Re: A(nother) Guide to Publishing Linked Data Without Redirects

Hi Harry,

On Nov 10, 2010, at 19:50, Harry Halpin wrote:

> On Wed, Nov 10, 2010 at 11:15 PM, David Wood <david@3roundstones.com> wrote:
>> Hi all,
>> 
>> I've collected my thoughts on The Great 303 Debate of 2010 (as it will be remembered) at:
>>  http://prototypo.blogspot.com/2010/11/another-guide-to-publishing-linked-data.html
>> 
>> Briefly, I propose a new HTTP status code (210 Description Found) to disambiguate between generic information resources and the special class of information resources that provide metadata descriptions about URIs addressed.
>> 
>> My proposal is basically the same as posted earlier to this list, but significantly updated to include a mechanism to allow for the publication of Linked Data using a new HTTP status code on Web hosting services.  Several poorly thought out corner cases were also dealt with.
> 
> I don't this solution cuts it or solves the problem to the extent that
> Ian Davis was proposing. To recap my opinion, the *entire* problem
> from many publisher's perpsectives is the use of status codes at all -
> whether it's 303 or 210 doesn't really matter. Most people, they will
> just want to publish their linked data in a directory without having
> to worry about status codes. So, de facto, the only status code that
> will matter is 200.
> 
> The question is how to build Linked Data on top of *only* HTTP 200 -
> the case where the data publisher either cannot alter their server
> set-up (.htaccess) files or does not care to.

Yes, I understand the tendency to think this way.  It is easy to implement and understand.

However do *you*, as a more knowledgeable individual, really think we can build a Web of Data (or whatever you want to call it) by overloading both the http:// namespace and the 200 status code?  I sure don't.  In fact, I think it is silly to try especially in the absence of any standard way to understand what we got back.

We are already running into serious problems trying to deal with physical and conceptual resources being given http:// URIs, but not being resolvable.  We are stressing the (very young) Web but not even solving basic problems.

In my opinion, we need a way to tie together (and yet allow to be separate) the Web of Documents and the Web of Data.  To me that means that we need a hook to separate an information resource (the 200 status code) from a general metadata description (what I proposed as the 210 status code).  My proposal doesn't have to be "it", but something does.  Without some separation at that level, we will continue to have practical problems.

If we forgo any status code separation, then we will have to introspect every 200 result to acquire any information about what we got back.  I don't think that is practical.

Regards,
Dave



> 
>> 
>> I look forward to feedback from the community.  However, if you are about to say something like, "the Web is just fine as it is", then I will have little patience.  We invent the Web as we go and need not be artificially constrained.  The Semantic Web is still young enough to be done right (or "more right", or maybe "somewhat right").
>> 
>> Regards,
>> Dave
>> 
>> 
>> 
>> 
>> 
>> 
>