Reliability of validation diagnostics
Robbo - W3C Rocks!
I submitted my "XHTML 1.0 Stict" page [ www.tesol-direct.com ] for
validation and got conflicting reports.
--- validator.3c.org says:
"This Page Is Valid XHTML 1.0 Strict!"
--- htmlvalidator.com says:
"CSE HTML Validator Lite generated 0 error messages and 6 warning
messages when checking this web page. These problems could damage this
web page's search engine rankings as well as cause viewing problems
for visitors. It is highly recommended that any problems be
corrected."
The "mistake" is that I twice used three unescaped chevrons to
highlight two menu options:
>>> Contact us
and
>>> Enrol now!
Which should I rely on? Will this "mistake" really "...damage this
web page's search engine rankings as well as cause viewing problems
for visitors"?
Vince Robbins
webado
But if even one of those chevrons had been the other way, like < ,
then you'd have had not only an error but missed some text.
I will venture to say that the validator looks for an opening marker
first and then hunts and pecks for the matching closing one. Having
the closing one only doesn't register as a tag marker.
Is this OK? Probably not. How bad? hard to tell.
I think this must be the reason why it's required or recommended to
escape special characters like this, to avoid any possible confusion
with markers that are part of html markup.
On Mar 10, 6:36 pm, "Robbo - W3C Rocks!"
Albert Wiersch
> The "mistake" is that I twice used three unescaped chevrons to
> highlight two menu options:
>
> >>> Contact us
> and
> >>> Enrol now!
>
> Which should I rely on? Will this "mistake" really "...damage this
> web page's search engine rankings as well as cause viewing problems
> for visitors"?
CSE HTML Validator tries to promote better style than strict
compliance and recommends that those characters be escaped.
--
Albert Wiersch
AI Internet Solutions
sup...@htmlvalidator.com
http://www.htmlvalidator.com/
Duncan Hill
code, for starters the code is not an area that should be semantically
used for any styling, that is purely CSS and/or any scripting you wish to
place on top.
The '<' and '>' symbols are standard international quotation marks.
You are continuing the Myth of Microsoft's policy of proprietary code and
standards for the web.
By working to strict standards, in creation and checking of the code
beneath a webpage, we can provide a more stable platform to display the
content to the widest possible audience. For a standard to be effective it
must be tightly defined and adhered to. Arbitrary decisions about coding
principles are not the remit of a code validator.
Leave the decisions about where and when to 'break' the code in the hands
of the author.
Duncan
Albert Wiersch
On Mar 11, 10:23 am, "Duncan Hill" <dhad...@mndhill.com> wrote:
> That surely negates the use of your software for viable checking of the
> code, for starters the code is not an area that should be semantically
> used for any styling, that is purely CSS and/or any scripting you wish to
> place on top.
before, most people are more interested in practical matters. Real-
world browsers are far from standards compliant. If they were strictly
compliant, then the vast majority of pages would never render because
there would be errors. We live in an imperfect world, not the perfect
world of perfect standards compliance.
Now, that's not to say that CSE HTML Validator won't help. Much of
what it checks for IS based on the standards. But it goes beyond the
standards based on practical matters and real-word browsers and tries
not to bring up issues that are not worth time addressing. But it is
very configurable so many of these issues can be adjusted depending on
what the author choosing to do.
For those who only care about being strictly standards compliant then
the W3C should suffice. However, CSE HTML Validator would also be
helpful because one can combine the output of a DTD validator with CSE
HTML Validator's own results if they simply check an option. Or they
can only use the DTD based validator if they so chose, but that would
defeat much of the purpose for using and usefulness of CSE HTML
Validator.
Duncan Hill
more standards compliance to the web. Even Microsoft who have succeeded in
holding things back for so long by the use of proprietary code and poor
adherence to W3C standards are forging ahead with improvements to IE8.
Nobody expects it to be perfect yet, but it now has the capability to meet
much of the standards in operation.
Of what value is any piece of software that cannot naturally recognise
those standards when it is declared to be validation software.
An experienced author would probably understand having to set the
validation to the level (s)he wants, but what happens at novice and even
intermediate levels where the user wants guidance from the outset to meet
standards.
It is a hard enough task to educate people to the need and advantage of
coding to standards without forcing them to learn the ins and outs of
configuring a validator so that it will skew the report in favour of their
code.
