> > Brian Kelly wrote: > > > > > Brian Kelly wrote: > > > > > > [long snip] > > > > > > > However I would agree that this is a legitimate > approach, although > > > > purists could argue that (a) Web resources should degrade > > > gracefully > > > > so we shouldn't be doing this or (b) we shouldn't be supporting > > > > fbrowsers with bugs in them. > > > > > > Are there any real-world browsers /without/ bugs in them? > > > > Perhaps I should have said significant bugs in them. For > example IE > > and Opera will do a good job for a lot (but not al) of the > CSS 1 spec, > > but Netscape 4 has major problems. > > > > What's your view? > > That we are constrained by opposing needs : > > 1) To conform to international standards > 2) To provide accessible documents > > The problem is that (2) is usually interpreted to mean > "Accessible by sighted, > partially sighted and blind users", whilst in the real world > it must also be interpreted to mean "Accessible from any > browser which has significant market penetration". Thus we > /cannot/ generate pages which are unreadable in NS 4.7+ > simply because we don't like NS's historical inability to > build browsers that conform to international standards. All > the while NS 4.7+ remains the browser of choice of a > significant number of potential visitors to a site, the > webmaster for that site has a duty to ensure that the site is > accessible to such visitors, in just the same way as he/she > has a duty to ensure that his/her site is accessible to blind > and partially sighted visitors. As a result of this, he/she > must /either/ program to a lowest common denominator, /or/ > serve page variants (possibly via variant CSS files, which in > turn can be generated on the fly using either client-side > JavaScript or server-side scripting) which ensure that the page is > legible on all common browsers. Thus, for the time being at > least (that is, all the while NS 4.7+ remains a commonly > found/used browser), CSS can be used only to /add value/ to a > page, not to convert it from gibberish to > a work of art. Hi Philip Many thanks for those comments. You've given a very clear and consistent argument. A couple of comments. How does one define "commonly-used"? Is this based on usage on one's own Web site or internationally (and if the latter are we not hostage to fortune to AOL?). Also at what level of usage does a browser cease to be commonly used. There is an argument that we should be developing richer, leaner, more accessible etc. services in order to provide motivation for people to upgrade their browsers (this reminds me of the time I worked at Liverpool University in 1990 and a small but vocal number of users did not want the VM/CMS mainframe to be switched off - if it had been up to them Liverpool, might still be running the system!) Also how do we reconcile this policy with other pressures to develop services using CSS, ECMAscript, etc. For example I understand that the popular VLE systems (WebCT, Blackboard) have minimum browser levels (correct me if I'm wrong). There's also the view of the browser vendors. I know that Netscape have stated that certain versions of Netscape are no longer supported (in some cases due to security problems). Brian --------------------------------------- Brian Kelly UK Web Focus UKOLN University of Bath BATH BA2 7AY Email: [log in to unmask] Web: http://www.ukoln.ac.uk/ Phone: 01225 38 3943 > ** Phil. >