On 4 Feb 2004, at 05:06, Daniel Florey wrote:
Am Mittwoch, 4. Februar 2004 05:46 schrieb Stefano Mazzocchi:On 3 Feb 2004, at 06:25, Daniel Florey wrote:Hi, I'd like to open a new thread where possible features/tasks for slide 2.x or slide 3.0 can be collected.
I strongly suggest not to mention 3.0. We need to market Slide as a solid platfrom that people can rely on. Talking about 3.0 when 2.0 is not even out the door yield the exact opposite thoughts.
Sorry for the subject, I just wanted to collect all issues that didn't made it
into 2.0.
No problem. Just let's avoid talking about "when this is going to show", but rather "what" we should show. They are different concerns and should be kept separate.
This are my favorite tasks for server side slide: -> Event handling (I'm on it right now)
good
-> Performance tuning
I wouldn't call these "tuning"
- Give stores the possibility to implement macro methods in a native way
Have a look at Michael Smiths mail. This is what I intended.
...avoid the creation of many temporary objects.
can't you write JNI in your classes and go with that? what does slide has to do with it?
- Improve caching (e.g. smart permission caching/checking)
yeah, well, this requires some thinking, don't you think?
- Avoid heavy object creation
are we creating any heavy object?
ok, that's different.
I think beside the workspace concept we need some methods like begin, commit-> DASL implementation with lucene
this requires a lot of thinking
-> Enable transactions that can span multiple requests via webdav-like
methods
i thought we had that with deltav workspaces. but I'm probably wrong.
and rollback.
there is no such thing in deltav. workspaces were invented exactly to create an abstraction for that. If this is a good thing or not, well, I really don't know since I don't know of any deltav client/server that uses workspaces (subversion is using a subset of deltav but not workspaces).
One thing I wouldn't want slide to to is to extend WebDAV in a proprietary way.
This would be very fast and a real benefit for many client
applications that just want to change multiple documents in a single
transaction. Slide is transaction aware so this should not be too much work.
Again, this is what workspaces are for:
1) you checkout a workspace (begin) 2) you work on it 3) then you save it (commit)
[forgot what's the way to rollback but it must be there somewhere :-) take a look at the deltav RFC for more details]
I'd like to spend a lot of time on these things if appreciated.
-> Event driven clusterable process engine
that'd be cool
-> Overall code cleanup (e.g. logging)
-> Clearify the concepts of domain/namespace/store/scope
- In which case should the user define mulitple
namespaces/stores/scopes?
-> Import/export of data
-> Documentation
-> Demo application to show how to build complex web applications with
slide
this is all good stuff. but kinda accademic if you don't get volunteers
to write those things :-)
Then I'm not the one who's going to stop you since all these things seem really good ideas.
The only thing, as I said, is the extension of WebDAV in a proprietary way: I would be against that. Other than that, you are free to submit all the patches you want :-)
-- Stefano.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
