Re: [RT] Getting rid of the table-based layout

[Miles Elam]

Stefano Mazzocchi wrote:

Robert Koberg wrote:

even if it is about .1% of the general Internet users? There might be an
argument if stats could be provided for the user base, but that does not seem to
be acceptable. What if you have 0 hits with Nav4? Do you still need to support
it?

Oh, god, we already had this conversation on forrest-dev a while ago and the figures say that xml.apache.org is hit by 10% of 4th generation (or less) browsers. So we decided to stick with 'some' HTML table layout.

I would love to go pure CSS as much as any of you, people, but we *must* understand that apache is a worldwide thing. And that 1% of the market we might be cutting off it probably the one that mostly needs our help to get out of their empasse.

Yes, it's a worldwide thing. It is also one of the biggest standard bearers for the W3C. While failure to provide information to all-comers is a big, fat mistake for any website -- especially the ASF's -- to continue publishing pages which don't pass spec is tacit approval of the practice to casual observers. Of all the sites on the Internet, aside from the W3C homepage, I can't think of any organization that should have standard, validating pages more than the ASF.

You can avoid the crashers (if you want). You can move your sites to CSS based
layout now it is a far superior way to handle page layouts! Last I saw IE has
around 97% of the market with the bulk being v5+. Still you can design the page
to degrade somewhat gracefully - the ns4 users are not that picky...

yes, as long as it doesn't crash, I agree: nobody is going to be picky if they don't support CSS.

Exactly!

I have been trying to get a CSS based site into forrest (along with some other
interesting ideas) but the prevailing order seems to be taking an ostrich
strategy toward many things (linking in particular)...

It won't be hard to have different stylesheets depending on the browser selection. So, let's use the power of cocoon instead of forcing CSS down everyone's neck.

Not always necessary. Especially if there is a well-marked demarcation line between standards-compliant browsers and pre-standard browsers, CSS can be made with little or no server-side client detection. For example, very old and text browsers don't know anything about the "link" tag so you hide CSS from mucking up their content. Obsolete browsers (Netscape 4-) don't understand the "@import" directive, so you can easily hide harmful CSS from them. The semi-standard browsers (IE 5.x for Windows) don't understand the ">" CSS selector. Basically what I am saying is that you can degrade gracefully with static files. This is of course a matter of opinion, but I have found that among browsers which support CSS (not trying to include Netscape 4), the differences are relatively minor: borders affecting widths and the like. Having three or four different stylesheets and doing client detection just seems like a whole lot of unnecessary redundancy to me.

But then again, this is all sophostry and rhetoric without something to look at or back it up with. So, getting to my point, I got bored today and made a mockup of http://xml.apache.org/cocoon/ in XHTML 1.0 Strict.

http://cocoon.iguanacharlie.com/

I hope this illustrates my point of view. Sure, it could use some tweaking (I just whipped it out), but it validates, looks pretty good for braille readers, looks pretty good for good browsers, and any browser should be able to get the content (including Mosaic holdouts who can't handle tables). Wasn't this the point of XHTML? Wasn't this the point of CSS?

I'm sorry if I'm starting another fight, but seeing non-standard pages on ASF sites has been paining me for some time. But rather than just complain, I'm trying an alternative and seeing how many people sigh. Given this, is it still worth avoiding CSS? (A hell of a lot easier to write XSLT for XHTML to be sure.)

I was also going to convert the GIFs to PNGs, but I'm quite sure that would've had me lynched before long because they most certainly don't show up in old browsers. Now pardon me while I duck the incoming bullets.

- Miles



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]