W3C home > Mailing lists > Public > xml-dist-app@w3.org > September 2001

Re: sparse arrays - too complex?

From: Tyler Close <tjclose@yahoo.com>
Date: Sun, 23 Sep 2001 11:09:34 -0400
Message-Id: <5.0.2.1.1.20010923102749.009ecda0@pop.mail.yahoo.com>
To: Jacek Kopecky <jacek@idoox.com>, Rich Salz <rsalz@zolera.com>
Cc: Alan Kent <ajk@mds.rmit.edu.au>, SOAP <xml-dist-app@w3.org>
At 03:44 PM 9/21/01 +0200, Jacek Kopecky wrote:
>SOAP doesn't even have native maps (associative
>arrays) and I value maps much higher than "diff" arrays.

Given this, why are sparse arrays in the spec? They seem like a very low 
value feature that create a disproportionate amount of complexity and 
confusion.

I read a previous post where it was said sparse arrays are useful for 
handling legacy RPC systems. If this is the only reason, then it seems like 
a poor idea to stick them right in the main SOAP spec. I think they would 
be better treated as just another application specific data structure (ie: 
"Generic Compound Type"). At best, the sparse array format should be put in 
an appendix for "The recommended way to encode type X from legacy system Y."

Specific application areas are full of custom types, so it seems unusual to 
give sparse arrays special treatment. Putting sparse arrays in the main 
SOAP spec forces all SOAP users to contend with complexity flowing from one 
application type. Good design does not let complexity flow in this direction.

Tyler


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
Received on Sunday, 23 September 2001 11:12:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:03 GMT