Re: Uniform access to descriptions

Pat,

Since we both agree there is a gray area. So, we should agree that there 
is ambiguous case your argument won't work and there is clear-cut case 
that my argument would be far-fetched too. Right?

It is not that I cannot understand the loose sense of IR or I don't have 
common sense. The issue is how we should make the web more generally 
useful and practical.

To invoke HTTP(x)=200 implies that there will be something in the gray 
area that will give people a difficult time.

Just think about why there is an httpRedirection-57 issue. Now, there is 
no recommendation to invoke the 200=IR logic yet, people already start 
trying to run away from that. Do you think it is healthy or do you think 
it is a good design?

I don't think so.  The key issue is to make a clear distinction between 
"representation" and "resource".  Any  self-description is a potential 
trap to a logical paradox. 

Sure, some representation can be *very* similar or identical to that 
resource, such as all those document. But there is no need  to confuse 
the entire web for the sake of these subset of things.  I have 
recommended a solution, such as designating another 2XX, say 208, code 
to show that the representation is a byte-copy of the original 
resource.  This allows you to return whatever HTML or Word document, PDF 
document etc., meaningful.  But 200 is the default OK behavior for all 
transported message, which relationship to the external resource isn't 
that *faithful*. When we design a system, we should not make the default 
behavior based on a set of particulars, right?  Open 200 for everyone 
and make 208/303 a pair.  Hence, if a user want and be sure to make such 
distinction, do it and they should know the consequence of being 
potentially judged by HTTP(208)=IR.  But please leave the default 200 
for someone who don't care or care but unable to care.  Isn't this 
approach much more reasonable than httpRange-14? 

Xiaoshu

Received on Wednesday, 9 April 2008 17:33:15 UTC