- From: Charles McCathieNevile <charlesn@sunrise.srl.rmit.edu.au>
- Date: Sun, 6 Sep 1998 20:33:02 +1000 (EST)
- To: WAI GL <w3c-wai-gl@w3.org>
Following Daniel's extensive comments... I agree that validating HTML/CSS is important, but ensuring it is valid can be done manually or by a validator. If the technique said: Check source code against the HTML 4.0 Spec [href] or use an automatic validator such as ... then I think it should be P1, since having valid code is P1. Bobby is not a P1 - one of the reasons is that it is not always correct. Manual checking is potentially better. Again, maybe that should be in the techniques somewhere. Column layout is a P1 problem. it ought to be UA, but the problem has been created already. It may (in the future) go away. But at present it is crucial for a relatively small group of people, and the likelihood that they will all get good software and machines which run it in the next six months may not be that high. The problem is the same as D-link vs LONGDESC - it ought to be a UA problem, but the legacy (for all users) of history means that it can only presently be solved by the author (using a D-link). I still think that table layout of columns is P1. Unfortunately, there is a wider problem. TABLE is there because we use a lot of tabular information (think of spreadsheets). It is possible to use an alternative format (linearise the table) and the algorithms are not even terribly complex in most cases (especially where the table has a structure, rather than being arbitrarily constructed for layout reasons). But they are not well-implemented (I understand there is an implementation for the w3 browser in emacs, but when I was playing with Jason White looking at websites he didn't have the table package, and as a consequence tables were a problem. (Jason, what do you think about this one?) Charles McCN
Received on Sunday, 6 September 1998 06:56:52 UTC