That isn't what validation is all about, it is ensuring that the code is
in compliance with the set and applied standards.
You could maybe class CSE Validator as a code review program, it would
serve that purpose well, but please if is a validator, make it adhere to
valid standards out of the box.
The reason that many pages would not render in a fully compliant browser
is purely because they are not using standards compliant code, is this
something that we must perpetuate, or should we perhaps move forward on
the web and work from and to acceptable standards.
None of the current top level browsers, other than the IE family will
break a standards compliant valid page at any level from transtional to
strict DTD in HTML or XHTML. Coded semantically the pages will also give
better rendering and more accurate styling with CSS, and last but not
least, will make perfect sense to text only browsers and assistive
technology.
Duncan
On Tue, 11 Mar 2008 15:45:20 -0000, Albert Wiersch <goo...@wiersch.com>
wrote:
Duncan Hill
Latest public release of Firefox (Windows, Mac, Linux)
with the following extensions installed:
Chris Pederick's Web Developer Toolbar (All platforms)
HTML Validator by Marc Gueury (All platforms)
Firebug by Joe Hewitt (All platforms)
they are all free and well supported, as a combination or individually
they will help keep your coding on track as you develop your website.
It is seen as good practice to test and develop using a browser like
Firefox and then pick up the bits that straggle about in IE. On a
standards based page, even IE6 is capable of not messing things up too
much.
Best wishes
Duncan
Albert Wiersch
On Mar 11, 2:28 pm, "Duncan Hill" <dhad...@mndhill.com> wrote:
> ALL of the major browser producers are working daily to bring more and
> more standards compliance to the web. Even Microsoft who have succeeded in
> holding things back for so long by the use of proprietary code and poor
> adherence to W3C standards are forging ahead with improvements to IE8.
> Nobody expects it to be perfect yet, but it now has the capability to meet
> much of the standards in operation.
reasonable to do so.
> Of what value is any piece of software that cannot naturally recognise
> those standards when it is declared to be validation software.
something is used that is not standard, often CSE HTML Validator will
point it out and recommend that the user use a more standards
compliant way of doing things. Also, it will point out deprecated
constructs and let the author decide what to do. It doesn't simply let
the significant non-standards compliant issues go by without some type
of notice.
> An experienced author would probably understand having to set the
> validation to the level (s)he wants, but what happens at novice and even
> intermediate levels where the user wants guidance from the outset to meet
> standards.
problems in their page and doesn't care about strict compliance and
doesn't want to be limited to that. However, like mentioned before,
CSE HTML Validator will often alert them and inform them of many non-
standard issues.
> You could maybe class CSE Validator as a code review program, it would
> serve that purpose well, but please if is a validator, make it adhere to
> valid standards out of the box.
Validator because that is not what most people want. They want
checking that is more thorough and not limited to DTD based validator
checking. A DTD based validator is much more limited in what it checks
for.
> The reason that many pages would not render in a fully compliant browser
> is purely because they are not using standards compliant code, is this
> something that we must perpetuate, or should we perhaps move forward on
> the web and work from and to acceptable standards.
greatly in that, making suggestions about using standards compliant
code and pointing out many non-standards compliant issues. It doesn't
simply let the significant non-standards compliant issues go by
without some type of notice.
Duncan Hill
be locked into the MS IE6 way of thinking about the web, yet even MS have
thrown that one out the window now (yeah! the pun is that obvious, but I
couldn't reist).
I don't deny that your software has some excellent points, I have used the
Lite version, but I still think you are handling validation of code as
though you are measuring something with a piece of string, and we all know
how long a piece of string is!
Be honest with people, call it an 'Evaluator', but unless you give default
standards compliant checking out of the box, the term 'Validator' is
misleading, because you have already made the decisions on what to accept
and what not to accept as permissible deviations from the standards.
Many things around us operate on standards, why is web-coding any
different if you want it to work everywhere as intended.
Duncan
On Tue, 11 Mar 2008 20:29:08 -0000, Albert Wiersch <goo...@wiersch.com>
wrote:
>
>
--
Duncan Hill
(DHadmin)
Robbo - W3C Rocks!
I have replaced the [>] with [>] and now both validators are happy.
Robbo