mo@maximo.UUCP (05/22/87)
From: <seismo!maximo!mo@mimsy.umd.edu> Date: Fri, 22 May 87 11:17:24 -0400 Mr. Fulton claims that creating good user interfaces requires great skill and artistic talent - I whole-heartedly agree. Ideed, I said so in my first comment. Further, he accurately pointed out that a "large, complex library of widgets" won't make good interfaces. Again, I completely agree. What he and several other respondents seem to be missing, however, is the vital importance that different applications running under the "Industry Standard" window system [X111,NeWS,etc, you pick] (I know, I know, but it will happen anyway) NOT, repeat NOT have different user interfaces!!!!! The last thing in the world I want is to have three different applications windows open on my screen, each with a different user-interface paradigm!!! THIS WILL BE FATAL! I believe one of the points to having windows in the first place is so one program doesn't take over the world, but yet, all these interfaces, and I infer their designers, believe their program will be the only one being used at any one time! This will lead to a replay of the current situation - every program has a different command language, only now it will be visual, not linear. A bit of apocrypha: A survey was done in the moderatly-recent past of IBM PC and Macintosh users regarding how much software they have, how much software they really USE, and how long they spent learning to use it. Note they interviewed REAL PEOPLE, not programmers. PC users generally have 8-10 or so pieces of software. They generally feel facile with only 2 or 3 pieces. They spend 20-40 HOURS learning to use EACH one. (This is probably why they don't learn more of them!) Mac users generally have 15-20 or so pieces of software. They generally feel facile with 10-13 of them. They spent 30 minutes to 1 hour learning the first one, but only about 20 MINUTES learning the additional ones. This ease of learning was universally ascribed to the commonality of the user interface BETWEEN programs. If software running on mid- and high-end workstations is ever going to be a broad commercial sucess, instead of an academic nicety or a select vertical market (these two are probably redundant!), this issue must be considered. Quite frankly, given the choice between a system with lots of individually good, but collectively randomly-behaved software, and a system which runs a lot of essentially-equally-powerful software with a very consistant interface, the randoms have little hope of commercial sucess. This is vitally important to those of us interested in writing and SELLING high-quality software so we can make a good living doing so. Oh yes, as Henry Spencer pointed out, and I cannot agree with more, There is no tool like the Right One for the job. But that doesn't make a hammer good for putting in screws because it does it so fast! Yours for the *right* tool, and clearer windows, -Mike O'Dell
jim@ci-dandelion.UUCP (Jim Fulton) (05/26/87)
I think that everyone is in violent agreement that having a good common user
interface is VERY important. The point that I was trying to make in previous
postings was that this is NOT something that should be built into the window
system specification. I strongly suspect (and hope) that commercial
applications written for the "Industry Standard Window Systems" will use the
common toolboxes and "play by the rules" so that they will work well with a
*variety* of User Interface Managers (which in themselves will become avenues
for commercial development). However, there is and always will be room for
growth and the development of new ideas.
The barbs being flung at X and NeWS should instead be aimed at the toolkits
that will reside on top of them. It is a much harder problem than first
appears given the wide range of display, CPU, and network hardware that will
used.
Jim Fulton
Cognition Inc.
900 Tech Park Drive
uucp: ...!{mit-eddie,talcott,necntc}!ci-dandelion!jim Billerica, MA 01821
arpa: jim@athena.mit.edu, fulton@eddie.mit.edu (617) 667-4800gnu@hoptoad.UUCP.UUCP (05/27/87)
Date: Tue, 26 May 87 03:06:47 PDT From: hoptoad.UUCP!gnu@cgl.ucsf.edu (John Gilmore) In article <8705221517.AA05422@maximo.uucp>, Mike O'Dell writes: > Quite frankly, given the choice between a system with lots of > individually good, but collectively randomly-behaved software, > and a system which runs a lot of essentially-equally-powerful > software with a very consistant interface, the randoms have > little hope of commercial sucess. I absolutely agree. I have had an old Unipress Emacs with mouse support for years. I cannot use its mouse support because it is different from the mouse support provided by the basic window system (SunView) and I can't remember what it does. So I end up copying text around into files and inserting the file in Emacs, because it doesn't have a Stuff command in the mouse menu like all the other windows. NeWSfolks at Sun, please remember to define a good user interface for NeWS. When the interpreter works and is complete and fast, you are only half done. If you don't provide a user interface at least as good as SunView, everybody will write their own, and we users will be mightily confuzzed. (SunView tends to suffer from too many frobs and too little effort put into simplicity and ease of use -- try the so-called editor sometime -- but its advantage is its ubiquity. Applications CAN'T override it so they learn to live with it.) This job has clearly not been done in the NeWS prerelease tape, though I don't fault the prerelease for the lack of it -- something had to go out the door to get people familiar with programming in NeWS. A very good skeleton is there; it just needs flesh.
mlandau@diamond.bbn.com.UUCP (05/28/87)
In comp.windows.news (<8705261006.AA14448@hoptoad.uucp>), gnu@hoptoad.UUCP (John Gilmore) writes: > >NeWSfolks at Sun, please remember to define a good user interface for >NeWS. When the interpreter works and is complete and fast, you are >only half done. If you don't provide a user interface at least as good >as SunView, everybody will write their own, and we users will be >mightily confuzzed. Last time I heard anything official from Sun, the party line was that NeWS would ship with a SunView toolkit, so that applications programmed to SunView specs would be source-code compatible with NeWS. Has there been some change in plan recently? -- Matt Landau "Waiting for a flash of enlightenment mlandau@diamond.bbn.com in all this blood and thunder..."
