W3C home > Mailing lists > Public > www-style@w3.org > July 2005

RE: CSS is doomed (10 years per version ?!?)

From: Ian Hickson <ian@hixie.ch>
Date: Tue, 12 Jul 2005 00:22:00 +0000 (UTC)
To: Mark Moore <mark.moore@notlimited.com>
Cc: www-style@w3.org, 'Orion Adrian' <orion.adrian@gmail.com>
Message-ID: <Pine.LNX.4.62.0507120015130.5916@dhalsim.dreamhost.com>

On Sat, 2 Jul 2005, Mark Moore wrote:
> 
> You're quite mistaken.  HTML introduced explicit versioning as of HTML 
> 4. See section 7.2 "HTML version information" in the HTML 4.01 
> specification. [1]
> 
> Versioning was included in the specification precisely because quirks 
> mode was unmanageable.

HTML has had versioning information in the file format since version 2 
(and maybe even in HTML+ or earlier). Certainly since WAY before quirks 
mode was invented.

This is distinct to what was being discussed, which is implementation- 
level versioning. The only implementation-level versioning in HTML is the 
quirks vs standards vs XHTML mode switching, which is only loosely of the 
actual claims made in the file (e.g. you can trigger quirks mode with an 
XHTML DOCTYPE, XHTML mode with an HTML4 DOCTYPE, and standards with a 
version-free DOCTYPE).

The point being there are no conformance criteria related to how to handle 
the version information in the file.


> It's actually inevitable that explicit versioning will be added to any 
> system given enough releases and any meaningful user base. [2]
> 
> In a very similar way, Intel resisted versioning their CPU's for years. 
> What sprang up was an ad hoc, convoluted routine used to identify the 
> various x86 versions and steppings based on specific quirks. [3]
> 
> This CPU sniffing algorithm grew and the legacy quirks were supported by 
> every new CPU model until Intel decided it was far simpler to create an 
> explicit CPUID instruction.

You are confusing two things; versioning of user agents, and versioning of 
the file format. The former is required due to the realities of software 
development and is independent of the language itself (as per Bert's 
comment recently). The latter is only needed if the the format does not 
remain backwards and forwards compatible.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 12 July 2005 00:22:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:39 GMT