- From: Smith, Brooke <Brooke.Smith@Butterworths.com.au>
- Date: Mon, 4 Nov 1996 13:50:42 +1000
- To: "'www-html@w3.org'" <www-html@w3.org>
This thread was conceived from the statements of Murray Altheim <murray@spyglass.com> in response Jason O'Brien" <jaobrien@fttnet.com> with the analogy of two restaurants. It came down to presentation v's content. I would like to introduce another analogy (outline) that should be kept in mind when discussing this - The seven o'clock news (on a non-commercial channel) and Married with Children. There are two sets of people - those who watch the news, and those that watch Married with Children. There would be an intersection, mind you most likely a small one. Now I'd connect MwC w/ the Restaurant#2 that Murray described; > Restaurant #2: There is no handicap access. The front of the store is > covered with neon and gaudy signs that scroll and <BLINK>, and the signs > are in some sort of code that only the wealthy can afford to decode by > paying some upgrade fee every six months. Once in the restaurant (if you > can get in at all), the menus are obscured by all sorts of flashing > lights > that make it difficult to read, much less eat in peace. There's BGSOUND > blaring over the speakers, so you can't have a decent conversation. And > because they spent all their time and money on hype, they couldn't > afford > to hire a decent chef, so the food is terrible, and expensive. And they > haven't spent much time cleaning the place, because they're busy keeping > up > to date with the latest marketing glitz, updating their facades to the > latest models, etc. If the 99 rating on the restaurant was made by > someone > judging the restaurant on its food, then the reviewer was bribed. After > all, this is where all the money is. and the News with Restaurant#1; > Restaurant #1: The floors are clean, everything is in order. If you're > blind or in a wheelchair, you can safely enter and navigate the > interior, > and actually eat a meal. You can see the menu, which is straightforward, > and displays all the items in an easy-to-read manner. (Why does a > minimal > Web page have to smell bad and be dirty? They've got plenty of time to > clean up -- they're not learning the latest JavaScript trick.) In fact, > the > menus are available in several languages, because the proprietor > actually > values foreign customers. It doesn't cost an arm and a leg, and the food > is > actually tasty: they can afford a good chef. You figure the reviewer > (rating this place a 52) wasn't particularly interested in the food > (maybe > he was impressed by the flashing lights of other restaurants), since it > is > actually quite good. But a lot of people like Married with Children, and many of those have no interest in watching the news. Once upon a time the socio-economic-educated status of such people (the Web equivalents) would have prevented them from accessing the Web (actually the Internet - before the WWW). Without meaning to show some kind of x-ism, I would also claim that the population of such people on the web is increasing - and the unfortunate side of democracy is that the majority rules, even if it means we have to put up with shite. This leads to a suggested solution - customisation. > >--- > >| And as soon as you design into a page something that shuts out a > >| substantial part of your intended audience, that's poor design. I > >have a > >| friend who frustratingly can't read much of WiReD due to some sight > >| problems -- he has no difficulty with The Economist. > >--- > > > >Actually, I have more trouble with the Economist than with WiReD, due > >to > >the age of my eyes and their preferred font/size choice. But I agree > >with what you say, so long as we note the "intended" in there. The > >designer may need to choose to serve some users at the cost of not > >serving some others, or to serve some better than others. > > > >I strongly agree with the value of including ALTs, alternative paths > >for less-powerful browsers, and content negotiation. > These are perfect examples of the strength of customisation. We have been discussing style sheets and how the client can over-ride the authors settings to suit there taste (including changing output devices such as a speech synthesizer). Brilliant - the basis is already there. I wrote for my honours thesis last year a customisation system for the web which allowed authors to write documents which had customisation texts that were aimed at a certain target audience. For example an author writes a document on 'C' programming, and has in mind a target audience of programmers of other languages. The document consists of generic text - not customised, and customised text which I called 'points'. When the author writes a point they break it up into 'snippets' which are the individualised customised text for the particular readers. There is no AI in this system, just a 'if you are such and such a person then you see this, else if you're that type of person then you see that etc...'. This system uses a User-Modeling system to maintain and manipulate data on users. I was quite pleased with my efforts - it worked and got me a first-class honours score. As I said in my thesis, this (I called the system 'Custor') is something that itself will never go past the Sydney University gates (Well it did on a floppy disk :), but the idea will go far. I'd like to refer you to a URL but alas I just have never got around to putting the damm thing on my home page (does anyone know of a DTD for RTF to SGML and an Omnimark (Cross-translate) script for this - people tell me that it is hard to maintain these because M$ keeps moving the goal posts in relation to the RTF spec. I know about RTF2HTML, but I was wanting to store my works in an SGML database). Web documents can be easily customised in ways similar to this. Try: The Web server asks the Client "should I include a Java script or provide an inline-gif/text to express the same information (if any)?", upon which the client looks in the user's user model and sends back (via the secure sockets layer) information to answer this. "Does the user prefer white text on silver, black on yellow,..., or should I let them set the style for <em> themselves?" upon which the appropriate style sheet is used (after the answer returns). Solid examples but fairly easily achievable. Now that I have posted this I'll make an effort to get my thesis (or a paper I've written since) on display. Brooke ~~~~~~~~~~~~~~~~~~~~~ Brooke Benjamin Oehm Smith Butterworths North Ryde Australia http://www.cs.su.oz.au/~bbos
Received on Sunday, 3 November 1996 22:51:22 UTC