of complaint letter and [107:55] there were some discussions between the [107:57] companies but then Microsoft and [108:01] hewlett-packard started doing Windows a [108:03] new wave and that was a much bigger [108:05] issue so Apple kind of stopped bothering [108:08] Digital research and instead got [108:10] involved in a big lawsuit with HP and [108:12] Microsoft Digital research at the time [108:14] was kind of a struggling company and [108:16] Apple didn't want to you know basically [108:19] pick on our little company so today we [108:21] picked any two companies that were too [108:23] big instead of picking on one that was [108:25] too little Bernard [108:41] yeah the the like the Xerox star and the [108:46] Lisa were the idea was that in both [108:50] groups was that people would only use a [108:52] few applications it's kind of true most [108:54] people only do use a few applications [108:56] but the the conclusion that therefore it [108:59] wasn't necessary to have third-party [109:00] applications was kind of ridiculous [109:02] actually in the Lisa it turned out we [109:04] did have a place for third-party [109:05] applications we thought for certain [109:08] vertical markets there would be [109:10] third-party applications and we'd have [109:11] accounting software and so on but even [109:14] then we were thinking very small scale [109:15] like there would be one page layout [109:17] application and there would be one [109:18] accounting application and and you know [109:21] a couple of vertical market things if we [109:23] ever had 20 applications or 30 that [109:25] would have been a very large number very [109:28] different from the Mac which were in the [109:29] very beginning was gonna try to do the [109:31] same thing as the Apple 2 and what the [109:33] IBM PC did which was go for lots of [109:35] third-party applications I think one [109:37] reason the Lisa was designed as it was [109:39] was to try to justify why there was this [109:41] project going on the company we already [109:43] had an apple - in an Apple 3 why do we [109:45] need this other thing and well the up [109:47] between the Apple 3 you know gonna have [109:49] all this third-party software but the [109:50] Lisa won't it was kind of a [109:52] differentiator it was a kind of a closed [109:54] system that was supposed to be [109:54] supposedly a plus turned out to be a [109:56] major - of course yeah well it's that [109:59] was not necessarily unique to Apple at [110:01] that point remember one of the leading [110:03] software companies of 1979 1980 was [110:07] personal software which became visit [110:08] Corp and their flagship product line at [110:11] that point was a failed product line [110:13] called busy on which was an attempt to [110:16] expand visicalc [110:17] into a comprehensive suite of [110:19] interacting applications a word [110:21] processor a database word star at that [110:25] point was also trying to expand and [110:27] everybody was trying to do what [110:29] Microsoft has successfully done with [110:31] Microsoft Office which is turn the thin [110:35] end of the wedge of getting one [110:36] successful application which Microsoft [110:38] had with Excel on the Macintosh and [110:40] turned that into a major suite that [110:43] delivers all of the application [110:45] computing needs of a given set of users [110:47] Microsoft has done it successfully and [110:50] one could argue that the default [110:52] computing platform right now is Windows [110:54] with Microsoft Office and not very much [110:57] else so I don't think that that [110:59] necessarily was a bad idea it may have [111:02] been many years ahead of its time but [111:04] Microsoft was selling the Microsoft [111:06] Office for Macintosh in 1985 I still [111:10] have box on my shelf where they supplied [111:11] the the word processor the spreadsheet [111:13] the filing program in the terminal [111:15] program in one box for one price for the [111:17] Macintosh so office suites are certainly [111:20] they certainly were not a dumb idea of [111:23] 1981 that's something that has really [111:26] monopolized the industry I think that [111:28] part's true thee sis on the ELISA what [111:32] was really one of the first Suites [111:35] and in addition to the ones that [111:37] personal software did and it was also [111:41] one of them it was the first probably [111:42] multitasking personal computer and but [111:46] basically ahead of its time and missing [111:49] the main point which was you needed a [111:50] low-cost platform that had reasonable [111:52] performance and that had lots of [111:54] third-party applications back there in [111:58] the aisle [112:24] the point is that we were testing in the [112:27] early years for totally naive users and [112:30] that over the years culturation has [112:33] taken place as I talked about the more [112:35] people approach computers with with less [112:39] fear and with more pre knowledge of what [112:40] they're doing is that your general point [112:48] yeah and the process came from a decade [112:50] of more of experimentation well I'll [112:52] tell you about an interesting user [112:53] experiment that I did in 1982 I took a [113:01] early prototype of the Macintosh and it [113:04] was then in a Plexiglas case just cut [113:08] panels of Plexiglas filed in glued [113:11] square we didn't have any of the [113:13] injection molded cases yet with the guts [113:15] inside it was running an early version [113:19] of mac paint which it which was called [113:22] max sketch at that point and it was [113:24] turnkey application you put the disc in [113:27] and booted it up and it went straight to [113:28] that application because there was no [113:30] finder at that point which was actually [113:32] the paradigm of most other machines [113:34] which was turned kyboot we put the [113:35] floppy in and you booted it and you got [113:36] into that one application and I took it [113:40] over to my girlfriend's parents house [113:42] for Thanksgiving where she had a [113:46] three-year-old nephew three year old [113:51] nephew and this was in 1981 video games [113:54] were even that hadn't even gotten down [113:57] to three year olds by 1981 that boy [114:00] spent four hours drawing with Mac's [114:03] sketch he learned it in no more than [114:06] fifteen or twenty seconds just you know [114:09] he pressed the button and draw things on [114:11] the screen and he was drawing things [114:13] with with an early version of Mac Paint [114:16] with virtually no tutoring and almost no [114:18] language at all with no color with no [114:22] acculturation at all so there were some [114:24] things that we got basically right about [114:26] human human factors if you can appeal to [114:29] a three-year-old and get a [114:30] three-year-old with no language skills [114:31] to [114:32] and how to do something productive in no [114:34] time at all that's not a cultural [114:36] statement you've done something right in [114:38] design and that's when I knew we were [114:40] really onto something I don't think it's [114:45] had no effect I think what acculturation [114:47] has done is it's it's gut it's made us [114:51] soft we don't address complexity anymore [114:53] the way we we sweated over complexity in [114:56] 1981 we were deathly afraid of [114:58] complexity we take it for granted now [115:00] you'll look at fries it's scary the kind [115:02] of things we get away with I think it [115:05] has made something's hard to change when [115:07] Microsoft announced they were gonna go [115:08] to single clicks instead of double [115:09] clicks on the desktop some journalists [115:11] said see they're leaving Apple in the [115:13] dust you know you're still in two clicks [115:15] and you're moving to one I said our [115:16] users on this doing this for 12 years [115:19] there's no way they're gonna change [115:20] we're never going to change it and maybe [115:22] the Microsoft users who've only been [115:23] doing this for a couple years can change [115:25] but as it turned out even a couple of [115:27] years using Windows 95 was too much and [115:30] and Microsoft had to back off on making [115:33] users to use single clicks in Windows 97 [115:36] because even two years was too much [115:37] people don't change we are out of time [115:39] now but there are some people here who [115:43] were involved in Lisa user-interface [115:44] work all the way from the Apple to the [115:46] Lisa and the Mac now maybe they could [115:48] stand up and a net Wagner especially who [115:52] was worked on the original Lisa look and [115:55] graphics and fonts and everything and [115:57] joy and Bob I mean once you all stand up [116:00] all you Apple user interface people I [116:03] think some of the users out here might [116:04] appreciate you pop glass [116:07] [Applause] [116:16] we're in a warehouse just down the [116:18] street to the right and Larry and Chris [116:20] would you walk through there to answer [116:21] questions right [116:23] [Applause] [116:26] [Music]