How important is validation?

[Scott Meyers]

I have a web site that, due to maintenance by several people, some of whom are
fairly clueless about HTML and CSS, etc. (notably me), has gotten to the point
where I'm pretty sure it's suffering from bit rot. Though the pages seem to
display okay under IE and FF, I really think it's time for an under-the-hood
cleaning. I recently received a copy of Molly Holzschlag's "Spring Into HTML
and CSS," and in the first chapter, she makes a big deal of producing pages
that validate cleanly. However, she doesn't explain why this is important,
e.g., doesn't say what the consequences of validation failure are.

I went to http://validator.w3.org/ and was unsurprised to see my home page
fail to validate. But then I got to playing around, and I found that the home
pages for none of the following validate, either: yahoo, ebay, google, artima,
and cnn. This makes me wonder whether validation is really something I need
to worry about. Morally, I'm all for standards, and given a choice between
pages that validate and those that do not, I'd choose validation, but I'm
going to have to find somebody else to do the work for me (somebody who DOES
know about HTML and CSS, etc.), and I'm worried that finding somebody who is
familiar with validation is going to be a lot harder and/or more expensive
than finding somebody who is not.

Can somebody please explain to me what the practical advantages of having
pages validate are? Also, I'm open to suggestions on who to consider hiring
to do the work at my site (which happens to be aristeia.com).

Thanks,

Scott

[Matt Silberstein]

On Mon, 15 Aug 2005 11:50:42 +0100, in
comp.infosystem s.www.authoring.html , "Alan J. Flavell"
<fl*****@ph.gla .ac.uk> in
<Pi************ *************** ***@ppepc56.ph. gla.ac.uk> wrote:

[snip]

Well, I'm sorry, but from my point of view, markup syntax validation
is a natural part of the QA process for web publishing; I don't feel I
need to justify it, any more than I'd need to justify checking the
spelling of the content, or making a reasonable effort to verify the
facts which I'm going to report on the page.

Everyone in this thread seems to be assuming that the code was hand
generated. If so, then I agree that validation is crucial. But if the
code is developed in either a home grown or standard high level tool,
then validation is less important. I am currently working on some
Perl/Javascript tools that are going to output web pages (html and
css). I want those to produce valid code, but it is more important
that it produces the right result (and result is defined as what the
customer sees). If I end up having some validation bug, but get the
right customer view, I will accept that. I am certainly not going to
go back and fix the resulting html, that is impossible.

A similar case applies for some standard tool. Yes, FrontPage is
garbage, but if, say, Dreamweaver produced some invalid code, that
would be fine. It is more important to have a product that is easy to
modify, which means no hand coding, than one that meets validation. (I
have no idea if Dreamweaver ever produces invalid code. It might not,
my point still stands.)
--
Matt Silberstein
And now our bodies are oh so close and tight
It never felt so good, it never felt so right
And we're glowing like the metal on the edge of a knife
C'mon! Hold on tight!
C'mon! Hold on tight!

Though it's cold and lonley in the deep dark night
I can see paradise by the dashboard light
Paradise by the dashboard light

Jim Steinman

[Alan J. Flavell]

On Tue, 16 Aug 2005, Matt Silberstein wrote:

On Mon, 15 Aug 2005 11:50:42 +0100, in
comp.infosystem s.www.authoring.html , "Alan J. Flavell"

Well, I'm sorry, but from my point of view, markup syntax validation
is a natural part of the QA process for web publishing; I don't feel I
need to justify it, any more than I'd need to justify checking the
spelling of the content, or making a reasonable effort to verify the
facts which I'm going to report on the page.
Everyone in this thread seems to be assuming that the code was hand
generated.

Not really. If I found that I was using a process which generated
invalid code, I'd do whatever was necessary to that process to make
sure it generated valid code. In fact I've often done it with
inherited CGI programs which emit HTML that proved to be unacceptable
to my QA standards.

