Local server HTML validator

[Albert Wiersch]

Re: Local server HTML validator


"Stan Brown" <the_stan_brown @fastmail.fm> wrote in message
news:MPG.1ae91f 46d31d3d7f98c20 7@news.odyssey. net...[color=blue]
>
> The intern has installed this CSE "Validator" on the library
> computer and is presumably using it to test, or develop, or
> something. Since she obviously doesn't know any better, how do I
> explain to her in words of one syllable what is wrong?[/color]

*Most* people, including her, are better off with CSE HTML Validator than
using a "real" validator. I wouldn't be too quick to discount a useful
product because a couple of people here have an issue with the name.

Be sure to read these pages:

http://www.htmlvalidator.com/htmlval/whycseisbetter.html

and

http://www.htmlvalidator.com/htmlval/testimonials.html

--
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Stan Brown" <the_stan_brown @fastmail.fm> wrote in message
news:MPG.1ae91e 247b90be8e98c20 5@news.odyssey. net...[color=blue]
>
> But if you're knowingly writing invalid HTML, doesn't the whole idea
> of "validating " it go out the window?[/color]

No, of course not. The exception is only if one is a purist and is limited
to only using standards and 100% technical correctness. Most of the world
isn't so technical and perfectionistic (and that can be a good thing!).

I get plenty of value and use out of validating HTML that contains some
"invalid" markup that I choose to use.

--
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Stan Brown" <the_stan_brown @fastmail.fm> wrote in message
news:MPG.1ae91e 247b90be8e98c20 5@news.odyssey. net...[color=blue]
>
> But if you're knowingly writing invalid HTML, doesn't the whole idea
> of "validating " it go out the window?[/color]

No, of course not. The exception is only if one is a purist and is limited
to only using standards and 100% technical correctness. Most of the world
isn't so technical and perfectionistic (and that can be a good thing!).

I get plenty of value and use out of validating HTML that contains some
"invalid" markup that I choose to use.

--
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Stan Brown" <the_stan_brown @fastmail.fm> wrote in message
news:MPG.1ae91e 8a323d5e1b98c20 6@news.odyssey. net...[color=blue][color=green]
> >I'm not a purist. I do not care if I "tell a lie"[/color]
>
> Then why should anyone be interested in anything you say?[/color]

Normally it is a good thing to try to understand what someone is saying by
taking it in context. Why not try requoting me properly?

If you don't care what I have to say, that's your choice. I don't mind.

Take care,
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Stan Brown" <the_stan_brown @fastmail.fm> wrote in message
news:MPG.1ae91e 8a323d5e1b98c20 6@news.odyssey. net...[color=blue][color=green]
> >I'm not a purist. I do not care if I "tell a lie"[/color]
>
> Then why should anyone be interested in anything you say?[/color]

Normally it is a good thing to try to understand what someone is saying by
taking it in context. Why not try requoting me properly?

If you don't care what I have to say, that's your choice. I don't mind.

Take care,
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Jan Roland Eriksson" <jrexon@newsguy .com> wrote in message
news:5pau70pm3d tkj032sujo2f0c5 dat32uraa@4ax.c om...[color=blue]
> On Thu, 15 Apr 2004 19:28:13 +0100, Steve Pugh <steve@pugh.net > wrote:
>
>
> From a formal SGML standpoint, we need to correct this statement :-)
>
> Comments in a "Document Type Definition" entity is not allowed to be
> read as "documentat ion" of semantics since the DTDefinition itself is
> only about markup syntax in the first place.
>
> Also the inclusion of a link, to the governing "Formal Prose" part of an
> application of SGML, is questionable. The SGML Declaration (if any) and
> the DTDefinition comes as a result of the "Formal Prose" definition for
> an application of SGML so in effect we should look at this as "traveling
> down a one way street".
>
> Formal Prose definition comes first, if that one calls for
> changes/additions to the SGML Reference Concrete Syntax -- write an SGML
> declaration to the fact. Finally, and still based on the prose, write a
> DTDefinition that expresses the syntactical rules for the markup that
> conforms to the prose. The rest of it, i.e. the semantic definitions for
> that markup, will still only be available in the "Formal Prose" part.[/color]

Good information!

So, if I interpreted your post correctly, it seems that it makes sense to
check HTML according to the "formal prose" as well as the DTD? Since a
"real" DTD validator only checks according to the DTD, wouldn't it be a good
idea to also check according to the "formal prose", perhaps by using a
program that goes beyond DTD checking?

