Archives! Two neat demos from my GSoC students
As some of you may know, Mailman had a bunch of Google Summer of Code students working on our codebase via Systers, which hired students to do improvements on their Mailman-based list infrastructure.
I was mentoring for the project which worked on the archives, where we had two students, Yian Shang who worked on the web interface for the archives, and Priya Iyer who worked on integrating search. They've produced some great demos that you all need to see, and some code that I'd really like to see integrated into Mailman 3.
I was going to forward their final summaries to the lists, but since they're quite long, here's the things I'd like you all to check out first and foremost. There's two neat demos here!
#1:
Yian's been working on a variety of UI infrastructure improvements (note that in this case, she's working at the code to support the UI, not the graphics and theming, so this may not *look* fancy but it *behaves* really well)
Her threaded conversation demo is here: http://dev.systers.org/pipermail/testing/all/conversation-3.html
and note that you can actually click through and read messages there, so try poking around those archives and see how you like them!
And you might be interested in hearing about the use cases that went into threading and quote detection that made that demo possible. Many of you may have answered the surveys I sent out on her behalf which were used to generate these use cases! She's got a whole series of posts on it, but here's one of interest: http://movicont.nfshost.com/blog/mailman-archives-ui-added-automatically-det...
#2:
Priya's search demo is here: http://lists.priyakuber.in/cgi-bin/mailman/mailocate/search.py
She'd originally hoped to find and integrate an existing search package, but licensing proved to be an issue for upstream contribution, so she wound up implementing her own. You can search a string, a phrase, wildcard searches with *, by author, by subject, and using some basic operators (AND OR NOT all work). It's got a basic spell check, though she expects that will need to be improved.
The documentation is here: http://systers.org/systers-dev/doku.php/doumentation-priya And the code is here: http://github.com/beachbrake/Mailocate/tree/master/demo/
I've cc'ed mailman-users on this email since many people there contributed to the surveys we used to develop use-cases, but follow-ups should probably go to mailman-developers!
Terri
I'm currently at UDS (Ubuntu Developer Summit) so I'm not able to look in detail at the moment, but to say I'm psyched would be an understatement! Really awesome work and kudos to everyone involved.
I'll play with the demos and provide more feedback when I get back, but in the meantime let me follow up on this:
On Oct 26, 2010, at 03:54 PM, Terri Oda wrote:
I was mentoring for the project which worked on the archives, where we had two students, Yian Shang who worked on the web interface for the archives, and Priya Iyer who worked on integrating search. They've produced some great demos that you all need to see, and some code that I'd really like to see integrated into Mailman 3.
Do you want to integrate all of the archiver and search code into Mailman 3 or just some integration code? Despite the way Pipermail's always been bundled with Mailman 2, I'm not sure that's the most productive way to structure the project going forward. In fact, I'd like to split Pipermail out of the Mailman 3 bzr tree (and started to go down that path at one point) and MM3's architecture allows for much easier integration with external archivers (even multiple ones at the same time). So it might make sense to structure the Syster's archiver work as a separate project too.
Please do submit merge proposals against the Mailman 3 trunk for the changes. If you're (meaning Terri, Yian, and Priya specifically, but also applying to anyone in general) unsure about how to go about that, guidelines for reviewable branches, or any other related issue, please don't hesitate to ping me on #mailman next week and I can help walk you through it. I'm UTC-4 and generally online during working hours.
We'll also need to get copyright assignments to the FSF. Email me off-line to get that process started, or if there are any problems doing so.
Cheers, -Barry
Barry Warsaw wrote:
Do you want to integrate all of the archiver and search code into Mailman 3 or just some integration code? Despite the way Pipermail's always been bundled with Mailman 2, I'm not sure that's the most productive way to structure the project going forward. In fact, I'd like to split Pipermail out of the Mailman 3 bzr tree (and started to go down that path at one point) and MM3's architecture allows for much easier integration with external archivers (even multiple ones at the same time). So it might make sense to structure the Syster's archiver work as a separate project too.
Right now, the archiver demos just eat Mailman mbox files for lunch (in theory they could be even integrated with a 2.1 installation) so splitting should be easy from a technical standpoint.
But there's only a handful of developers working on this, and I think the overhead, however small, is more than it's worth at the moment. So maybe we could keep them integrated for now until it's more clearly a win to split the tree off?
Mostly, I don't want my own project right now: I'm worried that I'd wind up being a blocker on something due to limited time over the next 6 months. I mean, these demos were done in August and I'm just posting them to the list now... We can revisit splitting it off when I have more time or more help. :)
Meanwhile, I'll get in touch with Priya and Yian and see about copyright assignment and getting those branches merged into somewhere appropriate in the tree.
Terri
Terri Oda writes:
But there's only a handful of developers working on this, and I think the overhead, however small, is more than it's worth at the moment. So maybe we could keep them integrated for now until it's more clearly a win to split the tree off?
What worries me is that pipermail's "integration" with Mailman has been a blocker (at the very least, sufficiently enthusiasm-sapping) for real improvements in Mailman's archiving system for about a *decade*. OK, so the new archiver/searcher is whizzy and nice, but realistically, people are going to find a million use cases that the new system doesn't do yet. Some people will prefer pipermail (if only from inertia). Some will want to use a third-party system. It needs to be easy for people to choose.
If we don't do the surgery now, when will it get done?
This is just incredible. Thanks so much!
On Oct 26, 2010, at 3:54 PM, Terri Oda wrote:
As some of you may know, Mailman had a bunch of Google Summer of Code students working on our codebase via Systers, which hired students to do improvements on their Mailman-based list infrastructure.
I was mentoring for the project which worked on the archives, where we had two students, Yian Shang who worked on the web interface for the archives, and Priya Iyer who worked on integrating search. They've produced some great demos that you all need to see, and some code that I'd really like to see integrated into Mailman 3.
I was going to forward their final summaries to the lists, but since they're quite long, here's the things I'd like you all to check out first and foremost. There's two neat demos here!
#1:
Yian's been working on a variety of UI infrastructure improvements (note that in this case, she's working at the code to support the UI, not the graphics and theming, so this may not *look* fancy but it *behaves* really well)
Her threaded conversation demo is here: http://dev.systers.org/pipermail/testing/all/conversation-3.html
and note that you can actually click through and read messages there, so try poking around those archives and see how you like them!
And you might be interested in hearing about the use cases that went into threading and quote detection that made that demo possible. Many of you may have answered the surveys I sent out on her behalf which were used to generate these use cases! She's got a whole series of posts on it, but here's one of interest: http://movicont.nfshost.com/blog/mailman-archives-ui-added-automatically-det...
#2:
Priya's search demo is here: http://lists.priyakuber.in/cgi-bin/mailman/mailocate/search.py
She'd originally hoped to find and integrate an existing search package, but licensing proved to be an issue for upstream contribution, so she wound up implementing her own. You can search a string, a phrase, wildcard searches with *, by author, by subject, and using some basic operators (AND OR NOT all work). It's got a basic spell check, though she expects that will need to be improved.
The documentation is here: http://systers.org/systers-dev/doku.php/doumentation-priya And the code is here: http://github.com/beachbrake/Mailocate/tree/master/demo/
I've cc'ed mailman-users on this email since many people there contributed to the surveys we used to develop use-cases, but follow-ups should probably go to mailman-developers!
Terri
Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://wiki.list.org/x/AgA3 Security Policy: http://wiki.list.org/x/QIA9 Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/sandhills%40gmail.com
participants (4)
-
Barry Warsaw -
Beth Morgan -
Stephen J. Turnbull -
Terri Oda