Below is the transcript of a committer's meeting held on #dspace at
irc.freenode.net at 20:00 GMT on May 27th.
We intend to hold the next meeting on #dspace June 3rd.
Anyone is welcome to attend.
You can see the upcoming schedule as well as other DSpace developer
events at:
http://www.google.com/calendar/embed?src=3mfp5qsv0kejvsbh558lmshujk%40group.calendar.google.com
Meeting Summary:
We discussed the recent activity around Duraspace, DSpace and Fedora,
including the utilization of the OSL Virtual Server currently
maintained in the DSpace Community as an extensive resource for
Duraspace and Fedora SCM and hosting activity. We reviewed some of
the interesting topics in the DSpace 2.0 workshop and how they relate
to doing DSpace 1.6 development planning.
There was some reporting on work areas and GSoC projects. We did
conclude it may be wise to organize GSoC projects to be worked on in
the OSL repository rather than the dspace-gsoc SVN so that the
community can be more aware of their activity. Likewise there was an
interested expressed in some of the work projects forking brnaches for
prototyping. However, there was limited success getting updates on
GSoC projects and 1.6 work areas and we suggested putting together a
meeting schedule to attempt to focus on those topics in finer detail.
Transcript: Attached as text file.
-Mark
--
Mark R. Diggory
@mire NV USA
http://www.atmire.com
[1:03pm] kshepherd: morning guys
[1:03pm] bollini: hi kim
[1:03pm] mdiggory: Hello, looking to see if Brad makes it in, but in the mean
time I volunteered to moderate
[1:05pm] bradmc left the chat room. (Read error: 110 (Connection timed out))
[1:05pm] kshepherd: i might run over to my office pc in the meantime.. afk for
~2min
[1:06pm] mdiggory: kshepherd: you've got one more minute
[1:06pm] mdiggory: So shall we gather our topics for this call?
[1:07pm] mdiggory: 1.) SCM Update
[1:08pm] mdiggory: 2.) Recap any topics from OR09
[1:08pm] stuartlewis: 3) VERY quick update on 1.6
[1:08pm] h4xOr joined the chat room.
[1:08pm] mdiggory: 4.) Duraspace and Development Resources
[1:09pm] h4xOr: Hi, i am wondering if there is a simple tutorial on creating
Aspects using custom SQL some where on the web
[1:10pm] h4xOr: even better i am looking for a Shopping Cart aspect for DSpace
[1:10pm] mdiggory: h4xOr: welcome, we are just starting a committers meeting
for the DSpace community perhaps we can answer these shortly afterward?
[1:11pm] stuartlewis: h4xOr: Have you tried:
http://di.tamu.edu/projects/xmlui/resources/DevelopersGuide.pdf (not sure
anyone has openly developed a shopping cart yet)
[1:11pm] mhwood: An Aspect can do what it needs to do, so long as it also
behaves as an Aspect. Those two aspects of the code should be separable.
[1:11pm] mdiggory: (or answer them now
[1:12pm] stuartlewis: SCM update?
[1:12pm] h4xOr: mdiggory: sure, i`ll wait until the meeting is done
[1:13pm] mdiggory: Ok, the SCM seems to be operating stable. We have managed
to get all users in place without any complaints on performance?
[1:13pm] stuartlewis: +1 - all seems to have gone smoothly. Thanks Mark
[1:13pm] mdiggory: We now have the dspace-sandbox in place in a "modules"
section.
[1:13pm] mhwood: Agreed. I checked out 2.0 to take a look and that went well.
[1:14pm] mdiggory: There is still some cleanup and redirection to be done ont
he SF SVN
[1:14pm] bradmc joined the chat room.
[1:14pm] mdiggory: The big news (4) is that theres much interest in hosting the
Fedora SVN repo in parallel on the same server.
[1:15pm] kshepherd: scm is working fine for me, too
[1:16pm] mdiggory: A discussion arose between Chris Wilper and other Fedora
members at OR09 and we concluded that it would be fairly easy to bring thier
SVN on board
[1:16pm] mdiggory: likewise, they are converting to using Maven in their build
process
[1:16pm] mhwood: That would make sense. If the organizations are merging then
there should be more borrowing across products, and having a common "memory"
should be beneficial.
