15 lines
1.4 KiB
Plaintext
15 lines
1.4 KiB
Plaintext
%META:TOPICINFO{author="GarryJolleyRogers" date="1259118871" format="1.1" version="1.4"}%
|
|
%META:TOPICPARENT{name="ImplementationsOfBDI.SDD"}%
|
|
---+!! %TOPIC%
|
|
|
|
I've been thinking about how SDD will be used in the context of CIPRes. I pressume users will be depositing matrices with up to hundreds of characters and taxa per matrix and sometimes multiple matrices per study. Under these circumstances, there would be a lot of bytes to transfer over the wire and we have to be careful not to send exceeding large SOAP messages. With this in mind, I propose that we decouple the terminology and externalDataInterface sections from the core description section of SDD to make it possible for users to submit multiple matrices belonging to a single study without having to repeat the same vocabulary and media resources assuming those are common between the matrices. This would mean that the first matrix would contain the entire vocabulation section, but subsequent matrices would simply refer to the shared voculabulary using some form of identification scheme that we would agree upon. The identification would refer to the entire vocabulary block defined in the first matrix rather a single taxon. Same goes for the media resource stuff.
|
|
|
|
What do you think of the idea of decoupling these sections in an SDD document and if so how would you recommend that we go about implementing these changes?
|
|
|
|
Thanks,
|
|
|
|
Shirley
|
|
|
|
|
|
|
|
-- Main.ShirleyCohen - 31 Dec 2004 |