[...] but if, say, Dreamweaver produced some invalid code, that
would be fine.
Not by me. On the rare occasions when our principal author of
official web pages (who used DW) turned-in code that failed
validation, I (as server admin) had a quiet word with him, and he
corrected it without argument. Recalling that we are bound by UK law
to produce accessible pages, failure of syntax validation is a black
mark against the WAI guidelines, even for those who can't find any
other reason for valid syntax. But his standards were higher than
that - anyone can make an occasional mistake (including me, of
course).
It is more important to have a product that is easy to
modify, which means no hand coding,

I'd rate that as petitio principii. Depends very much on
circumstances. Say, how many authoring packages do *you* know which
understand that a block of quoted text is a <blockquote> without the
author ever telling them?

best regards

[Matt Silberstein]

On Tue, 16 Aug 2005 18:55:22 +0100, in
comp.infosystem s.www.authoring.html , "Alan J. Flavell"
<fl*****@ph.gla .ac.uk> in
<Pi************ *************** ****@ppepc56.ph .gla.ac.uk> wrote:

On Tue, 16 Aug 2005, Matt Silberstein wrote:

On Mon, 15 Aug 2005 11:50:42 +0100, in
comp.infosystem s.www.authoring.html , "Alan J. Flavell"

>
>Well, I'm sorry, but from my point of view, markup syntax validation
>is a natural part of the QA process for web publishing; I don't feel I
>need to justify it, any more than I'd need to justify checking the
>spelling of the content, or making a reasonable effort to verify the
>facts which I'm going to report on the page.

Everyone in this thread seems to be assuming that the code was hand
generated.

Not really. If I found that I was using a process which generated
invalid code, I'd do whatever was necessary to that process to make
sure it generated valid code. In fact I've often done it with
inherited CGI programs which emit HTML that proved to be unacceptable
to my QA standards.

It is the "whatever was necessary" that I take (some) issue with. I
agree, of course, that good code is important and passing validation
is an important step. How much work to fix what problems is a question
of art.

That said, after I posted what I did, I thought more about it and I am
not sure how far I can go to defend it. In principle I think I was
right, but I can't think of a non-contrived test case. If I have made
code that then generates a page, and that generated page does not pass
validation, I would worry that some other bugs are in the generator.
Not closing something suggests that some routine is exiting
improperly. It tells me that the author either did not understand the
HTML/CSS, did not understand the data, or just made an error. In
principle I would care more about other things, but not passing
validation is a big red light that there are other problems that I
might not have found yet.

[snip]
--
Matt Silberstein
And now our bodies are oh so close and tight
It never felt so good, it never felt so right
And we're glowing like the metal on the edge of a knife
C'mon! Hold on tight!
C'mon! Hold on tight!

Though it's cold and lonley in the deep dark night
I can see paradise by the dashboard light
Paradise by the dashboard light

Jim Steinman

[Spartanicus]

Henri Sivonen <hs******@iki.f i> wrote:

Petr Nalevka has developed (building on James Clark's work) RELAX NG and
Schematron schemas that express more constraints than the HTML and XHTML
DTDs. His validator is available at http://badame.vse.cz/validator/

As I and many others here tell people who call CSE a validator, there is
only one requirement for a document to be valid: it must comply with the
requirements as set out in the document's DTD.

Although fundamentally different from CSE,
http://badame.vse.cz/validator/ is therefore equally not a validator and
shouldn't be labeled or referred to as such.

--
Spartanicus

[Henri Sivonen]

In article
<92************ *************** *****@news.spar tanicus.utvinte rnet.ie>,
Spartanicus <in*****@invali d.invalid> wrote:

there is only one requirement for a document to be valid: it must comply with the
requirements as set out in the document's DTD. .... http://badame.vse.cz/validator/ is therefore equally not a validator and
shouldn't be labeled or referred to as such.

"Valid" in the SGML and XML specifications is indeed defined to mean the
condition that the document conforms to the DTD it declares for itself.

However, the RELAX NG and Schematron specifications use the word "valid"
to describe the condition that a document conforms to a RELAX NG or
Schematron schema, respectively. http://badame.vse.cz/validator/ and
http://hsivonen.iki.fi/validator/ are validators in the RELAX NG and
Schematron senses of the word.

--
Henri Sivonen
hs******@iki.fi
http://hsivonen.iki.fi/
Mozilla Web Author FAQ: http://mozilla.org/docs/web-developer/faq.html

[Jan Roland Eriksson]

On Wed, 17 Aug 2005 00:04:15 +0300, Henri Sivonen <hs******@iki.f i>
wrote:

In article
<92*********** *************** ******@news.spa rtanicus.utvint ernet.ie>,
Spartanicus <in*****@invali d.invalid> wrote: [...]

"Valid" in the SGML and XML specifications is indeed defined to mean the
condition that the document conforms to the DTD it declares for itself.