[1:16pm] mdiggory: and we can assist them in using our maven repo and rsync
process to support publishing fedora artifacts
[1:17pm] bollini: so... are you proposing to rename scm.dspace.org in something
like scm.duraspace.org?
[1:18pm] mdiggory: yes, that would be a possibility
[1:18pm] mdiggory: likewise aliasing maven.dspace.org to mvn.duraspace.org
[1:19pm] bollini: mmm... I like this but... we have just now announced the new
location
[1:19pm] bollini: mdiggory: ok alias is good
[1:19pm] mdiggory: There would be aliasing and retention of older hosts for
backward compatability
[1:20pm] mhwood: I think people need some time to get used to the whole
Duraspace thing.
[1:20pm] mdiggory: then alternative is to still use separate domains for the
two projects
[1:20pm] kshepherd: `/names
[1:20pm] kshepherd: bah
[1:21pm] mdiggory: kshepherd: its so much more. Heres the opportunity to
reduce the distance between the projects and get developers interacting
[1:22pm] kshepherd: oh yes i agree, that line from me was a botched irc
command, not a comment
[1:22pm] bollini: I like the idea of an unique domain with the "old" locations
keep as alias
[1:22pm] mdiggory: kshepherd: oh sorry, misinterpreted
[1:23pm] mdiggory: I also think we should avoid getting in a naming model where
we have to create lots of subdomains for everything being managed
[1:25pm] mdiggory: For instance: dev.duraspace.org/svn dev.duraspace.org/maven
... might keep barriers low to providing resources to the communities
[1:26pm] bollini: scm, archive, www are relevant part of the project so I think
that subdomains are fine
[1:27pm] mdiggory: happy medium then
[1:28pm] mdiggory: So, at a certain point it would be good to get a few
committers more involved with sys admin activities on the server
[1:28pm] stuartlewis: Sounds good.
[1:29pm] mdiggory: I'm looking for one or two volunteers atm
[1:29pm] mhwood: Yes, it's more important to join the products than to scoop a
lot of distinct yet related resources onto one domain. I worry slightly about
not being able to separate services later without Yet Another Naming, though I
can't come up with a good reason why we would need to separate them.
[1:29pm] bradmc left the chat room. (Read error: 110 (Connection timed out))
[1:29pm] kshepherd: i could put my hand up for that, i'm pretty comfortable
with general sysadmin work
[1:30pm] mdiggory: mhwood: the proposal ATM is to maintain separate SVN repos
under the same domain
[1:31pm] mdiggory: I think anything to ease the Maven artifact publishing
process will benefit everyone. It would not be too difficult to separate those
package doamins at a later date
[1:32pm] mdiggory: per artifact publishing... we maybe want a better model then
scp deployment of the artifacts. I'd be interested if there were a
httpd+webdav solution available that could be setup so that release managers do
not require system accounts on the server
[1:33pm] mdiggory: and possibly the scm password would double as maven release
process.
[1:33pm] mhwood: I was thinking of dev.duraspace.org/svn, /mvn vs.
svn.duraspace.org, mvn.duraspace.org. The former ties separate services to one
host.
[1:34pm] mdiggory: mhwood: thats similar to bollini's recommendation
[1:34pm] mdiggory: I'm happy with it if Brad et al. are happy with it as well
[1:35pm] mdiggory: An interesting intersection point may emerge with Maven site
generation....
[1:35pm] mdiggory: http://projects.dspace.org
[1:36pm] mdiggory: We will see a space for javadoc, module documentation and
reports emerge for the community
[1:37pm] bradmc joined the chat room.
[1:37pm] mdiggory: Well we are at about the 40 minute mark, maybe it would be
good to move onto a new topic?
[1:37pm] mhwood: Hopefully we will see documentation move into that space,
unlike some projects that have elaborate machine-generated sites with very
little content.
[1:38pm] mhwood: Did we cover SCM and Duraspace developer resources together?
[1:38pm] mdiggory: mhwood: yes I feel so
[1:38pm] mhwood: So, OR09 recap, or 1.6 brief update, or...?
[1:39pm] stuartlewis: OR09?
