RE: Myth of Loose coupling

Comments inline.

> -----Original Message-----
> From: Doug Kaye [mailto:doug@rds.com]
> Sent: Wednesday, October 01, 2003 12:27 AM
> To: 'David Orchard'; www-ws-arch@w3.org
> Subject: RE: Myth of Loose coupling
>
>
> Dave,
>
> I like your list of loose-coupling properties. May I suggest
> two others that
> I don't think are covered by your list?
>
> 1. Data validation through published schema. (As opposed to
> "by convention"
> [brittle] or as part of the service [too fine-grained and noisy].)
>

In general, that makes sense.  I'm not quite sure yet the relationnship
between coupling and typing.  For example, is corba's typing more or less
coupled than html web or web services?  And I don't get the "data validation
as part of service".  How does that compare the html web vs xml web services
vs corba vs dcom vs smtp vs ?

> 2. Delayed binding. (Just in general.)
>

I'm not sure about that, but I think it ties into the trade-off between
loose coupling, typing and robustness.  Which starts us down the "are
compiled applications more or less loosely coupled than interpreted
applications" path that I don't actually want to go down.

Cheers,
Dave