[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Best Practice for URI construction?

• To: "Costello, Roger L." <costello@m...>
• Subject: Re: Best Practice for URI construction?
• From: Chris Burdess <dog@b...>
• Date: Fri, 9 Dec 2005 14:49:10 +0000
• Cc: "XML Developers List" <xml-dev@l...>
• In-reply-to: <B8415163A689094689542C617ECA036650F585@I...>
• References: <B8415163A689094689542C617ECA036650F585@I...>

Costello, Roger L. wrote:
> (1) The above URLs (I believe) are expressing the same thing - they 
> are identifying the same resource.  So which is "better"?

For practical purposes, the first is better, since there is no limit to 
the number of name components. In the second example, you can have at 
most 2 or 3 parameters before search engines give up and stop indexing 
your resource.

In theory, in the first case the semantics of each level of the 
hierarchy are implicit but you have the advantage of an additional 
semantic axis (taxonomic relationship), whereas in the second this axis 
is absent but the names of the semantic roles of each component are 
explicit. Also, in the first case we can make the fact that the 
resource represents a collection of other resources explicit 
(http://www.location.org/US/MA/), whereas in the second case there is 
no way to know whether it represents one resource or many just from its 
form.

> (2) As was noted at the top, the purpose of a URL is to "identify" a 
> resource.  Can every resource in the universe be identified using the 
> above two approaches? 

Of course. In fact, you don't need more than one name component to 
identify every resource in the universe, although this can lead to some 
pretty unwieldy names.

> Are there resources that do not lend themselves to identification 
> using the above two approaches? 

If there is a resource that doesn't fit into any kind of taxonomy, then 
presumably the first is not the best kind of representation for it, 
unless we assume that the "root" level contains names of objects of 
this kind.

If there is a resource that cannot be differentiated from other 
resources on the basis of explicit properties, then the second form is 
useless.  However, normally when this happens people just use a vacuous 
property name like "name" or "id" or "num", and assign the resource a 
taxonomic name or serial number.
-- 
Chris Burdess

• Follow-Ups:
  • Re: Best Practice for URI construction?
    • From: Tim Bray <tbray@t...>

• References:
  • Best Practice for URI construction?
    • From: "Costello, Roger L." <costello@m...>

• Prev by Date: Re: Best Practice for URI construction?
• Next by Date: Re: Need good C++ XML Library
• Previous by thread: Re: Best Practice for URI construction?
• Next by thread: Re: Best Practice for URI construction?
• Index(es):
  • Date
  • Thread