[1:39pm] mdiggory: mhwood: if we can get individual module maintainers to
publish their own documentation that would make life easier. But eventually we
need a more dynamic process using wiki+docbook to emerge.
[1:40pm] bradmc left the chat room. (Client Quit)
[1:40pm] bradmc joined the chat room.
[1:41pm] mhwood: It was good to meet some more of you at OR. I'm still
digesting my notes so I can find out what I heard
[1:41pm] mdiggory: OR09 was very exciting and I think the big take home was
that Duraspace is on our horizon and approaching, thus the community might want
to start to consider how we relate to one another under a new umbrella
[1:42pm] h4xOr: Duraspace == Fedora + Dspace?
[1:42pm] bradmc: Hi all.
[1:42pm] mdiggory: Likewise, there is much with the DSpace 2.0 Effort and
Fedora Communities that we should consider synergistic.
[1:42pm] kshepherd: yeah, was good to meet most of you... apologies if i wasn't
as outgoing/sociable as i could have been, first time over there and i was
possibly a bit shy
[1:42pm] mdiggory: Hi bradmc
[1:42pm] mhwood: Another way to look at it is that now I *must* know more about
what Fedora has been doing since we went down the DSpace route.
[1:42pm] bradmc: Duraspace = the merger of the two _Foundations_
[1:42pm] kshepherd: and the panel meeting at the end was great but my brain was
literally mush by then
[1:42pm] bollini: the GSoC project (Fedora DAO) will be a good starting point
[1:42pm] bradmc: It's up to the two communities to decide to get as close as
they choose.
[1:43pm] bradmc: Which would be a good thing, IMHO, but it's not forced by the
administrative change.
[1:43pm] • bradmc recovers from internet withdrawal.
[1:43pm] bollini: the dspace 2.0 data model fit well with fedora...
[1:44pm] mdiggory: mhwood: I'm think more along the lines that the communities
help inform each other and cross pollinate, that the tendancy will be to share
knowledge in the near future, not that you will need to work harder to
understand both platforms
[1:44pm] mhwood: There's value in having two products with different emphases,
and also value in being able to borrow good ideas from one another without so
much friction.
[1:44pm] mdiggory: kshepherd: as well was the panels
[1:45pm] stuartlewis: Was there an update on the duracloud system? What state
of development is it in at the moment?
[1:46pm] bradmc: The update on duracloud is that the pilot sites will be NYPL
and bioheritage diversity library, and one more TBD.
[1:46pm] mdiggory: I didn't glean a status of it yet, there is work in the
fedora repo on it?
[1:46pm] bradmc: Planning with those sites over the summer, initial
implementations in the fall, public services early 2010. No public code yed.
[1:46pm] bradmc: yet.
[1:47pm] bradmc: Process of opening code to begin over the summer. Process of
coupling code to existing repository stacks to occur in the fall.
[1:48pm] bradmc: Slide with pictures of Amazon, Sun, EMC, Microsoft, Rackspace
logos. Draw your own conclusions.
[1:48pm] mdiggory: bradmc: so should we consider a third svn repository for
that? It could be added to our support request for the fedora trac/svn stuff
[1:49pm] stuartlewis: Was there much discussion about 1.6 at OR09?
[1:49pm] bradmc: bollini: Yes, the GSoC project is promising re: DSpace /
Fedora.
[1:49pm] bradmc: mdiggory: No, it can go under the Fedora repo umbrella when
we release.
[1:50pm] mdiggory: bollini: bradmc The GSoC project is promising, which brings
up another topic. Activating the GSoC projects better so that theres more
community activity around them
[1:51pm] mhwood: What kind of activity?
[1:51pm] bradmc: OK, what's the call to action? Activating = ??
[1:51pm] bradmc: BTW, can somebody with the IRC log for the first 44 minutes
email it to me, please?
[1:52pm] mhwood: I can do that.
[1:53pm] bollini: mdiggory: bradmc: GSoC project is about DSpace 2 / Fedora or
DSpace 1.x / Fedora? or both (this could be difficult)
[1:53pm] bradmc: I think it's DSpace 1.x, as it builds on last year's GSoC.
