Login | Register
My pages Projects Community openCollabNet

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Catacomb] Proposed Road Map



Elias Sinderson wrote:

Sung Kim wrote:

This is proposed Road-Map so please share your ideas. I'd like to post it at our project homepage, after we have enough discussion.

GOALS
 - Implement all DAV protocols.
 - Support All kind of DBMS.

Wow. These are serious goals. Implementing all the DAV protocols will be a lot of work, but not impossible.

Got that right!  We'll make Sung do all the work. Heh. Just kidding.

Supporting all DBMS seems a little unrealistic to me - How will we support Oracle (for example) without even being able to pay for the license ourselves? It seems like scoping this goal to only support available DBMS is more reasonable...

How about: Support MySQL, PostgreSQL (the two biggest Open Source players.) But, more importantly, develop generic API for integration with other RDBMS's so those who have Oracle or Sybase or MSSQL or whatnot can develop their components, provide a common space for RDBMS component developers to share their code. (Akin to the JDBC model.) Or, another option, would be to piggy-back on top of another generic API (like http://libdbi.sourceforge.net ) but it's sure that SQL statements will be different between each RDBMS, depending on what features they have and how they implement them. (I.e. transactions, sequence #'s, table locking, etc. Always db-specific.)

Proposed Mile Stones

Add (at a high priority, IMHO) external file storage. The 16MB *total* size limit of a database will seriously hinder serious adoption of Catacomb.

 - DBMS Abstraction layer

Halfway there. ;^)

 - Discovery of Supported Query Grammars
- DeltaV Core Versinonig package (VERSION-CONTROL. CHECKOUT, CHECKIN, UNCHECKOUT, REPORT).
 - ACL
 - DeltaV Workspace Package (WORKSPACE, VERSION-HISTORY)
 - BIND
 - DeltaV remains

Not sure if these are in any specific order, but it appears they are. I'd place a higher importance on ACL than Delta-V. The specification is more mature, and much easier to implement. Also, what do you mean by the BIND milestone? How does this relate to Catacomb?

Depends entirely on what you consider ACL implementation. Catacomb will not be able to manipulate the permissions of documents in the filesystem, but it might be possible to map ACLs to permissions in MySQL. 'course, that would mean it would be MySQL-specific. I've developed ACL logic for other projects and it's not that hard (and we can steal the logic even though it's written in Java.) I would propose that ACLs be stored in tables in the db, and methods would have to enforce the ACLs. The protocol is fairly straightforward, but has a multitude of options that we might want to implement.