--
Albert Wiersch

http://www.htmlvalidator.com/

[Albert Wiersch]

Re: Local server HTML validator


"Jan Roland Eriksson" <jrexon@newsguy .com> wrote in message
news:5pau70pm3d tkj032sujo2f0c5 dat32uraa@4ax.c om...[color=blue]
> On Thu, 15 Apr 2004 19:28:13 +0100, Steve Pugh <steve@pugh.net > wrote:
>
>
> From a formal SGML standpoint, we need to correct this statement :-)
>
> Comments in a "Document Type Definition" entity is not allowed to be
> read as "documentat ion" of semantics since the DTDefinition itself is
> only about markup syntax in the first place.
>
> Also the inclusion of a link, to the governing "Formal Prose" part of an
> application of SGML, is questionable. The SGML Declaration (if any) and
> the DTDefinition comes as a result of the "Formal Prose" definition for
> an application of SGML so in effect we should look at this as "traveling
> down a one way street".
>
> Formal Prose definition comes first, if that one calls for
> changes/additions to the SGML Reference Concrete Syntax -- write an SGML
> declaration to the fact. Finally, and still based on the prose, write a
> DTDefinition that expresses the syntactical rules for the markup that
> conforms to the prose. The rest of it, i.e. the semantic definitions for
> that markup, will still only be available in the "Formal Prose" part.[/color]

Good information!

So, if I interpreted your post correctly, it seems that it makes sense to
check HTML according to the "formal prose" as well as the DTD? Since a
"real" DTD validator only checks according to the DTD, wouldn't it be a good
idea to also check according to the "formal prose", perhaps by using a
program that goes beyond DTD checking?

--
Albert Wiersch

http://www.htmlvalidator.com/

[David Dorward]

Re: Local server HTML validator

Albert Wiersch wrote:
[color=blue]
> I get plenty of value and use out of validating HTML that contains some
> "invalid" markup that I choose to use.[/color]

That is a contradiction. HTML can not be validating if it contains invalid
markup.

And the amount of value you get out if it is obvious if you can't remember
why you used non-standard markup to achieve something that can be achieved
easily with standard markup except *maybe* in some browser you can't
identify (thus removing the need for you to provide evidence that you
really did have a good reason).

(reference: <news:se-dnSHjmriZQePd4p 2dnA@august.net >)

--
David Dorward <http://blog.dorward.me .uk/> <http://dorward.me.uk/>

[David Dorward]

Re: Local server HTML validator

Albert Wiersch wrote:
[color=blue]
> I get plenty of value and use out of validating HTML that contains some
> "invalid" markup that I choose to use.[/color]

That is a contradiction. HTML can not be validating if it contains invalid
markup.

And the amount of value you get out if it is obvious if you can't remember
why you used non-standard markup to achieve something that can be achieved
easily with standard markup except *maybe* in some browser you can't
identify (thus removing the need for you to provide evidence that you
really did have a good reason).

(reference: <news:se-dnSHjmriZQePd4p 2dnA@august.net >)

--
David Dorward <http://blog.dorward.me .uk/> <http://dorward.me.uk/>

[David Dorward]

Re: Local server HTML validator

Albert Wiersch wrote:
[color=blue]
> That's not CSE HTML Validator generating those warnings. CSE has no
> problem with <p/>.[/color]

Shame. Browsers do.

--
David Dorward <http://blog.dorward.me .uk/> <http://dorward.me.uk/>

[David Dorward]

Re: Local server HTML validator

Albert Wiersch wrote:
[color=blue]
> That's not CSE HTML Validator generating those warnings. CSE has no
> problem with <p/>.[/color]

Shame. Browsers do.

--
David Dorward <http://blog.dorward.me .uk/> <http://dorward.me.uk/>

[Jukka K. Korpela]

Re: Local server HTML validator

"Albert Wiersch" <mrinternetREMV ETHISUPPERCASET ORPLYnews@wiers ch.com>
wrote:
[color=blue]
> Good information![/color]

We are not surprised at seeing that factual information about SGML was
new to the advertizer of the so-called "CSE HTML Validator".
[color=blue]
> So, if I interpreted your post correctly, it seems that it makes
> sense to check HTML according to the "formal prose" as well as the
> DTD?[/color]

Yes. Have to tried doing that to your markup?
[color=blue]
> Since a "real" DTD validator only checks according to the DTD,
> wouldn't it be a good idea to also check according to the "formal
> prose", perhaps by using a program that goes beyond DTD checking?[/color]

Maybe for some SGML applications, if someone wrote such a program. Surely
you cannot do that without knowing SGML, without knowing what "formal
prose" means here, and without actually reading the prose of the
specification of the SGML application in question.

