Hi,

[Felix Röthenbacher wrote]:

> I think, a JSR-170 repository can be much more adapted to the needs
> of Lenya -- or generally, a CMS -- than a filesystem. JSR-170 offers
> the tools to support a CMS storing and retrieving its content in
> an easy, standardized way.

I agree!

> Please let me know what you think about the draft moving
> Lenya away from the filesystem-centric approach to a more
> powerful usage of the possibilities JSR-170 offers.

+1

[Andreas Hartmann wrote]:

> Do we still want to support other repository implementations,
> especially the current source-based approach?

I don't really see the need. All the world is heading towards JCR, and
this is what JCR has been created for.

> If not, should the switch to a JCR-only approach happen in 1.4,
> or rather be deferred?

What value would a 1.4 release add in its current state? There has been a
lot of internal cleanup between 1.2 and the current 1.4, but what exciting
new features could we offer to our users with a 1.4 in the current state.

Even worse (thinking a bit in marketing terms): 1.4 is late anyqay. If it
comes out now without the long awaited repiository integration (I think
lots of people are awaiting this) we might disappoint the audience. My
opinion.

> If yes, do we want to require mirroring the functionality JCR offers
> in the other repository implementations, or should we introduce
> somthing like compliance levels (-1 for complexity reasons)?

-1 (I think it isn't worth the effort)

> (we can drop our "lightweigt repository layer" or reduce it to
> a mere facade to JCR).

+1 (drop it)

> How about the JCR block? It looks like it won't help us, because
> the whole source-based approach would be dropped, wouldn't it?

I agree. I played a lot with it recently and found that

a) I don't like some of the design decisions they made.
b) It forces you back into a file system metapher.
c) It is somewhat incomplete.

I understand it had been created over in Cocoon land to serve very
specific needs and I am quite happy that they made it available, but I
suspect it might get replaced or at least undergo quite an amount of
development (hopefully).

See: http://www.mail-archive.com/[email protected]/msg32862.html

In its current state, it's fine to access single elements from a JCR
repository, but it does not really offer a lot of value when it comes to
more complex content retrieval IMO. In other words: It will do a good job
if you have a site defined by other meands and you want to pull in very
specific pieces of content from a repo. But that's it so far. They also
keep discussing about the question what protocol to use to access content
from different repos, i.e. should it be jcr1://..., jcr2://, jcr:repo1://
and jcr:repo2:// or jcr://repo1/... and jcr://repo2/... .

Summa sumarum: I am not sure the JCR block (in its current state) is the
solid fondation for Lenya's core operations.

Having said all that, there is probably one question left:

What contract do we intent to implement between the sitemap and JCR in
Lenya? At some point in time, a Generator needs to access content from the
JCR repository.

Regards,
Torsten


> Felix Röthenbacher wrote:
>
> [...]
>
>> Please let me know what you think about the draft moving
>> Lenya away from the filesystem-centric approach to a more
>> powerful usage of the possibilities JSR-170 offers.
>
> In theory, it would of course be great to leverage the full
> power of JCR.
>
> Some open questions:
>
> Do we still want to support other repository implementations,
> especially the current source-based approach?
>
> If not, should the switch to a JCR-only approach happen in 1.4,
> or rather be deferred?
>
> If yes, do we want to require mirroring the functionality JCR offers
> in the other repository implementations, or should we introduce
> somthing like compliance levels (-1 for complexity reasons)?
>
>
> IMO the recent experiences with JCR suggest that the integration
> is rather simple and straightforward, so I'm already in favor of
> focusing on JCR only. This will certainly increase the power of
> the Lenya core (queries etc.) and even make things much easier
> (we can drop our "lightweigt repository layer" or reduce it to
> a mere facade to JCR).
>
> How about the JCR block? It looks like it won't help us, because
> the whole source-based approach would be dropped, wouldn't it?
>
>
> -- Andreas
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>


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

Reply via email to