W3C home > Mailing lists > Public > public-html@w3.org > August 2007

Re: [HDP] Other comments from RI

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Mon, 27 Aug 2007 23:36:05 +1000
Message-ID: <46D2D345.3080608@lachy.id.au>
To: Olivier GENDRIN <olivier.gendrin@gmail.com>
CC: Richard Ishida <ishida@w3.org>, public-html <public-html@w3.org>

Olivier GENDRIN wrote:
> On 8/23/07, Richard Ishida <ishida@w3.org> wrote:
>>> 3. Degrade Gracefully
>> In fact, in the statement "new features should work reasonably well 
>> in older user agents that do not support the functionality" what 
>> constitutes an 'older browser'?
> 
> Perhaps we can define an older browser as a browser that is older 
> than 5 years and jounger than 10. And ignore browser older than 10 
> years. Or rely on versions (older browser : version n-2, with n being 
> the most young stable version).

I don't think it's a good idea to try and nail down a specific set of 
qualifications for browsers to be considered, or at least not too 
specific.  It's a guideline only, it's not set in stone, and one needs 
to consider various factors.

One of the major factors is market share.  Obviously, the greater the 
market share, the more users who would be affected and the less likely 
it is that a feature would be adopted if it doesn't degrade gracefully 
in it.  This would obviously mean that IE 6/7 and Firefox are relatively 
important, since collectively they currently have the greatest market share.

Opera and Safari may (depending on which statistics you believe) each 
have relatively small market share compared with Firefox and IE, at 
least in the desktop browser market, but they both also hold a 
significant portion of the mobile market too.  However, older browsers 
with negligible market share like Netscape 4, IE  5.x, etc. are so 
insignificant, they're not even worth looking at.

Yet a small market share doesn't necessarily mean that some browser 
should be completely ignored.  Lynx, for example, may be used very 
rarely compared with other desktop browsers, yet even it is sometimes 
considered.  Similarly, screen readers and other assistive technology 
may have a relatively small market share, but are still quite important.

Another factor to consider is the age of the browser and its 
obsolescence.  IE6 is already 6 years old and even though it has been 
superseded by IE7 and is therefore not as important, it still can't be 
entirely ignored.

Firefox 1.0, on the other hand, is only 3 years old, yet it has been 
superseded by 1.5 and 2.0, and soon by 3.0.  Therefore, Firefox 1.0 
could be considered relatively insignificant.  Even Firefox 2, which is 
based on the Gecko 1.8 branch, is somewhat insignificant compared with 
Firefox 3.0 and Gecko 1.9.

It basically comes down to considering the cost vs. benefit of each 
specific change.  Some changes may turn out to be incompatible with some 
particular browsers, even if those browsers are relatively significant, 
yet sometimes the benefit far outweighs the cost.  For example, 
something may be incompatible with Firefox 2.0, but be fully compatible 
with the development versions of Firefox 3.0 and so it may be worth 
sacrificing a little bit of graceful degradation for the feature.

Given the time expected for this spec to be developed, and that new 
browser versions are being deployed quite regularly, it may be quite 
likely that today's browser versions will be obsolete before some 
feature even becomes implemented and used.  So the graceful degradation 
for some features in today's versions may not be particularly relevant 
compared with the next versions.

-- 
Lachlan Hunt
http://lachy.id.au/
Received on Monday, 27 August 2007 13:36:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:04 GMT