I think the harvester is a great project destined for DSpace.
However, I would like to get the community to start to think about
additions like this "quite differently".  Specifically, a significant
portion (if not all) of this work should be kept in a separate module
from the dspace core.  We need to cease with the idea of "cramming it
all into the dspace-api" as the appropriate way to deploy ones
customizations.  We need to stop thinking about what we want to get
into DSpace, and start thinking about what we "want to get out of it".
(In both possible meanings of that phrase)

It is critical that the developers working with DSpace start to
understand the Services model being employed in DSpace 2.0.  Once
understood, it can be employed even in 1.5/1.6.  And the strongest
lesson concerning this is a generalization that a "Service" maintains
its own "data model" and "persistence" separate from other services in
the system.  In the TAMU Harvester case...

https://source.tdl.org/svn/dspace/branches/dspace-1_5_0-with-harvesting/etc/OAI-ORE-harvester-1_5_0.patch

... this means that, rather than altering the DSpace core
Collection/Community api to insert its data about the
"harvestability", that instead it retains that detail in its own
service api/storage/data model.  In 1.5 we introduced the Event
Manager/Consumer api as a means to complete important tasks after the
core data model has been altered, in 2.0, services can register as
consumers and  listen to other services for such events. In this
regard, one should not have to alter the core data model objects, but
instead isolate that business logic in a Service with a consumer that
is activated when updates occur.

In the last commiter meeting I made a point that 1.6 should be a
stepping stone on the way to 2.0.  How the developer community
approaches customization is a critical "change of perspective"  that
will need to occur if we are going to all reach that 2.0 goal
together. I feel that 1.6 is a critical point to be making this change
of perspective.

I will be presenting on how to utilize Cocoon 2.2 Blocks to deliver
new servlets and aspects into the 1.5.2 XMLUI. It is clear that many
in the community are not getting how to organize their add-on projects
within the DSpace maven build ecology, I hope that this presentation
will bring some light to how to approach your work in this area.

The OAI Harvester is and excellent example of where the pain points
for developers customizing DSpace arise.  If we could take the OAI
Harvester and have a serious discussion about where to draw the lines
through its patch (core changes that would be good for 1.6, core
changes that should not be there and instead, part of a separate
service, packaged in a separate dspace-harvester-api artifact, and
finally XMLUI aspects/services packaged as separate cocoon blocks.  If
we can have such a discussion, I think it would become quite obvious
the path that DSpace 1.6 would need to take.  I postulate that path is
one of "decomposition", of starting to clarify what we want "to get
out" of the core DSpace API rather than further entangling into it.

Sincerely,
Mark

On Tue, Apr 21, 2009 at 8:46 PM, Kim Shepherd <[email protected]> wrote:
> (subject changed since this is a slight derail)
>
> I've been testing this new feature as well, and it's excellent. I have some 
> things I'd like to add/change:
>
> - The ability to pass null OAI Set IDs, so entire repositories can be 
> harvested... I've already tested this one on my test server
>
> - Some better handling of dc.creator vs. dc.contributor and dc.date vs 
> dc.date.* and dc.identifier.* when ingesting the new items
>
> - More error-tolerant metadataPrefix handling.. some repos return 
> invalid/erroneous namespace strings, even if they do support a certain 
> metadata format
>
> Definitely a useful feature, and I'd really like to see it make it into 
> Dspace 1.6.
>
> Cheers,
>
> Kim
>
>> -----Original Message-----
>> From: Paulo Jobim [mailto:[email protected]]
>> Sent: Wednesday, 22 April 2009 3:20 p.m.
>> To: Stuart Lewis; DSpace Tech
>> Subject: Re: [Dspace-tech] [Dspace-devel] DSpace 1.6 - Have Your
>> Say!http://dspacesuvey.info/
>>
>> The best thing I can think is the new oai harvesting from TAMU Texas
>> University
>> Paulo Jobim
>> On Apr 21, 2009, at 1:28 AM, Stuart Lewis wrote:
>>
>> > Hi!
>> >
>> > As you'll have seen from recent emails, the DSpace community has now
>> > released version 1.5.2 of the DSpace software. It has many new
>> > features, some enhancements to current features, and some bug fixes.
>> > Many of you will also know that a small team of developers have been
>> > working on DSpace version 2.0 which will bring with it many
>> > essential architectural enhancements to ensure that DSpace continues
>> > to fulfil the needs of the user community over the coming years.
>> > DSpace 2.0 is likely to be released early in 2010.
>> >
>> > In the mean time, the DSpace committers have decided to start
>> > working on DSpace version 1.6. By moving to 1.6 (rather than 1.5.3)
>> > we can add new features that require changes to underlying DSpace
>> > database. We can't tell you just yet what new features will be in
>> > version 1.6 because we haven't decided! And that is where you come
>> > in...
>> >
>> > We'd like you to tell us which three features you would like to see
>> > in version 1.6. To help you do this, we have created an online
>> > survey at http://dspacesurvey.info/. We know nobody likes to be
>> > bombarded with surveys, so we've kept this one really short. In
>> > fact, it asks only one question:
>> >
>> > "What should be in DSpace version 1.6?"
>> >
>> > The survey has three boxes to enable you to tell us what your top
>> > three new features would be. We can then look at all the survey
>> > responses to help decide where we should devote our development
>> > effort. We'll put all the commonly requested features into JIRA
>> (http://jira.dspace.org/
>> > ) to enable further commenting and voting. As always, if you want to
>> > develop your own new features, we'd love to work with you to get
>> > those features included provided that they are in scope, or if you
>> > want to work with us on the new features that the community votes
>> > for, please get in touch! The DSpace community relies on developers
>> > donating their time and expertise to help improve the software. If
>> > you want to join in, get in touch at the dspace-devel email list.
>> >
>> > But for now, what are you waiting for? Fill in the survey...
>> http://dspacesurvey.info/
>> >
>> >
>> > (Please complete the survey by the 28th of April as we'll close it
>> > then)
>> >
>> >
>> > ---------------------------------------------------------------------
>> ---------
>> > Stay on top of everything new and different, both inside and
>> > around Java (TM) technology - register by April 22, and save
>> > $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
>> > 300 plus technical and hands-on sessions. Register today.
>> > Use priority code J9JMT32. http://p.sf.net/sfu/p
>> > _______________________________________________
>> > Dspace-devel mailing list
>> > [email protected]
>> > https://lists.sourceforge.net/lists/listinfo/dspace-devel
>>
>>
>> -----------------------------------------------------------------------
>> -------
>> Stay on top of everything new and different, both inside and
>> around Java (TM) technology - register by April 22, and save
>> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
>> 300 plus technical and hands-on sessions. Register today.
>> Use priority code J9JMT32. http://p.sf.net/sfu/p
>> _______________________________________________
>> DSpace-tech mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/dspace-tech
> ------------------------------------------------------------------------------
> Stay on top of everything new and different, both inside and
> around Java (TM) technology - register by April 22, and save
> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
> 300 plus technical and hands-on sessions. Register today.
> Use priority code J9JMT32. http://p.sf.net/sfu/p
> _______________________________________________
> DSpace-tech mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/dspace-tech
>



-- 
Mark R. Diggory
http://purl.org/net/mdiggory/homepage - Bio

------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Dspace-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-devel

Reply via email to