W3C home > Mailing lists > Public > public-webapps@w3.org > January to March 2009

Re: [selectors-api] Selectors API I18N Review...

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Sat, 31 Jan 2009 22:53:38 +0100
Message-ID: <4984C862.8010906@lachy.id.au>
To: "Phillips, Addison" <addison@amazon.com>
Cc: "public-webapps@w3.org" <public-webapps@w3.org>, "public-i18n-core@w3.org" <public-i18n-core@w3.org>

Phillips, Addison wrote:
> Assuming for a moment that CSS3 does not resolve the problem with 
> Unicode normalization in selectors, users of selectors-api will be 
> potentially surprised by the results in certain circumstances 
> involving denormalized selectors.

I have discussed this with my co-editor, Anne, and we have a number of 
concerns about including such a note within Selectors API.

The core of the issue is the Selectors specification itself, which you 
feel inadequately addresses normalisation, and so the issue is likely to 
occur not only in Selectors API, but in all other uses of Selectors by 
other specifications.  This includes, but may not be limited to, CSS, 
XBL and CSSOM.

The note itself will only remain relevant and accurate for as long as 
Selectors does not include normalisation.  If a subsequent version of 
Selectors does include it, then the note will become misleading and 
inaccurate.  This will be of particular concern once Selectors API 
becomes a recommendation, at which point it will be too late to remove 
it in the event of Selectors adding support for normalisation in the future.

As such, if the CSSWG does not require normalisation in the current 
specification, it seems like it would be more beneficial for authors if 
the proposed note were instead included within Selectors itself, rather 
than individually within each existing or future specification that 
depends on it.

> Implementers will also benefit by not getting stuck on test cases
> that involve normalization.

I do not understand that claim because, since Selectors currently 
doesn't require normalisation, any test cases that relied on 
normalisation would be invalid tests.

-- 
Lachlan Hunt - Opera Software
http://lachy.id.au/
http://www.opera.com/
Received on Saturday, 31 January 2009 21:54:21 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:49:29 GMT