Re: VOQL-TEG Meeting #3

From: Benjamin Gufler <benjamin.gufler-at-in.tum.de>
Date: Thu, 25 Jan 2007 16:45:03 +0100


On 2007-01-25 15:16, Yuji SHIRASAKI wrote:
> - It is not covenient for both a client and service provider to have
> three-step operations for that:
> 1) create a table and insert a votable to the table, or just upload
> a votable and give a virtual table name to it,
> 2) execute a query,
> 3) delete the table and/or votable.

I do agree on that to a client it might be more comfortable to have this whole workflow available as a single operation.
>From the point of view of the service provider, however, this would
imply the need to provide the "atomic" operations (upload a votable, execute a query, delete a votable) as well as the combined operation. Furthermore, we would surely be able to find other combinations of the atomic operations which would be nice-to-have for the clients. I fear that the approach of providing combined operations would lead to a large list of operations a service provider had to offer, and most of them would be just be combined invocations of other services which are also offered.
Hence, what I would prefer to have is an (ideally rather small) set of atomic operations even in the service specification, which would nevertheless give client applications the opportunity to offer combined operations to the user and just invoke several of the atomic operations behind the scenes. This way, the clients would still have their comfortable application, but the server side doesn't get more complex.

Just my 2 cent.

        Benjamin

-- 
Dipl.-Inf. Benjamin Gufler       Lehrstuhl Informatik III
Tel. +49 89 289-17276            Fax +49 89 289-17263
Boltzmannstr. 3, 85748 Garching  Raum FMI 02.11.035
Received on 2007-01-25Z16:45:19