[1:53pm] mdiggory: Well, we were planning to use the devel list for status
reporting on the project. But I've been challenged with OR09 up until now. We
need to have some kickoff activities to get developers and students interacting
[1:54pm] mdiggory: bradmc: I skyped it to you
[1:54pm] kshepherd: i've got another meeting now, i'd better run
[1:54pm] • bradmc thanks mdiggory.
[1:54pm] bradmc: Is this IRC format useful? Scheduled kickoff times?
[1:54pm] bollini: bye Kim
[1:54pm] mdiggory: GSoC project last year was 1.5 I believe. I also think some
consideration went into using the DAO prototype
[1:55pm] mdiggory: But, the mappings for something like DSpace 1.5 and DSpace
2.0 would be very very different
[1:55pm] kshepherd: bradmc: these start times are good for us NZT folk anyway
and i'm an irc junkie anyway so i'm always on
[1:55pm] kshepherd: seeya Andrea
[1:55pm] mdiggory: yet thats ok either way IMHO
[1:56pm] bollini: I'm fine with the scheduling
[1:56pm] bradmc: Sorry, meant to suggest using this IRC format for kicking off
the GSoC discussions!
[1:57pm] mdiggory: It would be a good strategy initially.
[1:57pm] mdiggory: I've got emails coming to me in private that should be
addressed by the community and mentors more collaboratively
[1:57pm] bollini: I think that the mentor should decide
[1:58pm] bradmc: Are they best redirected to -devel, or to here, or both?
[1:58pm] mdiggory: bollini: while that is true, there is some accountability
here, mentors should be sharing with the community as well.
[1:59pm] mhwood: The ML automatically keeps a record.
[1:59pm] mdiggory: ML?
[1:59pm] bradmc: = mail list
[2:00pm] bollini: mdigorry: sorry, I want to say that we need to ask to the
mentor to schedule a calendar for some IRC discussion
[2:00pm] mdiggory: the challenge is that (1) most previous projects have ended
up simple proof of concepts because they are/were far removed from the core
developer community
[2:00pm] bradmc: bollini: Yes, I think that's the next step.
[2:01pm] mdiggory: bollini: agreed because trying to get everyone together at
once is always a challenge
[2:01pm] stuartlewis: Claudia and I have a GSoC IRC chat every Thursday at 8pm
(NZ time) - (11 hours from now) each week with our student.
[2:02pm] • bradmc realizes that stuartlewis is getting me back with the 4am
thing.
[2:02pm] bollini: stuartlewis: this is a good news for me
[2:02pm] mdiggory: stuartlewis: thats great, I might recommend these chats be
posted in a similar fashion to the commiter list discussions?
[2:03pm] stuartlewis: mdiggory: Possibly not - these are often as much
'menotirng' as they are dicussions about development, so perhaps not ideal for
public dissemination.
[2:03pm] stuartlewis: mentoring*
[2:03pm] mhwood: So, how can the community help?
[2:04pm] bradmc: At the risk of creating work, I think a weekly summary might
help, even if you don't post the actual transcript.
[2:04pm] bradmc: Something to collect reactions.
[2:04pm] stuartlewis: bradmc: Agreed
[2:04pm] mdiggory: stuartlewis: I would recommend posting into the wiki for
that stuents project then
[2:04pm] stuartlewis: In previous years we made students submit weekly reports.
Shifts the effort on to them.
[2:04pm] mdiggory: we need to get the pages in the wiki organized for this year.
[2:05pm] mdiggory: http://wiki.dspace.org/index.php/Google_Summer_of_Code
[2:05pm] stuartlewis:
http://wiki.dspace.org/index.php/Google_Summer_of_Code_2009_Submission_Enhancements
already there
[2:05pm] mdiggory: Vs
http://wiki.dspace.org/index.php/Google_Summer_of_Code_2008
[2:05pm] mdiggory: stuartlewis: excellent
[2:05pm] bradmc: Sorry, have to run, will read up on rest of conversation and
summarize it later.
[2:06pm] mdiggory: So in some cases we just need to "link up" to the top pages.
[2:07pm] mdiggory: stuartlewis: In previous years I had place the IRC chats in
the "talk" portion of the wiki page
[2:08pm] mdiggory: Slight tangent... Should we consider not using dspace-gsoc
for these projects an place them in the dspace scm instead?
