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

Re: Objects at REST...

• From: Rick Jelliffe <rjelliffe@a...>
• To: XML Developers List <xml-dev@l...>
• Date: Fri, 14 Mar 2008 01:57:14 +1100

Peter Hunsberger wrote:
> Sure, but the question then becomes; how do you define "important"?
>   
By specifying the use cases, a la UML. And/or the Information Units a la 
Maler and el Andeloussi. And/or the Configuration Items a la ITIL even.

Ultimately, by modeling what the relevant public's conception of the 
structures in the document set is.  For our legislation, it is the view 
of the legislation from the eye of lawyers. (Drafters are another use 
case, and they may have a different conception, and so if they were a 
significant target, you would also have a set of URLs for them.)

A real success would be if an end-user (who is aware of the conventions 
and the domain) could sit at the computer and write the PRESTO URL for 
any arbitrary numbered thing in a piece of legislation. So if they 
wanted the third paragraph of the fifth list item of the first list in 
Part 1, section 2.3 for the 1999 revision of the Child Protection Act of 
1927 they could type

   
http://www.eg.com/ChildProtectionAct1927/1999/part1/section2.3/list1/item4/para4

and the system would deliver to them the best granularity version 
possible. Or they could try
 
   
http://www.eg.com/ChildProtectionAct1927/1999/part1/section2.3/list1/item4/para4/application/pdf

to specify they wanted the PDF in particular. Or they could type

   
http://www.eg.com/ChildProtectionAct1927/1999/part1/section2.3/list1/item4/para4/incoming-references

and see as best a list as possible of the other documents or clauses 
that reference that para, item, etc.

(I am not saying that I expect end-users to type these. But at a pinch 
they could.)

---------------------------------------------------------------------------

Another way to look at the issue is this: if I am a programmer and I am 
told to have a table of contents of each of our zillions of WordPerfect 
files available, what URL should I use? At the moment, a programmer is 
quite likely to do something like
   
http://www.eg.com/sevlets/getToC4WP?file=http://www.eg.com/files/WP/100035.wp
and then it is unclear what a table of contents should be linking to: is 
there a scheme for accessing fragments of WorkPerfect data?

In the PRESTO approach, you would say
   http://www.eg.com/files/100035.wp/TOC
where you form the URL hierarchically by concatenating the parent 
object's  URL with the name of the output of the (now disappeared) method.

Or, better,
   http://www.eg.com/files/TheDarkAndStormyNight/TOC
and the table of contents would have links such as
  http://www.eg.com/files/TheDarkAndStormyNight/para36
and so on.


Cheers
Rick Jelliffe

• References:
  • Objects at REST...
    • From: "Andrzej Jan Taramina" <andrzej@c...>
  • Re: Objects at REST...
    • From: "Peter Hunsberger" <peter.hunsberger@g...>
  • Re: Objects at REST...
    • From: Rick Jelliffe <rjelliffe@a...>
  • Re: Objects at REST...
    • From: "Peter Hunsberger" <peter.hunsberger@g...>