HTML validators - are they reliable?

[Jeremy]

I am using Firefox with the developer tools for validating the
documents. I can get a clean bill of health on there but then when I
paste the document into (as an example) the validator here
http://validator.w3.org/ I get presented with a bunch more issues (such
as the document not being valid XHTML 1.0 Transitional.

That's fine, I have no problems with needing to correct the code, but i
do wonder about the reliability of these checkers - I would assume that
the one at w3.org would be the definitive?

I have also been playing with CSE HTML Validator Lite 3.5, which again
finds errors/warnings that the Firefox validator misses.

Is the Firefox validator just "not quite up to it"?

Appeciate any tips or references to other validation utilities.

Many thanks!

 

[ato_zee]

That's fine, I have no problems with needing to correct the code, but i
do wonder about the reliability of these checkers - I would assume that
the one at w3.org would be the definitive?

I use the W3C one. but also have CSE HTML Validator Pro v6.0 100.
W3C seems comprehensive, but you should also use their
stylesheet validator, for full validation.
CSE HTML Validator, as their website points out, finds things that
W3C misses. Like mis-spellings, 7 numbers for colour values etc.
That said I usually use CSE validator at the final stages, it tends
to overwhelm you with warnings and errors, it's main value is it's
batch mode which can check 30 (or more pages), with a few
mouse clicks, to select all the html's in a local folder, and produce
a nice report.
So once the local copy of the site is clean, I run it in batch mode,
before uploading, just to check I haven't made any typo's or
introduced errors. You can't do batch mode at W3C.

 

[Harlan Messinger]

I am using Firefox with the developer tools for validating the
documents. I can get a clean bill of health on there but then when I
paste the document into (as an example) the validator here
http://validator.w3.org/ I get presented with a bunch more issues (such
as the document not being valid XHTML 1.0 Transitional.

That's fine, I have no problems with needing to correct the code, but i
do wonder about the reliability of these checkers - I would assume that
the one at w3.org would be the definitive?

I have also been playing with CSE HTML Validator Lite 3.5, which again
finds errors/warnings that the Firefox validator misses.

The creator of CSE HTML "Validator" has argued speciously on Usenet as
to why he calls his application a "validator" despite the fact that it
isn't one under the technical definition of the word. He doesn't flag
syntax errors that he considers to be acceptable in real usage, while at
the same time providing warnings and error messages about issues that
may violate provisions in the spec that fall outside of the scope of a
DTD, and about other matters where best practice may call for changes.

 

[Bergamot]

I have also been playing with CSE HTML Validator Lite 3.5, which again
finds errors/warnings that the Firefox validator misses.

Contrary to what its name implies, the CSE product is not a validator.
If it were, it would give results comparable to the W3C validator.

 

[Toby Inkster]

You can't do batch mode at W3C.

But... http://www.htmlhelp.com/tools/validator/

includes "Validate entire site" option.

 

[NoWhereMan]

Is the Firefox validator just "not quite up to it"?

if you're using html tidy for firefox, browse to the developer official
site and download the beta version which features both the tidy validator
and the sgml validator (which is the w3c one)

--
NoWhereMan
-- NoWhereBlog: www.nowhereland.it
-- FlatPress: http://flatpress.nowhereland.it blog flat-file powered,
xhtml1.1strict compliant
-- deviantArt: http://nowhereland.deviantart.com

 

[Spartanicus]

I am using Firefox with the developer tools for validating the
documents. I can get a clean bill of health on there but then when I
paste the document into (as an example) the validator here
http://validator.w3.org/ I get presented with a bunch more issues (such
as the document not being valid XHTML 1.0 Transitional.

That's fine, I have no problems with needing to correct the code, but i
do wonder about the reliability of these checkers - I would assume that
the one at w3.org would be the definitive?

I have also been playing with CSE HTML Validator Lite 3.5, which again
finds errors/warnings that the Firefox validator misses.

Is the Firefox validator just "not quite up to it"?

Both CSE and the Firefox "validator" are not validators at all. IIRC the
Firefox "validator" extension uses HTML tidy, a proprietary linter who's
opinion on HTML is just that: an opinion of the people who created it.

Same applies to CSE: it's opinion on HTML is nothing more than an
opinion of it's creator, and that opinion is worth very little.