[2:09pm] stuartlewis: Would seem a sensible move.
[2:10pm] mdiggory: Just need to consider how we organize it I guess... Might
also be good to grab an existing snapshot of those previous years and organize
against that
[2:11pm] bollini: sorry for change the topic but I need to run, too
[2:11pm] bollini: do you know if there are any documentations available about
the dspace 2.0 demo? an how to get it up locally?
[2:11pm] mhwood: Yes, one of the problems with multiple SCM silos is that the
walls are so high. That's why I keep muttering about Mercurial.
[2:12pm] mdiggory: mhwood: walls are coming down. (I'll avoid biblical
references)
[2:12pm] mhwood: Yes, I had to dig around in 2.0 a bit to figure it out. I
*think* one moves into the top directory and runs 'mvn -Pall whatever'
[2:12pm] mdiggory: bollini: It should just be a matter of building the xmlui
war and deploying it
[2:12pm] mdiggory: its standalone and in memory
[2:13pm] bollini: well... I will try
[2:13pm] mdiggory: may need to catch up the CI on it for maven artifacts tho
[2:14pm] mdiggory: bollini: ping us with any issues/feedback
[2:14pm] bollini: mdiggory: of course
[2:14pm] mdiggory: we still need to work on build documentation
[2:15pm] bollini: good night, folks
[2:15pm] mhwood: Maybe that's something I can do, since I don't feel up to
speed on code yet.
[2:15pm] mdiggory: But the short version my demonstration only checked out the
"xmlui" project and built that
[2:15pm] mdiggory: everything else came from the maven repo
[2:16pm] mdiggory: accept the addon module I built
[2:16pm] bollini left the chat room. ("ChatZilla 0.9.84 [Firefox
3.0.10/2009042316]")
[2:16pm] mdiggory: seems we should finish up
[2:16pm] mdiggory: are there any 1.6 updates we need to address?
[2:17pm] stuartlewis: Quick 1.6 update:
[2:17pm] stuartlewis: I now have a working csv import/exort for batche editing.
Itis out with some interested parties for testing.
[2:18pm] stuartlewis: mhwood: How are disucssions going with stats?
[2:19pm] mhwood: Stats: very quiet. One vote for COUNTER compliance, one
suggestion to leverage Google Analytics, one discussion with someone having
requirements quite different from what I consider "usual", little discussion as
yet.
[2:19pm] mdiggory: I need to become more involved with the stats activities now
that OR09 is over
[2:19pm] stuartlewis: Should we offer larry / Christophe a scm branch for them
to work on controlled vocab?
[2:20pm] mdiggory: stuartlewis: +1
[2:20pm] stuartlewis: (I'm working on a Google Analytics solution - have got it
to nightly pull item view data from Google Analytics into a properties file -
now need to surface that in the UIs)
[2:20pm] stuartlewis: mdiggory: Ok - I'll write an email to them suggesting it.
[2:21pm] mdiggory: stuartlewis: I just want to make sure we are keeping modular
practices in our prototyping
[2:22pm] mdiggory: which I know is a challenge with JSPUI
[2:24pm] mdiggory: anyways, perhaps we should be holding some IRC planning
meeting around these atcivities as well?
[2:25pm] stuartlewis: Anyone heard any updates about embargoes from RRodgers?
[2:26pm] stuartlewis: Was there much talk about 1.6 at OR09?
[2:26pm] mdiggory: Thus a good reason to hold a meeting reviewing these topics
with the original volunteers present.
[2:26pm] mdiggory: stuartlewis: I did not hear much on 1.6
[2:27pm] mhwood: 1.6: not much that I recall aside from the panel. Mostly
framing it as stuff that could not go into 1.5 but should not wait for 2.0 IIRC.
[2:28pm] stuartlewis: Thanks for the update. Once we get closer to knowing what
is happening with stats and embargoes we can starting thinking about a timeline.
[2:30pm] mdiggory: AFAICR it was a limited discussion. Mostly about improving
OAI for DRIVER compatibility, there was another topic I am stuggling to recall
[2:30pm] mhwood: I've made a note to try again to stir up stat. discussion, and
make sure there's a place in the wiki for it.
