[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ih] URl construct

On Jul 5, 2012, at 12:30 PM, Dave Crocker wrote:
> On 7/4/2012 8:40 AM, John Curran wrote:
>> That's very much an IETF creation
> I think it wasn't.  I think it was in the original construct, based on some limited research I've done, such as:
>   http://infomesh.net/html/history/early/
> This is not to say that the milestone you cite is irrelevant or unimportant.

The notation itself was not an IETF creation; it's the co-opting 
of all other systems you referred to which was the result of the
standardization of it across information retrieval systems. 

> But I've been finding a difference between core technical points versus critical adoption (social) points.  From what I can tell, the basic construct of
>   <scheme> "://" <scheme details> was in the original Web design.

Correct.  Reference: <http://www.w3.org/DesignIssues/Naming>  (1991)

However, that was something implemented entirely within the w3 _client_,
whereby a w3 client could access objects via multiple protocols.

The real power occurred when it was standardized outside of just w3 clients 
with the "scheme:" becoming universal in interpretation for all info-systems.  
Until then, it was just shorthand for w3 clients on how to access digital
blogs referenced in w3 documents.

> The milestone you cite appears to have been the essential point of gaining agreement among disparate (and possibly competing) groups to use the same reference mechanism, with difference <scheme> values.

TimBL knew that he had to allow for other addressing schemes, and so included 
the capability in w3 clients including code to support accessing documents 
via gopher and WAIS (see <http://www.tranquileye.com/cyber/1992/AISandGopher.html>)
Again, the w3 client interpreted these w3-specified links in hypertext documents 
to find the information.  

The revolution required that the actual semantic meaning of the URL be 
understood by any client (not just w3), and hence IETF standardization.
The result was that any client (gopher, wais, www) could then include 
URLs in their documents and safely dereference...   I know that some 
of the gopher clients did such, and I began putting into one of the
WAIS clients (but by then it apparent that using HTTP servers and HTML 
documents was the best fit.)

So, let me try again:  The "<scheme>:" notation originated with the w3
work, but was adopted by others through standardization of the semantics
in the IETF.  One consequence of its adoption was the quick realization
that hypertext clients were far more flexible than clients of most other 
information retrieval systems and thus facilitated the domination of HTTP
as the dominant information retrieval protocol.