Print

Print


I agree that our server does it wrong and we will change it. On the
other hand I would support to allow  unpaired response/request messages
because I think that it allows much more flexible with respect to future
extensions.
In case you receive a "specialTheoRequest" (which is not supported)
what would be the appropriate response? And what would be the
alternative response on a searchRetrieveRequest with the "scanOnFail"
flag set?

Theo


>>> [log in to unmask] 29-2-04 19:15:20 >>>
In SRW, we currently have the response/request messages paired in the
WSDL. i.e. only the scanResponse is a valid response to a scanRequest
(ditto for searchRetrieve and explain). We could modify the WSDL to
allow any response to be sent in response to any request, but I'm
loath
to lose this additional syntactical validation.

So in SRW, Adam's server is the correct behaviour.

I can't see any particular reason why SRU and SRW should differ in
behaviour here.

Matthew


> -----Original Message-----
> From: Z39.50 Next-Generation Initiative [mailto:[log in to unmask]]
> On Behalf Of Adam Dickmeiss
> Sent: Sunday, February 29, 2004 8:32 AM
> To: [log in to unmask]
> Subject: Re: Operation not supported?
>
> LeVan,Ralph wrote:
>
> >I sent Theo's server an SRU scan request.  I got back a
> >searchRetrieveResponse with a diagnostic explaining that
> scan was not a
> >supported operation.
> >
> >Is that the right way to respond?
> >
> >
> >
> I just checked my server. It returns a scanResponse with the same
> diagnostic. Possibly an explainResponse would be more proper.
>
> -- Adam
>
> >Thanks!
> >
> >Ralph
> >
> >
> >
>
>