Re: Gloss standard terminology for resource/representation (ISSUE-81 Change Proposal)

On Apr 6, 2010, at 3:29 PM, Ian Hickson wrote:

> On Tue, 6 Apr 2010, Dan Connolly wrote:
>>>
>>> I believe that mentioning such differences will increase the level  
>>> of
>>> confusion, not reduce it, because few people are familiar with the  
>>> way
>>> HTTP uses these terms.
>>
>> Few relative to the population of the planet, yes... even relative to
>> the web development community as a whole perhaps. But relative to the
>> intended audience of this spec, especially sections such as the one  
>> on
>> Application caches, I believe they are a critical constituency, if  
>> not
>> the majority.
>
> The two major target audiences of the HTML spec are Web developers and
> browser vendors. In my experience the majority of both of these
> constituencies get confused when people start talking about abstract
> resources and resources having multiple representations. I'm certainly
> happy to hear from members of either group who would be less  
> confused by
> the spec talking about "resource representations" than the status  
> quo; if
> there are indeed such people, then I'd want to change the spec to  
> help them.

Speaking as a browser implementor (and with my chair hat off):

When different specifications that are relevant to my work use the  
same terms in related but different ways, I find it helpful when the  
difference is called out and clearly explained. I believe many of my  
colleagues at Apple and in the broader WebKit community feel the same  
way. HTTP, URL/URI/IRI, and HTML are all standards that we need to  
look at in the course of browser implementation. In general,  
additional true factual information is helpful unless it is so  
tangential as to be off topic.

On the whole I sympathize with Ian's view that it is simpler and more  
understandable to refer to the thing that you fetch from a URL as "a  
resource" rather than "a resource representation". In fact, WebKit  
internal class names are already aligned with this view; we have  
ResourceLoader as the object responsible for loading a stream of bytes  
addressed by a URL, and CachedResource as the class that holds such a  
byte sequence in the memory cache.

However, since HTTP and IRI usage is different, it is helpful to have  
an indication of the difference in terminology, as a warning to not  
apply a blind one-to-one mapping at the interfaces. Since HTML is  
logically built on top of HTTP and IRI, it's more logical to put the  
explanation in the HTML5 draft. To do otherwise would be a priority  
inversion.

I think it might be additionally helpful to clarify that many other  
W3C specs use "resource" in the HTML5 sense rather than in the HTTP  
sense.

Regards,
Maciej

Received on Tuesday, 6 April 2010 22:39:53 UTC