However, the RELAX NG and Schematron specifications use the word "valid"
to describe the condition that a document conforms to a RELAX NG or
Schematron schema, respectively. http://badame.vse.cz/validator/ and
http://hsivonen.iki.fi/validator/ are validators in the RELAX NG and
Schematron senses of the word.

Then stop using the words "valid" and "validator" in relation to those
products, it would help to avoid confusion for many.

--
Rex

[Spartanicus]

Henri Sivonen <hs******@iki.f i> wrote:

there is only one requirement for a document to be valid: it must comply with the
requirements as set out in the document's DTD.

...

http://badame.vse.cz/validator/ is therefore equally not a validator and
shouldn't be labeled or referred to as such.

"Valid" in the SGML and XML specifications is indeed defined to mean the
condition that the document conforms to the DTD it declares for itself.

However, the RELAX NG and Schematron specifications use the word "valid"
to describe the condition that a document conforms to a RELAX NG or
Schematron schema, respectively. http://badame.vse.cz/validator/ and
http://hsivonen.iki.fi/validator/ are validators in the RELAX NG and
Schematron senses of the word.

But http://badame.vse.cz/validator/ is incorrect to call itself a
*HTML* validator right?

(it's actually called *the* HTML validator, but lets be kind and put
that down to their rather limited command of English)

--
Spartanicus

[Chris Beall]

Scott Meyers wrote:

I have a web site that, due to maintenance by several people, some of whom are
fairly clueless about HTML and CSS, etc. (notably me), has gotten to the point
where I'm pretty sure it's suffering from bit rot.

Scott,

I can confirm your suspicion. I viewed your site with Netscape 7.1 on a
600 X 800 display with the browser at full-screen size and Text Zoom set
to 100% (original size). Looking at the navigation on the left side of
the screen, the last item visible to me was "Articles & Interviews", the
last word of which was clipped at the bottom of the window. There was
no way to scroll this area of the page. Only by using the browser to
reduce the text size was I able to see the next three links. The bit
rot has become visible from the outside...

Chris Beall

[Scott Meyers]

On Wed, 17 Aug 2005 01:03:16 GMT, Chris Beall wrote:

I can confirm your suspicion. I viewed your site with Netscape 7.1 on a
600 X 800 display with the browser at full-screen size and Text Zoom set
to 100% (original size). Looking at the navigation on the left side of
the screen, the last item visible to me was "Articles & Interviews", the
last word of which was clipped at the bottom of the window. There was
no way to scroll this area of the page. Only by using the browser to
reduce the text size was I able to see the next three links. The bit
rot has become visible from the outside...

Actually, this isn't bit rot, it's an implementation error that nobody
brough to my attention for five years. The nav frame was supposed to be
set up to sprout a scroll bar if it was longer than the available screen
real estate, but apparently that never got implemented. My guess is that
nobody noticed, because nobody in my world runs at 800x600 or less. That's
not meant to be an excuse, just an explanation.

My concern about bit rot has more to do with the mixture of <p> and <br>
and CSS and empty table rows and who-knows-what to achieve display goals,
e.g., desired vertical spacing between elements, font changes, etc. My
feeling is that all that stuff should be cleaned up and made consistent,
and as long as we're degunking the thing, we should take the opportunity to
address some likely design errors (e.g., use of frames), implementation
mistakes (such as the one you mentioned), and generally spiff things up.
Also maybe shoot for code that validates.

All I have to do is find somebody to do the work :-)

Scott

[Henri Sivonen]

In article <ig************ *************** *****@4ax.com>,
Jan Roland Eriksson <jr****@newsguy .com> wrote:

On Wed, 17 Aug 2005 00:04:15 +0300, Henri Sivonen <hs******@iki.f i>
wrote:

However, the RELAX NG and Schematron specifications use the word "valid"
to describe the condition that a document conforms to a RELAX NG or
Schematron schema, respectively. http://badame.vse.cz/validator/ and
http://hsivonen.iki.fi/validator/ are validators in the RELAX NG and
Schematron senses of the word.

Then stop using the words "valid" and "validator" in relation to those
products, it would help to avoid confusion for many.

Have you already contacted James Clark and complained about the use of
the words "valid" and "validator" in relation to the RELAX NG spec and
Jing? Or the ISO committees that accepted RELAX NG and Schematron?

--
Henri Sivonen
hs******@iki.fi
http://hsivonen.iki.fi/
Mozilla Web Author FAQ: http://mozilla.org/docs/web-developer/faq.html