RICHER@SUMEX-AIM.STANFORD.EDU (Mark Richer) (05/30/87)
Date: Fri, 29 May 87 11:06:32 PDT From: Mark Richer <RICHER@SUMEX-AIM.STANFORD.EDU> Subject: Re: Jim Fulton's comments To: hoptoad.UUCP!gnu@CGL.UCSF.EDU Cc: NeWS-makers@brillig.umd.edu Well I have been sitting back following this discussion, hesitating to get involved, but here's my two cents anyhow: I think there is a real inherent tension between the need for standardization (uniformity, conformity) and innovation. Similarly there is a tension between the desire/benefits of a standard user interface and the reality there is not a general consensus (or at least a sufficient one) on such an interface. Personally I think the Mac interface is a great success and despite whatever shortcomings it possesses, it is clear that Apple put a great deal of effort (intelligent and skilled effort) into its development. Still it's hard to imagine choosing an interface today that we could say should be a standard user interface for all windowing applications on all machines. Obviously we are barely ready to settle on a standard programming interface (i.e., toolkit) either. However, the point is well-taken that a bunch of different applications (regardless if they all run under NeWS, X, MS-windows, or even Mac windows) with different user interfaces is a real lose. ONe perspective that helps is to separate out the needs of the research community from the commercial end-user. The research community wants to be able to experiment with different user-interface designs within the same windowing paradigm (NeWS and/or X), but the commercial end-user (or student trying to get some work done) couldn't care less about such flexibility. A uniform interface leads to far more productive use of application software than a bunch of applications all with their own unique, clever user-interface designs which often conflict even at the keystroke or mouse selection level. So my conclusion is that the modularity/layering that NeWS/X provide is nice for allowing experimentation and future innovation. It give vendors and individual programers more flexibility (not necessarily less headaches). BUt I think that at minimum each vendor should establish clear and thorough user interface guidelines (perhaps applications should be certified) so at least applications on the same piece of hardware will work in a similar way. Of course, many places (especially universities) desire a heterogeneous hardware environment to meet a diverse set of needs and it would be nice to know you can move from machine to machine and not have to deal with a different and potentially conflicting user-interface. I have already experienced the pain of swithching between several windowing environments where conventions for using the mouse conflicted. After a while you don't care which way is better anymore, only that there would be just one way so you could rely on one automatic/unconscious habit to work transparently. So this seems to call for standards across the board for commercial applications (buck the standard at your own risk --- some Macintosh applications have been severely criticized for not following the user-interface guidelines and the word gets around to consumers). Of course all this at the risk that future innovation may be delayed (I'm not convinced this is a big factor anyhow regarding commercial products), nobody will be able to sue anyone else for "look and feel," and normal people might actually enjoy using computers rather than moaning and groaning about them. Mark -------
jww@sdcsvax.UCSD.EDU (Joel West) (05/30/87)
In article <12306285733.82.RICHER@SUMEX-AIM.STANFORD.EDU>, RICHER@SUMEX-AIM.STANFORD.EDU (Mark Richer) writes: > Personally I think the Mac interface is a great success and despite whatever > shortcomings it possesses, it is clear that Apple put a great deal of effort > (intelligent and skilled effort) into its development. So much so that they view it as a competitive advantage and strongly proprietary. So much so that IBM (supposedly, it's all vaporware to me) decided to mimic the idea of a std interface with their OS/2 Presentation Manager, which looks a lot like MS-Windows. > The research community wants > to be able to experiment with different user-interface designs ... > So my conclusion is that the modularity/layering that NeWS/X provide is > nice for allowing experimentation and future innovation. It give vendors Here, here > BUt I think that at minimum each vendor should establish clear and thorough > user interface guidelines (perhaps applications should be certified) so > at least applications on the same piece of hardware will work in a similar > way. For each system running X, NeWS or whatever, there needs to be a standard Window Manager supplied, with standard guidelines on the 'right' way to do it. As Apple demonstrated, guidelines can be very effectively communicated using well-designed demos. Any window system that doesn't have a complete window manager functionality, guidelines, etc. is as incomplete as a programming language that decides basic I/O is "implementation dependent". Yuk! -- Joel West {ucbvax,ihnp4}!sdcsvax!jww (ihnp4!gould9!joel if I ever fix news) jww@sdcsvax.ucsd.edu if you must
gcj@maths.qmc.ac.UK (Gordon Joly) (06/02/87)
From: Gordon Joly <gcj%maths.qmc.ac.uk@Cs.Ucl.AC.UK> Date: Tue, 2 Jun 87 11:00:13 BST > [...] After a while you don't care which way is better >anymore, only that there would be just one way so you could rely on >one automatic/unconscious habit to work transparently. [...] Mark Richer. Mark, It would hard to make transparent change from the SUN mouse to the MAC mouse, since they have a different number of buttons:-). Gordon Joly, gcj@maths.qmc.ac.uk gjoly@cs.ucl.ac.uk