> > Apart from it isn't a completely different protocol -
> For someone who's built his client using a web-browser and 
> style sheets, or using LibWWW-Perl and and XML parser, it's a 
> completely different protocol.
> > So the amount of code changes to the client would be 
> relatively minor.
> Only if you built your SRU client, from the beginning, using 
> technology that most potential implementors would consider 
> ludicrously overpowered.

To convert an SRU client to SRW you need (at a minimum) to do two things

a) all the code for creating the UI, getting user input etc. remains the
same up to the point where you build the request URL - instead of build
a URL and doing a GET you use the same data to build an XML doc and POST
b) when you get the response back strip the top most few elements, and
then handle is as before.

That's hardly working from scratch and most of your client code remains

> Well, sure -- if someone wants a Web-Service version of the 
> protocol because they want to use Web-Service stuff with it, 
> then they need the Web-Service version instead of the URL 
> version.

Someone who has invested time in developing an SRU client is going to
impacted just as much (or as little) discovering that he should have
been developing SRW whatever the reason (complex metadata requests,
authentication system integration, etc.)