[thelist] Validation > pressure on software companies??

[James Aylard]

Sebastian,

> I personally clearly prefer a page that works and depreciates nicely over
a
> page that sticks to some standard which causes display errors. Please help
> me to understand. I feel really funky, because everybody makes such a big
> deal about validation and then in some instances the page does not even
> display correctly! What's the point then??

    Obviously, there is not much value to a page that validates but will not
display in the browser. In that sense, there is nothing magical about
validation. However, validation can be useful in locating structural errors
in the HTML (such as a missing closing tag or a malformed table) or in the
CSS (such as a missing semicolon or closing curly brace), depending on which
validator you are using.
    Valid HTML and (especially) CSS are more likely to display, and to
display consistently, in more browsers than markup which is not valid.
Again, no guarantees here as no browser supports every standard fully, and
not all browsers interpret every standard in the same way. But if it's
valid, your chances are generally better.
    There's also the sense that valid HTML and CSS are more "forward
compatible". While this is hard to predict, and probably not as critical as
some suggest, the tightening of standards-compliance in IE 6's
standards-compliance mode does give some weight to this argument (for
instance, border-width: 6 ; won't do any more in IE 6's standards-compliance
mode).
    And lastly, writing valid HTML and CSS are in some ways the mark of a
professional: someone who knows the standards, first of all, and someone who
exercises the discipline to actually implement them.
    As someone who readily argues that, in web design and development,
virtually all rules have exceptions, I am not very comfortable when the
position in favor of standards compliance takes on the air of a theological
argument. While I typically code to the standards, I also use proprietary
elements, attributes, and properties at times, and I don't apologize for it.
But I usually know which rule I'm breaking when I do, and hopefully have a
decent concept of the potential pitfalls of doing so. And that, I think, is
a pretty reasonable approach.

James Aylard