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

RE: Traditional RPC

• To: "'Simon St.Laurent'" <simonstl@s...>, xml-dev@l...
• Subject: RE: Traditional RPC
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Tue, 19 Feb 2002 09:15:00 -0600

As I said yesterday, between objectivism and selflessness 
is "do the right thing for the right reasons and it usually 
turns out alright".  It isn't the most secure middle ground 
one can imagine and seems to depend as much on luck as 
rigor.  On the other hand, a balance point is not known 
for rigidity; it is known for sensitivity to perturbations.

1.  Al makes a good point that packing arguments and names 
into URIs and RPC calls look a lot alike above the mapping. 
The difference is how one goes about extending the application.

2.  The Lovett article http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnexxml/html/xml10162000.asp
makes the point that XML As API works as long as one is willing to 
give up features, often features unnecessary for the task at hand. 
That is what "coarse grained" means.  The fine-grained object API 
has a lot of neat stuff, but the cost is hard to maintain code for 
little used features.

3.  Both point to the context of use as the determinant for choosing.

The "automagic hookup" hype does not account for the exceptions 
where the little used feature is useful.  It depends on a world 
of standard XML vocabularies with the exceptions hammered out. 
The URI/HTTP world depends on eliminating choices of verbs and 
saying, "if you have exceptions, you have to build them up 
from these primitives".

I don't see much difference except in picking who gets to 
be vindicated and who gets the pain.

len

-----Original Message-----
From: Simon St.Laurent [mailto:simonstl@s...]

2 and 3 are useful pluses for protocol success, but I don't think either
inherently reduces the risk of a popular but poorly implemented protocol
creating an enormous mess.  On the bright side, more pain MIGHT mean
more people willing to put the time in to fix that pain.

There's got to be a balance point there somewhere...

• Prev by Date: Re: Streaming Transformations for XML
• Next by Date: Re: Re: [namespaceDocument-8] 14 Theses
• Previous by thread: Re: Traditional RPC
• Next by thread: RE: Traditional RPC
• Index(es):
  • Date
  • Thread