W3C home > Mailing lists > Public > www-ws-arch@w3.org > January 2003

Re: Dynamic invocation vs. late/dynamic binding

From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Date: Wed, 8 Jan 2003 01:21:22 +0600
Message-ID: <022201c2b681$f87886c0$6f00a8c0@lankabook2>
To: "Mark Baker" <distobj@acm.org>
Cc: <www-ws-arch@w3.org>

Hi Mark,

> Late/dynamic binding means being able to manipulate squares and circles
> with the Shape interface.  Dynamic invocation means being able to
> construct, for example, a "displaySquare" message without compile-time
> knowledge of the full Square interface.

That's fine - WSIF can handle that using something called JROM we
created (see alphaWorks again) to represent arbitrary schema typed
values. 

Clearly, in the absence of magic the information about the interface
(namely the data type defs) is needed at runtime at least (possibly
using xsi:type), so once that's available you're on easy street.
 
> The former enables a client written to access Shape objects, to later
> access triangles, ovals, hexagons, you name it.  The latter doesn't.

I guess we're back to the REST vs. WS debate; your program cannot
manipulate those shapes in a meaningful way without an understanding
of what an oval is vs. a square. In the BPEL scenario the BPEL process
would need to understand all the shapes and then the underlying
infrastructure of WSIF/JROM can handle everything just fine.

Sanjiva.
Received on Tuesday, 7 January 2003 14:25:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:12 GMT