|Issue with markup validator fixing an issue causes more errors|
| 6:16 pm on Jan 3, 2012 (gmt 0)|
I am currently checking my website to see how valid it is to xhtml 1.0 transitional using W3C markup validator. On one page I had 1 error, a missing closing </ul> tag. I thought cool thats easy to fix, I checked the code and it was missing so I put the closing tag in and revalidated, however much to my surprise for the same page I now have 8 different errors, crazy errors like attributes contruct errors, tag mismatch errors and extra content at end of document errors.
I use a CMS and some of the tag mismatch errors would be in the CMS theme code and i just dont see them and these errors dont replicate on other pages and they should as the theme is site wide.
I am confused at why fixing an error would bring up more errors which are not there, at least I cannot see them and the validator does not see them on other pages.
Is the validator broken, or is something else throwing these errors?
| 6:56 pm on Jan 3, 2012 (gmt 0)|
Well I think i have fixed the validation errors, turns out there was an attributes construct error and possibly that was causing the mismatch errors, basiclaly there was no space between rel="" and class="" once i fixed that i got the page to validate.
I would still love to know why these errors never showed in the validator until i fixed the closing </ul> tag.
| 7:26 pm on Jan 3, 2012 (gmt 0)|
It is possible (particularly with [X]HTML) that the validator was simply unable to get passed your initial error (missing a closing </ul>), so fixing this error enabled the validator to continue...
| 7:38 pm on Jan 3, 2012 (gmt 0)|
On the flip side, there are many times where a cascading error with multiple error messages is fixed with just one edit.
| 10:42 pm on Jan 3, 2012 (gmt 0)|
Getting over a hundred errors is good. It means you made one mistake near the beginning; once you fix it, all the others will disappear. Getting a dozen or two errors is bad. It means you made a dozen or two separate mistakes :)
My personal record is 45,000 bona fide errors. But that wasn't from a w3c validator; it was from a specialized tool that also looked at things like incorrect punctuation and capitalization, unfamiliar letters and mismatched brackets. In a 500-page work produced by the Early English Text Society, this kind of error can mount up very very fast. Fortunately you can also ignore the report very very fast ;)
Beware of Tidy. I have never found a version that simply reports errors without "fixing" them, no matter how emphatically I tell it not to change anything.
You can also get interesting results in "validate by upload" if your finger slips and you feed in a plain-text file instead of HTML.
| 12:02 am on Jan 4, 2012 (gmt 0)|
I have seen it do this often. I liken it to peeling an onion. Once you have one layer off another emerges.