WebAIM - Web Accessibility In Mind

E-mail List Archives

Real Validation [was] Re: Do doctype declarations change browser interpretations?

for

From: Holly Marie
Date: Mar 16, 2002 2:47PM


----- Original Message -----
From: "Terence de Giere"
Subject: Do doctype declarations change browser interpretations?


| One point to follow with Doctypes - I have noticed some developers
will
| put in a doctype declaration, but do not test the resulting page with
an
| SGML parser to check that the HTML conforms to the specification
| referred to by the doctype declaration. .

While at it, we should let people know the differences there are out
there regarding True validation and not.
Some sites that check pagework (what some people think is validation)
are really *lints*, etcetera - True validators use
SGML and check the code more strictly for parse errors or correct
syntax.

I believe there are only 3 that are widely known. This was brought up a
year or two ago on a web design and developent list. And the information
was easy to recall because one tool is called the "real validator"./

True Validators [use SGML]
[1] http://validator.w3.org
[2] http://www.htmlhelp.com/
[3] http://arealvalidator.com/

Bobby is not really considered a true validator.
A *very* good and short summary on the differences between SGML
validators versus Lints or HTML page checkers, is at the Real Validator
Web site, on this page - http://arealvalidator.com/real-validation.html

LINTS
Examples of lints - these may report false errors, and also may on the
other hand give other ideas and more information that True validators
might miss. What I believe, is BOTH forms of page checkers are needed.
Extra information that might be given, are messages to check for alt
descriptions and the content of these, color contrast, page load time,
file sizes, link popularity, meta tag checking, broken links, etc... and
other hints about a web page that may be overlooked by the Real
Validators. However, when it comes to clean code it is important to
check the code in the Real validator and use the lints for extra hints,
ideas or checks. Be aware some of the error reporting may be mistakes.

Lints more are listed on the differences page listed in the previous
paragraph, these are common ones:
CSE HTML Validator - http://www.htmlvalidator.com/
HTML TIDY - http://www.w3.org/People/Raggett/tidy/
Bobby - http://www.cast.org/bobby/
Weblint - http://www.weblint.org/
Web Site Garage - http://websitegarage.netscape.com/
NetMechanic HTML Check - http://www.netmechanic.com/
[these links were taken from the real validator pages, and all were not
checked]

| A number of graphical HTML editors put Doctype declarations in a page,
| but do not produce HTML to conform to it.

Without being a snob or sounding like an idiot here. I have yet to find
a user friendly Graphical editor that will put out very clean code, much
at all, or without the operator at the keyboard knowing what to fix.
Just because an Editor, either WYSIWYG or Text web editor, offers up ALT
attrribute boxes or areas. The text editor I use even gives warnings on
a lot.


I have not seen one impress upon the user to:
[1] fill in this box now
[2] fill in this box with proper descriptive text.



I guess this could go for captions on all sorts of media, too.
THERE ARE very few editors, either type that will write valid code to
xhtml standards. They are all still catching up. I think HS and a couple
of others have XHTML standards built in now.

Even then, people have to know how to:
[1] use the inside validation, BUT also check at a True validator like
W3.org before committing.
[2] know how to fix the mistakes, that tools can make
[3] be responsible enough to check it all, anyway, and then even
more{color blind checks, content, usability, navigation, etc]

One wonders how this Valid code area might be enhanced?
Maybe, a page cannot be saved to .html or other format unless it
[1] First validates to the proper DTD in the document at the W3.org???
[2] Can only be written in HTML 4 Strict or even XHTML strict or even
newer
[strict mode generally trashes all the deprecated items and
presentational items from the page code]

Interesting thoughts, maybe pages could not be saved in the future if
the ALTS on images are not filled in and all the valid criteria are
met,. however, I doubt that will ever happen. But it might be a neat
thing to make tools that will. Maybe as 508 and guidelines become even
more important, tools like these have a place in the market for sale.

Still many things can only be approached with checks by human factor.


| Some even get the Doctype
| declaration wrong - certain parts of a doctype declaration are case
| sensitive, so it should be used exactly as specified, as changing the
| case will make it impossible for the parser to select the correct DTD
| for checking. I have found that HTML coded to specification generally
| displays more consistently across more browsers, especially
| non-Netscape/Microsoft browsers, but this does not mean the page will
| look exactly the same in all these browsers.

*** ALSO back to a DTD***
Not using a doctype or DTD or one of the older ones in many browsers
will trigger the old browser display of a web page. Called DTD bug mode
by Mozilla, NN, and also Internet Explorer.
Transiitonal may trigger this in some of the newer browsers, but no DTD
will trigger bug modes for sure.
But modes refers to displaying the pages with old buggy non compliant or
up to date browsers. Though many are not quite really up to date just
yet, anyway. But getting better. I would have to look for the direct
links to these items for each browser, but the information is available.
Even MS has a page on the information for it IE6.

holly



----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/