Microsoft Usability (Jef Raskin) : LUSENET : Joel on Software : One Thread

Jef Raskin asked for specific examples of bad UI practice in Microsoft products - well all he has to do is read Alan Cooper's book 'About Face' (I have no interest in advertising this book).

Amongst Alan's list of serious complaints about Microsoft products: 1) unnecessarily bothering the user with the file system 2) Over-use of dialogue boxes (esp. annoying things like asking if you want to save changes if all you've done is print) 3) No orchestration in the interface, i.e. related functions accessed by very different aspects of the interface 4) Over-use of toolbar buttons, rather than using them in an appropriate way to access the most commonly used functions. 5) The use of wizards that help you out once, but give you no insight into how to help yourself again 6) No attempt to learn about the user and their way of working 7) Catering for all that is possible, rather than what is probable (e.g. being asked every time you close a document if you want to save changes - 90% of the time you do; with a decent undo all facility this question would never be necessary). 8) Blindly following the implementation model

The good thing about Cooper's book is that he sticks his neck out and makes concrete suggestions about how these areas could be improved.

As an aside, in the UK the The Open University has a HCI course; a lot of students who do this course choose a re-design of a microsoft feature (e.g. styles in Word) as a project because they know there will be lots of room for improvement. I know of at least 10 good re-design projects relating to Microsoft Office products - most of these improvements didn't exactly take a stroke of genius to come up with ;0)


-- Anonymous, May 30, 2001


yes, I read the Humane Interface

I read Raskin's book. It's an interesting book, but not an important one. The UI that Raskin describes there is NOT more usable than the UIs we have now. It's so radically different that it would be extremely hard for existing users to understand and use correctly.

It might be nice to invent something radically different if people didn't already know how to use computers. Unfortunately, that's not the world we live in. In this world, Raskin's hopelessly radical UI would just be impossible for people to use and would fail (deservedly) in the marketplace. If Jef Raskin proved that putting the accelerator on the left and the brake on the right was "better" usability wise, it would still be stupid to make such a change. (This is something he understands, and it's even one of his examples, so it's odd that he still goes on to suggest blatently arbitrary changes like eliminating typing-replaces-selection).

It's also hand-wavy of Raskin to describe an arbitrarily invented UI without ever building it and testing it. (No, he is NOT describing the CanonCat UI when he says users should type their own menus.) Some of the ideas he likes so much are so obviously bone-headed it's pretty obvious to me he's never done even a tiny bit of testing. Example: he thinks that when new email arrives, it should be inserted at the insertion point. That's right. If you don't want your new email in the current document, well, you can cut and paste it somewhere else. He also seems to think that all that anybody does with computers is type text... he thinks the way to save a file is to type the word "Save" right in your document and press a command key. Nice. What if I'm listening to an MP3? What if I want to save a JPG that I'm editing in Photoshop? Do I have to switch to the text tool? Come on. This is supposed to be better than Ctrl+S?

Raskin's is so obviously not a good UI it's a miracle he has the guts to write it up. Like, what if I don't WANT to be interrupted by email going at the insertion point? What if I'm actually typing? Jeez.

-- Anonymous, May 30, 2001

Joel wrote:

"Designing UIs for a program as complex as Word or Excel is extremely hard and the very best designs are still often difficult to use, and that applies to any program with a reasonable level of complexity."

Word and Excel represent the very best UI designs? Thou aren't serious, art thou? ;^)

The simple truth is that Word and Excel are way too friggin' complex. I'd go back to Word for DOS V4 if I had a machine that would run it.

-- Anonymous, May 30, 2001

Jef Raskin did design an efficient word processor, part of the Canon Cat personal computer. In my opinion it failed in the market because it was just too different from the way other computers worked.

-- Anonymous, May 30, 2001

Joel said (on Raskin and other MS interface detractors)...

"The dishonest part is that they never seem to offer any specific improvements."

This remark is incorrect and unfair - especially wrt Raskin. His book "The Humane Interface" describes a very real, deep, compelling alternative to modern GUIs. Yes, it describes something very different to MS' interfaces, but a fundamental shift in models and metaphors is probably required to manage (reduce) the overwhelming complexity of current apps & systems. Surface-level 're-design' rarely does more than shuffle the complexity around a bit. Raskin's book is an important work. Surely you read it before commenting on his 'offerings'?

I think you (surpringly) picked a poor example this time.


-- Anonymous, May 30, 2001

improvement within the current paradigm

Joel criticised Raskin's work as being too radical a change for users to cope with; this is why I value Cooper's work: he shows how you can radically improve the interface whilst sticking with common conventions (toolbars, buttons, menus, dialogs, etc).

I believe that 'design is in the details' , rather than requiring a grand paradigm shift. It's about orchestrating the elements of the interface and bearing in mind how users actually use seperate functions to perform meaningful tasks / goals. It is alos about making the computer a little less dumb, and making it work harder to make the users' life easier.