In the case of HTML, the possibilities are fairly limited. You could, for
example, report <p></p> as valid but incorrect (at least at some level -
it violates a specific "should not" statement), but even <p>&nbsp;</p> (a
common product of wysiwyg software) is questionable, since the prose is
obscure in this issue. (Is a paragraph consisting of a no-break space
empty?) And although <p class="heading2 ">Introduct ion</p> rather
undeniably violates the prose of the HTML specifications, while
unquestionably valid, it is hardly possible to capture such cases, except
perhaps desultorily on a heuristic basis. And the more cases you try to
cover with heuristics, the more errors in error messages you generate.

The "CSE HTML Validator", which is not a validator (and not even an
imaginary validator) but is being sold and advertized as a validator,
despite several mentions of this fact in public through years,
does not even come close to _trying_ to be a checker that consists of a
validator and additional checks against the prose of the HTML
specifications. It just reflects its author's ideas, opinions, and
tastes, which may occasionally coincide with the formalized requirements
in HTML DTDs, or the prose of HTML specifications, or other W3C
recommendations - but its author hasn't bothered distinguishing such
issues from rules he just made up.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Jukka K. Korpela]

Re: Local server HTML validator

"Albert Wiersch" <mrinternetREMV ETHISUPPERCASET ORPLYnews@wiers ch.com>
wrote:
[color=blue]
> Good information![/color]

We are not surprised at seeing that factual information about SGML was
new to the advertizer of the so-called "CSE HTML Validator".
[color=blue]
> So, if I interpreted your post correctly, it seems that it makes
> sense to check HTML according to the "formal prose" as well as the
> DTD?[/color]

Yes. Have to tried doing that to your markup?
[color=blue]
> Since a "real" DTD validator only checks according to the DTD,
> wouldn't it be a good idea to also check according to the "formal
> prose", perhaps by using a program that goes beyond DTD checking?[/color]

Maybe for some SGML applications, if someone wrote such a program. Surely
you cannot do that without knowing SGML, without knowing what "formal
prose" means here, and without actually reading the prose of the
specification of the SGML application in question.

In the case of HTML, the possibilities are fairly limited. You could, for
example, report <p></p> as valid but incorrect (at least at some level -
it violates a specific "should not" statement), but even <p>&nbsp;</p> (a
common product of wysiwyg software) is questionable, since the prose is
obscure in this issue. (Is a paragraph consisting of a no-break space
empty?) And although <p class="heading2 ">Introduct ion</p> rather
undeniably violates the prose of the HTML specifications, while
unquestionably valid, it is hardly possible to capture such cases, except
perhaps desultorily on a heuristic basis. And the more cases you try to
cover with heuristics, the more errors in error messages you generate.

The "CSE HTML Validator", which is not a validator (and not even an
imaginary validator) but is being sold and advertized as a validator,
despite several mentions of this fact in public through years,
does not even come close to _trying_ to be a checker that consists of a
validator and additional checks against the prose of the HTML
specifications. It just reflects its author's ideas, opinions, and
tastes, which may occasionally coincide with the formalized requirements
in HTML DTDs, or the prose of HTML specifications, or other W3C
recommendations - but its author hasn't bothered distinguishing such
issues from rules he just made up.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Jukka K. Korpela]

Re: Local server HTML validator

Stan Brown <the_stan_brown @fastmail.fm> wrote:
[color=blue]
> The intern has installed this CSE "Validator" on the library
> computer and is presumably using it to test, or develop, or
> something. Since she obviously doesn't know any better, how do I
> explain to her in words of one syllable what is wrong?[/color]

That's quite a challenge, but it is useful to try and learn to express
oneself with monosyllabic words at times.

It [the "CSE HTML Validator" - sorry I can't _name_ it using words of one
syllable :-) ] just tells what its author likes. Don't use it if you
don't know the real rules well enough to tell what's right and what's
wrong in what it spits out. It's known to make false claims 'bout things
being wrong when they are in fact right. It works by rules that someone
just made up.

(Sorry about the occasional bisyllabics.)

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Jukka K. Korpela]

Re: Local server HTML validator

Stan Brown <the_stan_brown @fastmail.fm> wrote:
[color=blue]
> The intern has installed this CSE "Validator" on the library
> computer and is presumably using it to test, or develop, or
> something. Since she obviously doesn't know any better, how do I
> explain to her in words of one syllable what is wrong?[/color]

That's quite a challenge, but it is useful to try and learn to express
oneself with monosyllabic words at times.

It [the "CSE HTML Validator" - sorry I can't _name_ it using words of one
syllable :-) ] just tells what its author likes. Don't use it if you
don't know the real rules well enough to tell what's right and what's
wrong in what it spits out. It's known to make false claims 'bout things
being wrong when they are in fact right. It works by rules that someone
just made up.

(Sorry about the occasional bisyllabics.)

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html