Re: [ElementTraversal]: Feature string for DOMImplementation.hasFeature(feature, version)?

On Sun, 18 Oct 2009 01:01:02 +0200, Doug Schepers <schepers@w3.org> wrote:

> Hi, Anne-
>
> Anne van Kesteren wrote (on 10/17/09 2:33 AM):
>> On Sat, 17 Oct 2009 01:46:56 +0200, Doug Schepers <schepers@w3.org>  
>> wrote:
>>> Sorry for the tardy response.
>>>
>>> This was an unfortunate oversight. I've now added this to the proposed
>>> errata [1]. Please let me know if this suits your needs.
>>>
>>> [1] http://www.w3.org/2008/12/REC-ElementTraversal-20081222-errata#S1
>>
>> Didn't we explicit decide against this because you could easily feature
>> test it?
>
> I don't recall that, and can't find any reference to it in the  
> archives... do you have a link?  It's possible it was discussed in some  
> telcon of F2F that I don't recall, but was not minuted.

We get this request for lots of things, and often we have decided not to
support it.

> In any case, I don't believe that adding a feature string is harmful or  
> introduces significant implementation burden.  If it is reported  
> accurately, it is useful, and in non-browser environments, where there  
> may be different DOM implementations available, it is necessary for the  
> DOMImplementationRegistry (as Michael mentioned).

In this case, I think the spec is pretty easy to get right. So I don't see
a lot of harm coming from supporting it, although I am not convinced that
it is a brilliant way (in general) of figuring out what works -
unfortunately in complex cases there is far more likelihood of it
something only partially working, which means the feature string becomes
useful only for very simple cases.

It is possible to check whether something is supported (and how well)  
without the feature string, in many cases, by testing for a method or even  
testing that it does the right thing - certainly the overhead of doing  
this for ElementTraversal, while non-zero and harder than checking for the  
feature string is also not really complicated for most cases I can think  
of. That said, it seems that the case of querying the implementation  
registry probably requires a feature string.

cheers

Chaals

-- 
Charles McCathieNevile  Opera Software, Standards Group
         je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com

Received on Saturday, 17 October 2009 23:22:28 UTC