There are various proper validators, both online and offline. Only these
should be trusted.

 

[Nikita the Spider]

Yes, it is. It's the Gold Standard. That doesn't mean that (a) it's
perfect, (b) it can't be bettered or (c) you don't benefit from
supplementing its reports with other syntax checking tools. But if you
only bother with one validator, that's the one to use.

I use the W3C one. but also have CSE HTML Validator Pro v6.0 100.
W3C seems comprehensive, but you should also use their
stylesheet validator, for full validation.

I haven't used it myself, but I've read many negative opinions about the
quality of the CSS validator. I'm not trying to discourage you from
using it, but caveat emptor.

CSE HTML Validator, as their website points out, finds things that
W3C misses. Like mis-spellings, 7 numbers for colour values etc.
That said I usually use CSE validator at the final stages, it tends
to overwhelm you with warnings and errors, it's main value is it's
batch mode which can check 30 (or more pages), with a few
mouse clicks, to select all the html's in a local folder, and produce
a nice report.
So once the local copy of the site is clean, I run it in batch mode,
before uploading, just to check I haven't made any typo's or
introduced errors. You can't do batch mode at W3C.

I know of two validators that will check more than one page at a time.
The HTMLHelp validator is free and is based on the same technology as
the W3C validator.
http://www.htmlhelp.com/tools/validator/

Click the "Validate entire site" option to get multiple pages validated.
The last time I used it, it stopped after 100 pages.

Alternatively, there's my spider (see my sig). It is also based on the
same technology as the W3C validator, has no page limit and will do link
checking for you too. It's free right now, eventually I plan to charge
money for it.

HTH

 

[John Hosking]

I am using Firefox with the developer tools for validating the
documents.

I'm very confused. I, too use the validator in the Web Developer Toolbar
(that's what you mean, right?). When I do, a window opens up showing a
report for the subject page at
http://validator.w3.org/check?verbose=1&uri=http://www.example.com/

I am sure I never changed the default validator, or had to specify
w3.org. How come the OP sees a difference, and someone like Spartanicus
says, "Both CSE and the Firefox "validator" are not validators at all?"

What am I missing?

 

[cwdjrxyz]

I am using Firefox with the developer tools for validating the
documents. I can get a clean bill of health on there but then when I
paste the document into (as an example) the validator here
http://validator.w3.org/ I get presented with a bunch more issues (such
as the document not being valid XHTML 1.0 Transitional.

That's fine, I have no problems with needing to correct the code, but i
do wonder about the reliability of these checkers - I would assume that
the one at w3.org would be the definitive?

I have also been playing with CSE HTML Validator Lite 3.5, which again
finds errors/warnings that the Firefox validator misses.

Is the Firefox validator just "not quite up to it"?

Appeciate any tips or references to other validation utilities.

To get a more complete discussions concerning validators and some other
related issues that have not been brought up in this thread, the W3C's
page http://validator.w3.org/docs/help.html may be helpful.

 

[David Dorward]

I'm very confused. I, too use the validator in the Web Developer Toolbar
(that's what you mean, right?). When I do, a window opens up showing a
report for the subject page at

http://validator.w3.org/check?verbose=1&uri=http://www.example.com/

That is the W3C Markup Validation service ...

I am sure I never changed the default validator, or had to specify
w3.org. How come the OP sees a difference, and someone like Spartanicus
says, "Both CSE and the Firefox "validator" are not validators at all?"

.... when people talk about the Firefox validator they usually mean
http://users.skynet.be/mgueury/mozilla/ which (unless you use the beta and
turn the right options is) is based around Tidy, which isn't a validator
(and is rather buggy).

 

[John Hosking]

... when people talk about the Firefox validator they usually mean
http://users.skynet.be/mgueury/mozilla/ which (unless you use the beta and
turn the right options is) is based around Tidy, which isn't a validator
(and is rather buggy).

Thanks, David, for the tip. I've never been *there* before. I see (by
back-researching) that it's an extension for Firefox, listed on the
extensions page as "Html Validator 0.7.9.5". Not something I'd ever
place too much faith in.

