Re: Are standards a bad or good thing?

>The problem as I see it in narrowly limiting validation to only those times
>when one has a formal definition (I beleive that is provided by a DTD) is
>that I don't beleive that the major browsers HAVE formal DTD's. (Supposed
to,
>Should and Ought to however). I looked at the DTD's recognized by the W3
>validator and don't see Netscape or MSIE 4.0 listed there.


Well, technically a UA doesn't HAVE a DTD, but interprets BASED ON a DTD.
However, the DTD they interpret by is usually not the same as the one
provided ina HTML standard.

The DTD's the major browsers obey are supersets of the HTML 3.2 DTD (let us
call them the "messy DTDs"). Technically, the browser should determine which
DTD to obey based on the !DOCTYPE element. Most browsers ignore the !DOCTYPE
and obey one static DTD (their own messy DTD), which incorporates the latest
official DTDs.

There is of course nothing wrong with this approach, as long as the
browser's messy DTD fully incorporates the most recent HTML DTD and obeys
the rules about unrecognised elements or attributes; it's when the messy DTD
incorporates changes to the official DTD (like Netscape's and IE's to a
lesser extent) that you run into trouble.

>A page might conform to the DTD's that are published, but then blow dead
>bunny's out of it's exhaust pipe when run through the software that a
browser
>uses. I would guess that there is a major disjunction between the DTD's and
>what happens in reality (though the gap is narrowing)

That is often the case with (especially with Navigator, but also with IE).
However, given a document that conforms to a W3C-endorsed DTD, that is the
mistake of the UA, not of the author of the document.

There is also another issue to consider; a UA is supposed to ignore
unrecognised elements and attributes. Given the use of elements and
attributes from a newer DTD than supported by a given UA, the UA might
render a document differently than the author expects (still, however,
conforming to the correct rules). The author might experience this as a flaw
in the UA, when in fact it is not. You must be careful not to confuse these
issues.

>So, for a web author, what is ''valid'', if one does not have a reliable
DTD
>to validate againt for the popular software that will show our page written
>in HTML?


Again, you must make a distinction between actual mistake, as in deviance
from the standard HTML, and ignoring of unrecognised elements or attributes.
You cannot expect older UAs to correctly render newer contructs and authors
should be prepared for this happening.

You have a guarantee when using official HTML that your document will remain
renderable and readable in older UAs, possibly at the cost of some fancier
markup or stylistic decor. That is still within the bounds of the goals of
HTML. You can't expect your document to "look" the same everywhere and you
certainly can't make the "look" of your documents on your own screen the
standard for "validation" (slight abuse of the term here).

>Well, one loads the page in question into a browser. I don't know of any
>other way to validate or derive validation without a formal DTD.
>
>Or am I speaking of which I do not know and should just shut up?


I think you are expecting more than HTML can offer. HTML offers smooth
readability and accessibility across platforms through the use of the
abstract UA mechanism. HTML was never intended to port a "look and feel"
from one place to another. Certainly not stylistic markup or exact details
of the rendition of a document. All you can expect from HTML is that your
document, if written in correct HTML, will be rendered in an acceptable,m
accessible, readable fashion on every supporting UA. What you are looking
for with your browser-tests is a way to transport your SCREEN from your
monitor to the next. HTML is not meant for that; if you want to do that, I
suggest something like Exceed and direct access to your machine for
everybody.

Ben Z. Tels
optimusb(at)stack.nl
http://www.stack.nl/~optimusb/
UIN:2474460

"The Earth is the cradle of the mind, but one cannot stay in the cradle
forever."
                                        --Tsiolkovsky

HWG hwg-theory mailing list archives, maintained by Webmasters @ IWA