[thelist] Validation > pressure on software companies??

Steve Lewis slewis at macrovista.net
Fri Jun 14 14:54:03 CDT 2002

Kostal Design|Sebastian Kostal wrote:

>>>I've seen valid code that breaks on NN 4.5 Mac. So the validator seems
>>>to me no measure to assure correct page display.
>>As it shouldn't be. If that was happening then it would be the html standard
>>following browser foibles instead of browsers trying to follow the standard.
> I have mixed feelings about that. Netscape 4.5 is still a broadly used
> browser and one of the browsers I always test on.

Lachlan is spot-on here.  If you want to 'validate' correct page display
for some old buggy browser (NN 4.5 Mac) you will need to use your eyes.
   If you want to catch all your table data cells that do not have a
closing tag, you should use a validator instead of your eyes because the
validator will tell you where in the output the error appears, and will
always find it.  Some browsers do not have problems when this tag is
missing so your eyes will not necessarily catch it.

> important as a foundation. On top of this you can build non standards
> features that wont show, if they are not suported or not incorporated.
Customizing content to the requesting user agent string is a bad thing.
a) you must maintain duplicate versions of your content with only subtle
formatting differences
b) user agent strings are easy for users to change
c) user agent strings are not forward-looking so there are going to be
arcane magics you will need to perform over your site when a new browser
comes out--plus regression testing!
d) degratability has been a goal in this thread.  consider the number of
hours you will spend making these numerous browser-dependant versions of
the page all degrade nicely.
e) invariably if you are using these browser-specific tricks, and/or
trying to achieve pixel-precision, you need to make yet one more version
  of your page that is printer friendly.

> Still, a site that does not work on a relatively recent browser is not a
> good design, even if it validates, because the goal is not to meet a
> standard in the end, but serve people a website that functions.
Who claimed that validation was the end-all-be-all of design or
implementation of a website?  While I cannot argue with this statement,
I don't recall anyone here claimimg anything to the contrary.  Microsoft
Word, with spelling and grammer checkers on, will still give it's seal
of approval to some rather poorly constructed language.  There are some
well known examples passed around in education circles to show students
the value of learning to spell and use correct grammer on their own.
You still need to have some skill at working with your language to
communicate clearly, because spelling/grammer validation does not
guarantee correctness of the product, just correct application of
structures and a reduction in the number of typos.

Validation is a tool, nothing more nothing less.  If you want to create
a horrible websit that validates, you can.  If you want to create an
elegant degradable website that validates, is printer-friendly, and
conveys all of your content clearly in a variety of browsers and
platforms (which IMO should be the goal) you can.  If you apply
validation arfully you can achieve some very impressive designs.


More information about the thelist mailing list