RE: MarkUp Validation Service Problem

> -----Original Message-----
> From: www-validator-request@w3.org 
> [mailto:www-validator-request@w3.org] On Behalf Of Jukka K. Korpela
> Sent: 26 mai, 2003 18:23
> To: www-validator@w3.org
> Subject: RE: MarkUp Validation Service Problem

> On Mon, 26 May 2003, Denis Boudreau [ CYBERcodeur.net ] wrote:
 
> > With all due respect, I have no idea how to read a DTD either and I 
> > frankly don't care about reading it, just like I don't need to know 
> > how to break down a motor to ride my car.
 
> As a consequence, you inevitably get confused with 
> validators' messages.

Maybe the first few times around, but as years passed by i got to
understand them just as well as anybody else. The way i see it, it
didn't take me more than a few tries to understand the underlying
process associated with validation -- therefore really quickly i
understood what needed to be done to correct my errors. As time flew by,
the number of my errors lessened and my comprehension of the error
message grew. Now, i only rarely make mistakes, but i always know where
to look for them. And i still have to see a DTD with my own two eyes.
 
>> Why wouldn't validating be done just for it's own sake, because it's 
>> the *right* thing to do, because it ensures better operability and 
>> accessibility among other things ?

> Validation is a minor tool for checking your documents 
> against some rules that you know, and rather confusing if you 
> don't know the rules. It has rather small, though not 
> negligible, impact on operability and accessibility.

The only thing that really matters is actually reading the specs to see
what can and can't be done. Everything else is only a matter of personal
opinion, no matter how strong one feels about them. Quite amusingly, the
number of my errors lessened as soon as i started reading the
instruction manuals provided by the W3C... along with a few errors here
and there, i have grown quite capable of writing valid (x)html from
scratch... like anyone who really wants to do that. Understanding how a
DTD works really has nothing to do with it.
 
>> It's not personnal either, but instead of asking that 
>> question every so often, you should maybe ask yourself why you 
>> ask it in the first place

> That's a good question. Maybe I'm just too worried about 
> other people's problems when they get cryptic validation 
> messages that are mostly irrelevant - often indicating 
> something that should, in principle at least, be fixed, but 
> little indication of what the problem really is.

Only one thing can settle this problem; experience. Experience with
html, experience with the undelying rules associated with the language.
Experience also taken from answers on mailing lists from people like us
who can provide respectful answers to questions asked by less
experienced html authors who simply want to understand (while maybe
being without a single clue as to where to look for help). 

Remember that HTML, as opposed to other more complex web languages like,
say, xml for example, is not meant to be overly complicated. If html is
to remain the language of the people, one has to be able to pick it up
and code almost immediately. Asking them to learn how to use DTD might
be a good thing because it'S always better to know more than less, but
it is in no way mandatory -- neither should it ever be.

While i agree with you that xml is not overly complicated either it *is*
a lot less usable for the common html author than html is or will ever
be. I'm pretty sure that if I picked up a DTD i would understand it
quite fast. However, i see no reason to do so and therefore, am not
doing it. However, that doesn't mean i produce evil web documents.
 
> Actually I'm more worried about the phenomenon that people 
> _believe in_ validation without understanding what it is - 
> and ignore far more important issues in Web site design, when 
> they think that their pages are "valid" (i.e., everything OK, 
> cross-browser operable, accessible, standard, and whatever) 
> when they pass validation.
 
Issues like what ? Validation should be the end that justifies the means
for all web authors. If any, it should be the only thing that really
matters. If everybody cared about creatring valid documents, it would be
sufficient for the web to become a better place for everyone. While
validation may be trivial by itself, it opens up a world of
opportunities for authors to become better at coding. And that is
priceless.

> In this case, for example, it is completely irrelevant 
> whether the border attribute in <input> is valid (under some 
> DTD). Validation was useful only to the extent that it raised 
> the question _why_ the attribute is not in the specification 
> you validated against. It is pointless to ask _whether_ it is 
> there, since it can be checked, and if you don't check it 
> yourself, then someone else has to do that _or_ give you a 
> mere guess, and you still don't know whether the answer is 
> true. Actually, although validators may have errors, you can 
> be pretty sure that it reads the DTD properly and gives a 
> technically correct report - so sure that the only way to get 
> more certainty is to look at the DTD.
 
But do you really need to know why or how ? As purists, we might full
the urge to understand, but the average web author should have no
interest in such thing. As long as they understand the interests of
validation, what it can bring to a web site, everything else is trivial
at best.

> The question _why_  it isn't there naturally goes beyond the 
> scope of validation. On the practical side, most of the HTML 
> constructs that violate HTML 4.01 Transitional syntax should 
> be fixed. But such fixes are generally less productive than 
> people think. For example, it's easy to violate that syntax 
> by using tag soup with <font> tags here and there, but it 
> would be much more productive to remove the <font> tags (or 
> to rewrite the markup completely) than to make them comply 
> with the formal syntax.
 
Yes it goes beyond the scope of validation and it is perfect that way.
In my opinion, as long as a web author takes the time to read the specs,
they will know all they need to know about writing valid code.
Thereafter, the validation tool is only there to help clean up the mess
when there's a mess. I'll even go further than that and say that if the
validator was even more powerful, it would be able to provide contextual
correction tips and reasons that are easily understandable, making the
process even more easy for neophytes who feel like tackling the html
beast. Only then will validating really be something that every author
can add to their toolbox.

>> The only thing that matters is producing valid pages because we 
>> feel it's important, because we understand what the W3C tries to 
>> achive and we want to be a part of it all :)

> The importance of validation is grossly overrated and the 
> need to _understand_ its essence (if you use it) is seriously 
> underrated. Oh well, I think I need to mention the long rant 
> on this: http://www.cs.tut.fi/~jkorpela/html/validation.html

Your rant is interesting, but again, it fits exactly in the kind of
mindset that you have, which does not reflect the average user's in the
least. Again, why should an author want to understand ? I don't need to
understand how a phone works to use it like a pro. I believe you are way
too caught up in a programmer's vision of the whole thing. If you start
looking at the big picture from the averag user's point of view, you
would see that it matters not at all. It does matter for people like you
who would help build the tools, but for people like us who only care
about using an efficient tool it bears absolutely no  interest.

Denis Boudreau
CYBERcodeur.net

Received on Monday, 26 May 2003 20:38:44 UTC