OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   RE: Namespace and Versioning

[ Lists Home | Date Index | Thread Index ]
  • From: "Bullard, Claude L (Len)" <clbullar@ingr.com>
  • To: Henrik Frystyk Nielsen <frystyk@microsoft.com>, xml-dev@lists.xml.org
  • Date: Wed, 30 Aug 2000 16:12:20 -0500

Because the application types keep changing both 
in time and location.   One URI won't do it.  
I have to know the version.  We are not the 
only ones changing the schema.

The police databases, for example, have interesting 
problems not dissimilar to others but a little exagerated:

1.  There is only one standard for the information, 
NIBRS, and it is Federal.

2.  Each agency gets to futz with it by say, city 
and by state.  Actually, the state may clean it up 
on the way to the FBI but below that, it is agency 
specific.

3. And just as an aside, simple validation is not good enough.   
Why?  Well, it is one of the few database apps where trying to put 
false information into it is not only a sport, but a 
survival tactic for the databasees.  Lieing is not only 
possible, but expected (AKA tracking, driver's who lie 
about the vehicle owner, etc).  The entity relationships are 
pure hell.

So we have several versioning issues some of which 
we handle in the local database using lookups (agency stuff), and 
some of which we have to handle by versioning the 
base schemas.  Yes, it can be done with a URI but 
only if the URI has version number on the end of it.  
Not complex, but I think necessary.  My only reason for 
citing SOAP originally was that I had a sample in front 
of me.  The issue is really for schemas that are 
quickly evolving but are all of which (major/minor 
changes) are still in use.  

We say we sell COTS but that is wishful thinking.

Are you saying, the URI is used but then the system 
negotiates for the version?

As for names, there is that thing about having power 
over that which can be named.  On the other hand, 
it hasn't worked on my wife yet and is slipping 
quickly with my kids.  Classic three body problem... :-) 

Len Bullard
Intergraph Public Safety
clbullar@ingr.com
http://fly.hiwaay.net/~cbullard/lensongs.ram

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Henrik Frystyk Nielsen [mailto:frystyk@microsoft.com]
Sent: Wednesday, August 30, 2000 2:54 PM
To: Bullard, Claude L (Len); xml-dev@lists.xml.org
Subject: RE: Namespace and Versioning


> Ok for messaging and again, for central standards shared
> across the uberSystem.  Probably not so
> good for database management of applications hosted on that system.

Why isn't it an advantage to have a single URI instead of a (URI,
version-number) tuple for the envelope as well as for any "application"
which can live within a SOAP message?

> This is a juicy topic for those who like distributed chaos
> and systems
> that run best on the edge of turbulence.

Naming is always juicy, no?

Henrik




 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS