From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Mon Jun 16 2003 - 04:39:10 EDT
From: "Doug Ewell" <dewell@adelphia.net>
> Philippe Verdy <verdy_p at wanadoo dot fr> wrote:
> > It's a great news. It will force websites to stop using Microsoft
> > specific features and caveats, and adopt the real standards.
> > ...
> > If web sites start using the real standards, people will upgrade for
> > a more standard browser, and Microsoft will consider correcting
> > its IE to follow the evolution of websites, removing the unnecessary
> > features that Microsoft wants to promote, such as proprietary web
> > fonts for CSS2, or the Office "smart links".
> I'm sure it is also true, as Philippe states, that IE supports
> proprietary features that aren't in the HTML specifications published by
> W3C. (Netscape was famous for this in the early days with the <blink>
> tag.) To the extent IE supports tags and features and behavior that are
> not in the HTML DTDs, and hence are not legal HTML, one can say that IE
> has not fully adopted, or is not fully conformant with, the "real
> standards."
I did not spoke about Netscape or AOL that also are using proprietary
or non-conforming "features". I could speak about Apple too, that also
uses proprietary fonts, or incoherent decompositions of Unicode characters
in its VFS driver for HFS+ filesystem.
Microsoft is not alone, and my intent is not anti-someone, but I applaude
when Microsoft will no longer try to impose its view (and I won't call it
standard) on what a browser can or should be, because its too huge
installed base of IE creates a de-facto standard that ruins the efforts
done to standardize the web, and forces other browsers and web site
authors to adopt its quirks just to emulate IE.
The problem is that when there's a de-facto monopole with no real
alternative, the existing standard gets threatened, and as long as
MS does not correct its browser, no other browser will correct theirs.
Then if MS corrects its browser, all web sites need to be updated to
correct their content so that they display correctly (for example see
the many quirks that was introduced in websites displayed with IE6,
and the big difficulties encountered to support cross-browser CSS
support, even between IE5.5 and IE6 which do not handle the
attributes inheritance the same way.
Do to this situation, websites are considered bogous instead of
browsers used to visit them. I have the same bad feeling with
Netscape 4.x bogous CSS support, or bogous Unicode support,
something that was corrected only in the excellent Gecko engine
of Mozilla and Netscape6.
Conformance to an existing standard and anticipation to its
evolution or corrections is something that developers of browsers
should consider as bugs in their browsers, instead of in web sites,
simply because it is much more complicate for content authors to
have to test and patch their content face to bugs or quirks in some
browsers.
The basic requirement is that the documents displayed
must be rendered at least correctly,even if the exact layout is not
supported because of font differences. This requires that browsers
correctly interpret at least any text which is correctly labeled
and encoded, and then any correct markup (some of which should
be ignored but should not forbid the rest of the markup to be rendered
correctly.)
> I guess my problem is that Philippe makes it sound as though IE doesn't
> support HTML at all, but rather some proprietary HTML-like
> Microsoft-specific markup language;
That's an excessive interpolation. Don't assume things I did not say.
In fact I admit that it's difficult to implement a standard, but this should be
done by coherent steps, so that conformance to a given version is required
before trying to implement and interpret the next level of the standard.
The proprietary aspect of browsers is the fact that they often do not follow
an evolution that respects strictly the path indicated in conformance
requirements of the standards they wish to support. This situation will
continue for several years until there's a real conforming XML-DOM engine
that passes the conformance levels needed to parse contents without
introducing "quirks", that browsers often want to maintain instead of
correcting them.
This was true for Netscape4.x, and it is still true for IE, and
for other browsers like Opera that want to emulate IE, or for AOL's
proprietary "features" that were added to IE on Windows, and that
will now be added on top of Safari or Netscape on Mac OS; on the
opposite, the Mozilla's Gecko engine is built considering that any deviance
is not a quirk to support but a bug that needs to be corrected.
We know what this means now for AOL: it looses customers because it
maintains its proprietary features in its mandatory connection software.
This is true also for cable access providers (which were often in
situation of monopole for the cable access until DSL started to explode
and threaten their monopole: they have started to remove their exclusive
alliance as AOL distributors, offering alternative with their own connection
kit and support, with total freedom of choice of the browser, and of the
connection hardware, that can use the same set of software and hardware
add-ons that work now with DSL, with the only exception of the modem,
that they now offer with their subscription)
Thanks to the standards, they mean more customer choice, more
competition, better services, and more technical alternatives for
Internet access, and reduced prices on basic features.
This discussion is really going too much out of topic of Unicode,
even though it speaks about what standards should and should not
be: a standard is not the adoption of a technology supported by a
unique vendor, even if it's in a near monopole situation. All
monopoles need (and should) be complemented by alternative
technologies without impacting heavy costs for those that wish to
make this step. Open standards are there to ease and accelerate
the adoption of newer technologies with lower development costs
and better ROI, not to create new constraints.
This archive was generated by hypermail 2.1.5 : Mon Jun 16 2003 - 05:31:10 EDT