On Wed, Mar 02, 2005 at 04:13:00PM +0100, Ralph Slooten wrote: > Umm, *cough* *cough* ;-) ... not nearly close to being remotely true > here .... Slashdot is so bad that they have even blocked the W3C > validator as users were complaining about it. Try it... save the index > file (main page) of /. and upload it to the validator: > > File: Slashdot.htm > Encoding: utf-8 > Doctype: HTML 3.2 > Errors: 115 > > 115 errors on one page is definitely not "perfectly valid html" :P > > Actually the reason I bring this up is not to drill your opinion into > the ground (nothing personal), but it goes back to when I often used > dillo to surf with. Dillo, like many other browsers had big issues > rendering slashdot. We did a lot of searching around then and found it > to have terrible code. There were several compaints sent to slashdot to > at least try fix their code. It seemed then only to get worse. > Eventually we gave up, and they seem to have blocked validation services > from showing just how bad their code actually is. > > All I'm trying to point out here is that you cannot base *any* browser > on slashdot's code... that is unless you are testing for crashes ;-0 > > Greetings > Ralph >
Actually, running the /. homepage through htmltidy showed that it is completely valid: 136 warnings and no errors. The warnings are all of the type empty <font> or <p> or <font> elements not being closed (but shouldn't really matter with proper scopes). Some others are reminents of /code, or tags missing suggested fields (for example, "summary" for <table>, I personally never used that...). As I said, any competant engine should be able to render the HTML (it works just fine in lynx and links (and it actually renders BETTER in links than in firefox for that matter)), which I take as meaning further evidence that the html is VALID. Note that VALID doesn't mean IDEAL. The w3c validator AFAIK makes noise as long as you don't follow their suggested practices (why should I put an ALT tag for an image if there is a caption clearly labeling the picture as such?). Of course, it is best to conform to the standard as much as possible, but slight and minor deviations like in the case of slashdot does not harm its validity. (Actually, the w3c validator also screams at news.com.com.com.com.com, and it renders just fine on most any browser I use). (going on a tangent here: the HTML validator really isn't the best tool here. For example, if you had bothered reading your 115 errors, you'd have realized that the vast majority is due to the & characters used in URLs, that cuts it down to around 50. And if you ignore the one's due to w3c's deprecation of certain older attributes, and the ones generated from the use of inline style sheets (which unfortunately, HTML 3.02 does NOT support officially, but 4+ do), and the ones generated from silly comments like "%" is not allowed in "width" or "#" is not a "literal character", you are down to 20 or so. Now some of them are just weird: what the heck is <nobr> anyway? But the rendering engines are supposed to ignore tags that they don't know about anyway (think lynx). The rest are issues with scopes: what tags are allowed to go inside of what else, which hardly anyone follows to the letter, and shouldn't affect rendering... In fact, none of the errors the validator produces has anything to do with BAD HTML code*, nor anything to do with why /. does not render properly in firefox 5 out of 6 tries. If you are a web designer, you are much better off using HTMLtidy to check your source to make sure it is sane and parseable...) However, just because the code is valid doesn't mean that it would look nice, or that it would render properly (as in, "the way we like it") on all agents. It's just like coding in any other language: you can write something syntactically correct that doesn't do jack when it comes to solving the problem it is supposed to solve. In this case /. chose an extremely bad design and very incompatible with Firefox, but in a sense "it rendered like the code told firefox to render it"... What I am trying to say here is not to disagree with the fact that /. is broken, but that it is broken not because of Invalid HTML, but because of a bad design philosophy. It renders. But often it renders poorly. And thus should not be something to test your browsers with. Sorry to make such a big rant on the pedantics of code validity, and of course, nothing personal, but I really want to make a point distinguishing VALID code and PROPER code. Remember: just because it compiles doesn't mean it does what you want it to. Best, W *BAD HTML code here means "stuff that can seriously fuck up the display agent" -- ---------------------------------------------------------------- * Address: 45 Spelman Hall, Princeton University 08544 * * Phone: x68958 AIM: AngularJerk * * E-mail: [EMAIL PROTECTED] From: sep.dynalias.net * ---------------------------------------------------------------- "Where's that fucking term!?" ~DeathMech, Some Student. P-town PHY 205 Sortir en Pantoufles: up 7 days, 10:12 -- [email protected] mailing list
