Posted by John Bokma on 04/12/06 01:32

"Alan J. Flavell" <flavell@physics.gla.ac.uk> wrote:

> On Tue, 11 Apr 2006, John Bokma wrote:

>> Lets also not forget that MS is one (if not the only one) that
>> supports outdated OSes as long as they do. By the time Vista is
>> released XP is over 5 years old.
>
> You seem to be tacitly assuming the truth of the testimony that MSIE
> is an operating system component. I would rather have a WWW browser,
> leaving the OS component to be used for applying the vendor's own
> fixes.

I consider making an application work on an outdated OS supporting that
OS, nothing more, nothing less.

>> Yes, MS did quite some things wrong, but other browsers have flaws
>> too or keep flaws in existence. Ditto for the holy "standards
>> organization" W3C.
>
> Oh, sure. But deliberate violation of a mandatory requirement of an
> IETF standards-track protocol (RFC2616) can hardly be laid at the door
> of the W3C alone - even if the RFC in question was authored with W3C
> participation.

RFC is *not* a standard. Several RFCs have been "violated" by several
companies, IIRC Apache has still a fix for a fu by Netscape on board.

>> It's funny how a "flawed" company like MS did create a nice way to
>> do things conditionally, in an early stage admitting that things
>> could change drastically. IIRC, at that time CSS was draft, and
>> unclear.
>
> The CSS specification mandates clients to ignore CSS syntax which they
> don't understand. But no: MSIE had to know better, and overrule that
> mandatory requirement, and screw up when doing it. Things have only
> changed superficially since then.

How does a conditional comment "screw" up? MS foresaw issues, issues
that now require to exploit parser flaws for other browsers.

> I suppose it's nice, in a perverse sort of way, that they've granted
> us a non-standard

There we go again: CSS is a recommendation, *not* a standard.

> feature to defeat their non-standard features. I
> still prefer a specification-conforming WWW-compatible browser myself,

Then download and install Amaya and join the happy few who have done so
already :-D.

> but as an author I need to be aware of users out there who wouldn't
> recognise a specification-conforming browser if it bit them in the
> bottom, sadly.

standard, now specification: get it right: w3c recommendation, or w3c
draft, or maybe even RFC depending on what exactly you're talking about.

> Not that the Mozilla-family, or Opera, or Lynx or whatever, are
> entirely free of bugs, but at least one can have reasonable confidence
> that their developers have some commitment to remedying those bugs
> when they are identified, instead of (as appears to be the MS policy)
> replacing known bugs with unknown bugs,

That's normal in software development: when you fix one bug, you often
introduce new code or change existing code, and hence new bugs might pop
up. The code base of an application rarely gets smaller. I am sure you
can find reintroduced bugs in all browsers you named, or bugs that went,
and introduced new ones.

Opera is core business for the company that produces it.

Lynx, Mozilla, whatever are Open Source. The latter, Mozilla & Co, still
suffer from the gigantic shitty code Netscape produced some time ago.
Have a look at the history format for example (or better: don't).

> and requiring those who desire
> specification-conforming behaviour to research obscure workarounds.

Even if MS dropped dead today your problem wouldn't go. The reason why
things slowly seem to settle is that IMO there is hardly any progress in
the web development.

--
John Perl programmer: http://johnbokma.com/perl/perlprogrammer.html

TextPad+TortoiseSVN:http://johnbokma.com/textpad/textpad-subversion.html

[Back to original message]