 

[NoWhereMan]

... when people talk about the Firefox validator they usually mean
http://users.skynet.be/mgueury/mozilla/ which (unless you use the beta and
turn the right options is) is based around Tidy, which isn't a validator
(and is rather buggy).

latest beta of the extension (very stable) has two (selectable) engines.
tidy and the SGML parser. SGML parser is developed by w3c and it should be
actually the libaray on which the online validator is based

--
NoWhereMan
-- NoWhereBlog: www.nowhereland.it
-- FlatPress: http://flatpress.nowhereland.it blog flat-file powered,
xhtml1.1strict compliant
-- deviantArt: http://nowhereland.deviantart.com

 

[Spartanicus]

latest beta of the extension (very stable) has two (selectable) engines.
tidy and the SGML parser. SGML parser is developed by w3c and it should be
actually the libaray on which the online validator is based

IIRC the validator option is based on openSP. openSP was developed by
James Clark independently of W3C.

 

[NoWhereMan]

IIRC the validator option is based on openSP. openSP was developed by
James Clark independently of W3C.

yes, but:
http://validator.w3.org/docs/help.html#how


download of the ffox extension:
http://users.skynet.be/mgueury/mozilla/preview_080.html


--
NoWhereMan
-- NoWhereBlog: www.nowhereland.it
-- FlatPress: http://flatpress.nowhereland.it blog flat-file powered,
xhtml1.1strict compliant
-- deviantArt: http://nowhereland.deviantart.com

 

[Jukka K. Korpela]

Scripsit Harlan Messinger:

The creator of CSE HTML "Validator" has argued speciously on Usenet as
to why he calls his application a "validator" despite the fact that it
isn't one under the technical definition of the word.

He has made some appearances in HTML-related groups, typically _only_ to
advertize his product and to repeat the same babble about it in response to
criticism. Having lost his face so many times, I wouldn't be surprised if he
tried using a fake identity and pretending to be a satisfied customer.

 

[ato_zee]

It's an imperfect world. Editors can generate code that validates, but doesn't
work, W3C validator can miss some things. Some browsers work with
some pages and not others.
Personally I use W3C validator at the design stage, and CSE validator
at the end to see if it picks up anything I've missed.
DW8 is useful for some things, to see nesting/layout, find orphened
files, check links, but it doesn't pick up a lot of errors.
So I code in a text editor.
There is no single tool that does everything.

 

[Albert Wiersch]

He has made some appearances in HTML-related groups, typically _only_ to
advertize his product and to repeat the same babble about it in response
to criticism. Having lost his face so many times, I wouldn't be surprised
if he tried using a fake identity and pretending to be a satisfied
customer.

I still have my face, thanks.

No need to pretend to be a satisifed customer. First, I wouldn't do that
because it's wrong. Second, there are plenty of satisfied customers already
who have expressed their satisfaction on the CSE HTML Validator testimonial
page.

As for the original question, HTML validators are great tools that should
always be used, but they're not perfect. They can't tell you that a page you
wrote will display perfectly as intended. Some validators are better than
others.

Albert

 

[Jukka K. Korpela]

Scripsit Albert Wiersch:

As for the original question, HTML validators are great tools that
should always be used, but they're not perfect.

Nonsense. They are primitive tools that we use just in lack of more useful
checkers. Besides, there's really no such thing as "HTML validator". It's an
SGML validator, or XML validator, or both, and if you use it for HTML
documents, that's your decision.

But at least validators are what they claim to be, modulo a few bugs
perhaps, and not half-baked checkers sold under an intentionally misleading
name like "CSE HTML Validator".

 

[Albert Wiersch]

Scripsit Albert Wiersch:


Nonsense. They are primitive tools that we use just in lack of more useful
checkers. Besides, there's really no such thing as "HTML validator". It's
an SGML validator, or XML validator, or both, and if you use it for HTML
documents, that's your decision.

Call it a primitive tool if you want. It doesn't matter. A hammer is still a
"primitive" tool but still quite useful.

Have you notified the W3C that there's no such thing as an "HTML Validator"?
It seems they have a link on their front page called "HTML Validator" and
have had that for quite some time. Try searching the web for "HTML
Validator". It seems like you live in your own world with your own
definitions.

But at least validators are what they claim to be, modulo a few bugs
perhaps, and not half-baked checkers sold under an intentionally
misleading name like "CSE HTML Validator".

Again, you continue to say more things that aren't true.

Albert