> Hello,
>
> Thanks for the interest in Delta-V.
> We (Software AG) are volunteering for a Delta-V implementation in Slide.
> If you agree, we have following plan:
> Our time frame would be, to have the Delta-V standard at least partly
> implemented until mid of 2002 (including workspaces).
> Currently we are implementing a prototype, to find out basic concepts and
> the areas in Slide, which we need to touch for a Delta-V implementation.

It's quite obvious that the "content" package will have to go through some
changes.

> We will share this information (design and code) with you next week.

Sounds good to me.

> Our plan is to proceed with the prototype towards a working version, which
> shows most of the concepts and their possible architecture and
> implementation, again publish the results in the Slide community.
> In parallel we will prepare the final architecture (based on the prototype
> results) and ask for feed-back (this phase should be completed until end
of
> this year). Then the final implementation will start for all associated
> layers (client API, servlet, CM Api, stores etc.). Possibly this will be
an
> iterating process.

Yes.

> We will be very happy, when additional people are volunteering to help us
in
> this big task (e.g. Jiantao Pan expressed her/his interest in testing,
Remy
> and Dierk in the kernel, etc.).

Obviously, I'll help.

I'll also be experimenting with the core if I can, to try to make it
workflow based. If I have the time to do it, it will happen as an
independent proposal (since it will be considered highly experimental; I'll
add a "proposals" directory for this).
The goals are to:
- be compatible (or at worst require little changes) with the current Slide
API from the WebDAV servlet perspective
- be compatible with the current stores
- add approvals / confirmations / preconditions, which should provide a lot
of flexibility to extend the processing of commands
- add indexing (probably using the workflow-like capabilities mentioned
above), as well as a searching interface
- be a bit faster than the current core by avoiding some redundant
operations (mainly security and lock checks)

One thing I was also considering which would require an important API change
is strenghtening a lot the notion of ownership, by adding an 'owner' field
on the node (as a property, it's very hard to access it easily from the
security layer, and it's also very hard to treat it as a special protected
property).

> We will prepare a task list (completed end
> of 2001), where people can volunteer themselves. Especially in the store
> areas (file and JDBC) we are happy for helper, but also in all other
areas.
> Does this make sense?

Yes, that's great :)

Remy


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to