Process, Slide 1.x, Slide 2.0

[Remy Maucherat <re...@betaversion.org>]

Quoting jon ewing <jo...@mediasurface.com>:

> Hello,
> 
> I'd like to find out more about the current ideas that people have
> concerning the role of Process within Slide.
> 
> I would like to help define and build the Process aspect of Slide, so
> any
> thoughts would be much appreciated!

I can rule out both a process engine and a process helper for Slide 1.x. It is 
not designed to operate asynchronously, which is mandatory since an approval 
process can take a while to execute.
For Slide 2, my idea was to have a process helper which would be very similar 
to the lock helper or the security helper : it would allow to set processes 
triggered on specific actions, and would be used by the other helpers 
transparently (like security and lock).
I was thinking about basing the Slide 2 core around a process engine. I don't 
know yet how generic this process engine can be. Some internal services will 
also be added to provide pluggability and robustness (low level locks, journal, 
id generator, transaction manager ...).
Eventually, I would like the helpers to operate like this :
- check security
- check locks (the high level ones)
- check approvals / notifications (from the new process helper)
- create the appropriate process
- schedule the process for execution in the process engine
- assume it will go ok and return
A new structure in the SlideToken will allow to keep track of the process 
status.
Indeed, everything would be a process. For obvious reasons, it can't be done in 
Slide 1.x. Note that although the core is different, the process engine will 
still make use of all the Slide 1.x stores to store / retrieve objects. Except 
for the helpers and parts of the core, I expect Slide 2 to have a lot of code 
in common with Slide 1. Hopefully, it will also be possible to have binary 
compatibility between the two versions, the Slide 2.0 API being an extension of 
the Slide 1.1 API.

Slide 1.0 is nearly done (feature wise). It still needs :
- finish a first version of the DAV command line client
- add DAV ACL support (both client and server)
- finish the JTA support (that's a last minute feature addition, and it should 
be done very soon)
After that, it will go in beta, and any DAV issues / bug will have to be 
addressed.

Slide 1.1 should add :
- index/search helper
- at least one index provider (like Lucene)
- DASL (also in the client)
- Delta V (also in the client)
- additional stores (that's a requirement for Slide to be useable in the real 
world)

Remy