I think this is commonly where we go wrong: looking for new metaphors and concepts, rather than tidying up the clumsy work that has gone before. The Office suite would be a damn fine set of productivity tools if Microsoft just took pause in their next realease to improve the interface as well as adding new tools and a new look.

The complexity of the office suite also illustrates the difficulty of designing for a very broad user base - something that Cooper again addresses in 'The Inmates Are Running The Asylum'.


-- Anonymous, May 31, 2001

I'll certainly go with the overcomplexity complain - my wife refuses to use anything newer than Word 5.1 for the Mac, which in her opinion is the pinnacle of word processors, and she's got thesis documents done in it to back her up. Although she uses LaTeX for some of her academic work, since she finds trying to do IPA in Word more trouble than it's worth.

Personally, I find the amount of layout stuff in Word a royal pain in the arse. I hate Word wanting me to fiddle with layout when I don't want to, and in a past life I had to deal with people who had been seduce into believing that Word's layout tools made it a good pulishing system. Since at that stage there was no way to do colour seperations from Word (and still isn't in the package), they ground to a shuddering halt if they wanted a large print run.

-- Anonymous, May 30, 2001

Joel said:

"yes, I read the Humane Interface"

Thanks for replying Joel. I'm not being antagonistic here - I eagerly read everything you write and this is the first substantial disagreement I've felt.

Yes there are certainly some oddities in Raskin's book (the email insertion thing is ludicrous, and smacks of an imminent publishing deadline), and yes there is a disproportionate emphasis on "typing"- style activities (but that covers a LOT of users!). However the fundamental emphasis given to locus of attention and habitualizable 'monotony' *does* reportedly give a quantum benefit in useability. And and yes, the gain is apparently worth the pain of unlearning/relearning some habits - at least according to the very favorable testimony from users and reviewers of SwyftCard and the Canon Cat. This is not the same as swapping pedals in a car - the use of those pdeals is _already_ uniform, habitualized and internalized, which is just what Raskin wants, so he would never change it. Raskin recognizes that *all* user guestures are arbitrary to some degree; he just wants to reduce the complexity to help the users cope. (Most of the "Ctrl+S" combinations in the world go unused, or have unpredictable effects from app. to app.) I for one would *love* for all of the 6+ editors that I use daily to behave in exactly the same utterly predictable way. The "leap" navigation system is a wonderful development of incremental searching. The requirement for dedicated physical control buttons is in direct agreement with Donald Norman's "affordances". There is much of value here. Even Bruce Tognazzini calls the Canon Cat a "really good abstract interface".

My point was that Raskin *has* designed and implemented a real alternative. It may yet be far from perfect, but he (more than most) has the right to aim criticism at current UIs.

BTW: On a *totally* unrelated note... You set great store by Tom DeMarco (Peopleware) - as do I. A while back I read your doubts about pair programming in XP (some of which I share). Have you heard that Tom has now embraced XP as (reported quote) "Not just the best game in town, the ONLY game in town" (his emphasis). The apparent contradiction between Peopleware's quiet and XP's pair programming is explained as some new group dynamic - a new phenonmenon (I can't remember the exact quote). I'm sure you can find the article/link on the net quite easily.


-- Anonymous, May 30, 2001

OK, I'll bite.

the single biggest usability flaw with MS Office is the eternally morphing file formats that produce the lockstep upgrades that make MS so much money.

You can bet using a backward and forward compatible file format gets pushed way down the list at every feature priority setting meeting, but it costs users millions in lost productivity as they try to work out why they can't read each others documents.

At 5:18 pm -0400 30/5/01, Joel Spolsky wrote: OK, so, how do you add features to the product and make old versions of the app work on documents using the new features? for example -- if Word N adds tables for the first time, how do you represent this in Word (N-1)'s file format?

Try looking at an HTML table in a pre-tables browser - what a shock - the
etc tags are ignored, so the table looks ragged, but the text is there.

First you define an extensible file format from the start, like IFF or *ML. This isn't hard, they were old in the '80s. Worst case the user of the old versions loses some new feature; if you're smart you allow alternative representations for fallback. I never design a file format for anything, even an Applications private prefs file, without using a structure like this.

You then architect the software so that the features are added by Components, and make the components downloadable and extensible.

QuickTime is the example of this par excellence - I have run apps off a CD-ROM from 1991 that are able to open streaming movies using brand new codecs, because the API & format were designed with extensibility and flexibility in mind, and backward and forward compatibility have been design goals throughout.

-- Anonymous, May 31, 2001

Kevin Marks wrote:

--the single biggest usability flaw with MS Office is the eternally morphing file formats that produce the lockstep upgrades that make MS so much money--

Umm, you do realize that the file formats for Word and Excel have been constant over the last five versions of Office (Office 97 for Windows, Office 98 for the Mac, Office 2000 for Windows, Office 2001 for the Mac, and Office XP for Windows), true?

-- Anonymous, June 02, 2001

Moderation questions? read the FAQ