Posted by Faust on 10/10/24 11:36

Wow! You are correct. That will do it. Sorry, Raj, I missed that,
drowned in information. If you had just added that one sentence it
would have been most helpful.

What I didn't bother to mention, because I didn't think that it was
relevant, was that I have already written code to invoke the W3C
validator and to parse the result. But I can certainly replace it with
this. Thanks to both of you.

Now we have the practical solution of *how* to validate an html page,
and we come to the philosphical questions of *whether* and *when* to
validate it.

Is it a good approach to validate a page each time that it loads?
Probably not, since this adds overhead, both to the page load and to
the W3C servers.

Perhaps each time that it changes? But what if the page doesn't change,
but some include, or its include does? How to know that?

Aha! perhaps I can compute a checksum of the dispalyed page and
validate if that changes. Hmm, no, this is no good on pags with hit
counters or any other variable content which can change but will never
break validation?

Maybe I should just have a cron job to validate my site overnight, or
once a week and email me if anything does not validate.

See what I mean? That's the sort of thing that I would like to disucss.
What do you think?

Thanks in advance

[Back to original message]