Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!linus!decvax!ittvax!wex From: wex@ittvax.UUCP (Alan Wexelblat) Newsgroups: net.cog-eng Subject: Re: User-Friendly Re-Defined Access-Efficient Message-ID: <985@ittvax.UUCP> Date: Thu, 1-Sep-83 09:20:30 EDT Article-I.D.: ittvax.985 Posted: Thu Sep 1 09:20:30 1983 Date-Received: Thu, 1-Sep-83 17:08:04 EDT References: <976@ittvax.UUCP> ulysses.590 Lines: 43 Hmmm. First, Gary, let me apologize for misspelling your last name. I hate it when people goof mine, and I'm sorry I wasn't more careful with yours. Second, let me explain where I'm coming from. I agree that I overgeneralized. That's wrong in itself. But most of my experience in interface design hasn't been on nice powerful UN*X machines, with libraries and directory trees and tools galore. Most of the businesses that I'm familiar with are at the level of a PC or two, or perhaps slightly above. I'm lucky if I have a Pascal to work with. If I'm really unlucky, I'm stuck trying to customize/improve something that someone else has written, in assembler or (YEECH -- are you listening, Mr Pournelle?) BASIC. On machines like these, space and speed definitely ARE factors to be considered. Furthermore, most of the companies I'm doing this work for are only going to use it in-house; so, the cost of my time is quite significant. Also, they're not likely to upgrade equipment all that fast (they usually don't need it, or can't afford it). I think this is a significant problem. Look at how many businesses own PC's vs how many own VAXEN. Just because they have a micro is no reason for them to have a bad user interface, but there's a limit to what I can do, given that it is a micro. And this problem is going to get bigger, not go away. Lastly, I want to defend my statements on the programmer as human-factors designer. I'm sorry if I over-generalized, but I'll bet dollars to doughnuts that I know better than the customer HOW to build a user interface. I need to hear suggestions and criticism about WHAT goes into the interface, but the methods should be left to the programmer. Of course, there are exceptions. If I build something and the customer says "I hate it; it's hard to use," then I've failed and should redo it. No doubt about that. But far more often, the customer insists on "a menu system," or something (as in Laura's example) that is really counter-productive. If the programmer does not talk the customer out of such ideas, then s/he is not doing the job properly. Maybe you're right. Maybe the average programmer does build lousy interfaces. In that case, I apolgize again, since I put a "we" where I meant to put "I". I think I design good systems, and the people I have worked for seem pleased by them. --Alan Wexelblat (until 9/2: decvax!ittvax!wex) (after 9/12: ucbvax!wex.UPenn@UDel-Relay)