As I said above, “I’m going to be uncharacteristically modest and not ask that the Samurai amend their document because I do things differently sometimes. We all do, and that’s OK … Here’s how my personal best-practice differs from that described in WCAG+Samurai”.

This isn’t gladiators, my/ our criticism is constructive.

(I had said:)

<noscript>

Banning noscript is absolutely correct from a philosophical standpoint. However, it makes life much harder for me, as an author who’s not a scripter. As Chris Hunt writes,

“I use a noscript to hold a link to multimap for those whose lack of JS means they can’t use an inline Google Map. I suppose I could use a div which is hidden by a bit of script instead, but it all seems a bit cumbersome when there’s already an element out there to do the job.”

I do exactly the same and don’t know how to do it with script alone. I’ll continue to do so; I don’t support compromising accessibility to make the web author’s life easier, but I see no inherent inaccessibility in this use of the element.

* Browser is JS capable
* Firewall strips some (or all) external JavaScript

Hence, the content isn’t displayed and the JS does not get triggered.

The problem with noscript is old JavaScript browsers. Take Netscape 4 – it doesn’t support modern W3C DOM JavaScripts, but as soon as you use the noscript tag you ask it “Do you support JavaScript at all?”, and it’ll answer Yes, since by its own lights it supports JavaScript.

So the few remaining Netscape 4 users will not get the scripted interface, but neither will they get the noscript interface. That makes the site inaccessible.

True, this group of users is not very large, but it’s often relatively easy to avoid use of the noscript tag.

I’m not sure if this is what the WCAG Samurai meant, but it’s the reason I don’t use noscript any more.

You say “In some browsers invalid hacks can cause the rest of the style sheet to be ignored.” – do you have a reference to which browsers and which hacks these are? (And are disastrous bugs like a combination of characters causing rendering failure/ browser crash accessibility problems, except in the least-technical sense of the word?)

The argument against requiring people to use conditional comments is that they themselves are against specification (although they validate fine, and the argument isn’t a particularly compelling one to me).