W3C home > Mailing lists > Public > public-html-a11y@w3.org > September 2012

In defence of iCab as an example...

From: Charles McCathie Nevile <chaals@yandex-team.ru>
Date: Mon, 17 Sep 2012 09:07:08 +0200
To: "Joshue O Connor" <joshue.oconnor@cfit.ie>, "public-html-a11y@w3.org" <public-html-a11y@w3.org>
Message-ID: <op.wkrp16yyy3oazb@chaals.local>
In another thread, where iCab was cited as an example for supporting  
On Sun, 16 Sep 2012 10:50:55 +0200, Joshue O Connor  
<joshue.oconnor@cfit.ie> wrote:

> With all due respect to iCab, I kinda wish people would stop referring  
> to it as some kind of model of best practice.

I don't think anyone suggests in this group that it is the best browser  
available. But it is perfectly reasonable to suggest it does certain  
things best. (TL;DR: And it is a real browser for the full web, so it's a  
reasonable example of a genuine product.)

In the longdesc case, Opera's implementation was inspired by iCab's which  
was regarded as best current practice for a browser. That wasn't a tough  
call, given the alternatives on offer - would you argue that Firefox 5  
years ago (the text of the link is shown in a context menu) or Safari (you  
can get it from the developer tools) represented best practice?

> I don't know anyone who uses it

I use it. And you know me. I even pay money out of my own pocket for the  
license. I don't think it is unequivocally the best tool, but it certainly  
does some things better than other browsers - and did a decade ago (image  
maps spring to mind...)

> and (I'm not singling you out here Leif, it's just something that has  
> been on my mind for a while) and just because it has some cludge for  
> handling @longdesc - to me, that doesn't necessarily mean that this is a  
> model of best practice.

In the case of longdesc I would bet a pint that you can't find a better  
native implementation in a browser. (I don't think it is the best possible  
implementation, as I have said before. But that's a different statement).

On the other hand iCab used to be pretty good for accesskeys, having  
figured out earlier than most other browsers (although about the same time  
that mobile browsers got it - i.e. last century) that letting them  
interfere with the default user interface was a bad idea. I can no longer  
figure out how to activate them at all - but in any case there are better  
and much better implementations around now.

> More so, not something we should hang our hat on as we try to
> engineer a solution for the use cases that @longdesc or its
> successor need to cover.

Why not? iCab is a reasonable browser, and it is a genuine product  
designed to support browsing the whole of the Web. According to the  
current "emerging consensus" it qualifies, for example, as an  
implementation when deciding if a feature has been implemented, and there  
is no reason that a given feature implementation in iCab is not best  
practice. The fact that few people use it (apart from the hundreds of  
people you know, it appears there are hundreds of millions of others who  
don't, even taking into account that it actively hides itself to avoid  
site-compatibility problems) doesn't mean anything for that - it merely  
suggests that it doesn't have the combination of marketing, overall best  
practice for the things that the people you know care about, and luck,  
that might make it more popular.

To turn your argument around, many people I know use IE, Chrome, or  
Safari. That doesn't mean their implementations of any given feature are  
any good. My own browser of choice demonstrates a combination of best  
practice implemented before anyone else, matching what others do, and  
being sadly behind.

Such is life...



Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
       chaals@yandex-team.ru         Find more at http://yandex.com
Received on Monday, 17 September 2012 07:07:42 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:30 UTC