[2:30pm] mdiggory: gets out note...
[2:30pm] mhwood: Communities and collections are monolingual?
[2:31pm] mdiggory: oh... there was a qestion about multi-lingual
Community/Collection fields
[2:31pm] mhwood: Don't extend DC; create your own namespace?
[2:31pm] mdiggory: yes, you triggered that neural path...
[2:31pm] mhwood: Licenses attached by reference, not by copying?
[2:32pm] stuartlewis: OAI for DRIVER - IIRC that requires us to expose an OAI
'set' for fulltext items. Not sure how we can cleanly do so unless items are
tagges as such. How would DSpace decide if an item is 'full text'?
[2:32pm] mdiggory: So we talked about 1.6 as a stepping stone for 2.0
[2:32pm] mdiggory: stuartlewis: that was an argument that started to happen...
true
[2:33pm] mdiggory: its not just full text... its assuring it meets Open Access
requirements
[2:33pm] stuartlewis: Even harder for DSpace to be able to decide.
[2:34pm] mhwood: Sorry, but I have the impression that access control has just
sort of happened along the way. Maybe that can be improved in 2.0.
[2:35pm] mdiggory: My proposal is that OAI should have its own store (lucene?
solr?) and be standalone... then updating items becomes a Consumer and the
available formats are serialized into the lucene document or something similar
[2:35pm] stuartlewis: And performance would e loads better!
[2:35pm] stuartlewis: be*
[2:35pm] mdiggory:
[2:35pm] mhwood: But then there's access control.
[2:36pm] stuartlewis: Its a little on the, shall we say, 'sluggish' side right
now!
[2:36pm] stuartlewis: mhwood: Access control to what?
[2:37pm] mhwood: Embargoed bitstreams. Decisions about what metadata to expose.
[2:37pm] mdiggory: The presence of restricted/public records in OAI would be a
consumer config on whether or not to add it.
[2:37pm] stuartlewis: In OAI, or somewhere else?
[2:38pm] stuartlewis: Having the browse system be able to take a context, and
return results acorrding to the authZ of the context would be awesome.
[2:39pm] mdiggory: mhwood: it would be upto the consumers configuration which
formats were present for a Item, this configurtion could take into
consideration Access Control, Item type, full text extracts, etc
[2:39pm] mhwood: Hmmm. One issue that's been raised more than once is that
browsing, export, PMH, etc. all have their own separate notions of
accessibility.
[2:41pm] mdiggory: I assume they will continue to
[2:42pm] mhwood: But then that leads to surprises. "I thought I marked that
nonpublic. Why can I get it anonymously through this interface?"
[2:42pm] mdiggory: TBH browse and search in the UI as mechanisms to manage the
content need to be seperated from search and browse utilized by websites for
exposure/discovery
[2:44pm] mdiggory: mhwood: the goal would be to have the application
responsible for mediating that requirement for protecting the resource, while
allowing tools like OAI to do what they do without hacking them up to support it
[2:45pm] mdiggory: thus why an Event Consumer would be written for the
application to manage that exposure rather than tearing open OAI and trying to
make it more sensitive to permissions
[2:46pm] mdiggory: seems I am not the greatest moderator... (or maybe I am
should we close out the meeting and continue the topic outside it?
[2:46pm] mhwood: Bring this topic back next week?
[2:47pm] mdiggory: We need to pull in those "stakeholders" that were in the
workshop... does anyone have contact info?
[2:47pm] kevDabu09 joined the chat room.
[2:48pm] mhwood: Drat, I didn't get names.
[2:48pm] kshepherd: back
[2:48pm] mdiggory: Lets try to bring the topic up next week (I have contact
details I can retrieve)
[2:48pm] kevDabu09: anyone here setup dspace on a mac? i could use some
direction...
[2:49pm] mdiggory: Ok, I call the meeting adjourned
[2:49pm] mhwood: 'bye all.
[2:50pm] mdiggory: have a good week.
[2:50pm] kshepherd: bye------------------------------------------------------------------------------
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, &
iPhoneDevCamp as they present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://p.sf.net/sfu/creativitycat-com
_______________________________________________
Dspace-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-devel
