- From: Jukka K. Korpela <jkorpela@cs.tut.fi>
- Date: Tue, 31 Aug 2004 14:27:53 +0300 (EEST)
- To: Tomasz Orczyk <tzok@poczta.onet.pl>
- Cc: www-validator@w3.org
On Tue, 31 Aug 2004, Tomasz Orczyk wrote: > Your validator indicates this as an error "<a href="main.htm">=> > Strona glowna</a>". It "dislikes" > code, but I don't find anything wrong > with this, it's supposed to produce link "=> Strona glowna". There's nothing wrong syntactically with that link (though, to sidestep a bit, it would be better to put the arrow outside the link text and present it as a graphic or as an arrow character rather than "Ascii art"). > My page is www.tzok.republika.pl OK, this shows the problem. There is no DOCTYPE declaration. Hence, technically, no validation is possible, since validation _means_ testing against a document type definition. However, the validator - trying to be helpful - tries to guess the intended document type. This is where it goes a bit wild. It tells No DOCTYPE Found! Falling Back to HTML 4.01 Transitional but later says Line 1, column 0: no document type declaration; will parse without validation These are mutually contradictory, of course. And the latter describes what the validator is actually doing. It only performs generic SGML parsing, reading tags but not checking any nesting rules, acceptable attributes, etc. This isn't quite consistent, though; it reports e.g. Line 1, column 41: "DEFER" is not a member of a group specified for any attribute because such shorthand markup (which is technically short for DEFER="DEFER" in this case) is permitted in special case only, depending on how the possible values of attributes have been defined in the document type definition. But the validator shouldn't really say a word about this if it is really in "parse without validation" mode, since it does not know what the document type definition is. On similar grounds, Line 2, column 5: document type does not allow element "HTML" here means that the validator is secretly doing something else than it says. The issue with > is caused by "parse without validation" mode. In SGML itself, no entities - including > - are predefined. They need to be defined in the document type definition (and they have been defined in HTML DTDs). So what you should _you_ do? Well, fix the syntax, starting from the DOCTYPE declaration, as far as possible. But as a quick and dirty fix, you can use the validator's Extented Interface http://validator.w3.org/detailed.html so that from the Doctype menu you select "HTML 4.01 Frameset". There will still be error messages, but more understandable (caused by the <script> element appearing before <html> and some common frameset problems). > (first 3 errors are not caused by me and I > can't do anything to avoid them). Well I hope _someone_ can. The Web server administration ("webmaster") should be competent enough to understand and fix the error, once someone tells about it. -- Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Tuesday, 31 August 2004 11:28:26 UTC