As some of you might already know, my next step, which has had me bouncing off the walls for the last month, is to join the great folks at Collabora Multimedia working on the PulseAudio sound server. I’ll be working from home here, in Bangalore (in your face, 1.5-hour commute!). It is incredibly exciting for me to be working with a talented bunch of folks and actively contributing to open source software as part of my work!

More updates as they happen. :)

The intimation about doing this was sent at 2357 hours today (yesterday, to be precise) to all. The notification basically stated that because of “undesirable activities”, Internet will, with immediate effect, be disabled from 0000 to 0600 every day. And that’s it — poof. The hostel network is disconnected from the rest of the Institute, thus making sure that nobody can access the Internet (or even the Institute’s own computing facilities). To compensate, the Computer Center (with a capacity of <200 computers) is to be kept open 24×7.

Of course, this was unacceptable, so a bunch of us decided that something needs to be done. There are 2 issues — the decision, and how it was implemented. While the decision itself needs discussion (more about this later), the implementation is of immediate concern. People were not prepared, and work on several people’s theses were affected. Plus, this has been done just a little after the end-semester exams, when most students are not on campus. This sort of fascism usually rears its head under precisely these circumstances. We decided that what needed to be addressed right now is the implementation — the Internet has to be made available this night.

A couple of our student representatives spoke to the Dean of Student Affairs (the DoSA — the official channel between the students and the administration). The DoSA basically said that they, the various Deans and the Director (and Deputy Director?) have made the decision at nothing would be done about it. More precisely, the Director, as the highest power in the Institute has taken the decision and that’s that. Further discussion may be taken up with him.

About 60-70 of us went to the Director’s house at about 2:30 (the entire process was peaceful — there was not shouting or slogans). We met with the security, who called the Head of the Computer Center (CC) and the DoSA to the place after some attempted dodging.

The CC Head turned up first and started asking what our problem was. He offered such resources as a vehicle to transfer us from hostel to CC as well as as many pen-drives as we require to transfer data from our machines to the CC machines. The DoSA just said that we’ve given you 2 years to think about whether this should be implementing it, and now we will be implementing it, so there.

Our student representatives (who did a pretty good job), after some dialogue, got the connection reinstated for tonight. They will be further taking up the issue later today.

The decision itself is extremely foolish, of course. Moreover, the dictatorial way in which this is being done is just as shocking. Let’s see how things pan out in time. Perhaps sense and sanity will prevail.

In addition, pkgcore 0.4.4 has my patch to support HTTP proxies for rsync. This was a fun patch to write, small as it is. The code is beautiful, and Brian Harring (ferringb) and Patrick Lauer (bonsaikitten) walked me through a lot of it. Good stuff!

I’ve also been working on splitting the gnome-python* ebuilds to make the dependency trees for packages that use these bindings a lot saner. This has been longer and more painstaking that intended. It wouldn’t even have happened if Jim Ramsay (lack) hadn’t made an excellent start with the gnome-python-desktop split, since all subsequent work was based on that. Hope this is useful to someone, though. :-)

As I said, a good week.

And late as they may be — the photos!

We started with Piyush continuing the KDE track, this time for developers. It was a good talk — the quick ‘n dirty PyQt demo was the icing on the cake. It really showed people how easy working on desktop apps can be.

We had a BoF session in the afternoon that rotated around Open Standards and Open Formats. We had about 20 people there, and after a 5-10 minute intro, we got around to a pretty interesting discussion. We started talking about whether open standards are really a “good thing”. The general consensus was that companies should be able to make money from their products, but vendor lock-in is bad. People agreed that the Adobe’s PDF model of keeping the format open and making money off the tools seems to work well. We then progressed to talk about how we can impact open format adoption. We settled on:

• Use open formats
• Accept only open formats in communication, to what extent is possible
• Get people to use open formats
  • Start with people who are willing to experiment and change
  • Then move to people who don’t care, and show them how open formats are better for them
  • Keep trying, no matter how bleak it looks
• Keep bugging companies that use closed formats to open them

The discussion was lively, and went all around the board — good fun.

Finally, Chaitanya Gupta of Cleartrip.com gave (actually, is giving) a talk on Common Lisp. People are sitting in there instead of standing in the queue to listen to Strings, so I’m guessing something is going right. Grin

Oh, and did I mention that we distributed 200+ Ubuntu/Kubuntu-KDE4/$distro_of_choice CDs just today?

So it’s been a fantastic four days. Neither the weather nor colds, sore throats, and $illnesses could get in the way. We had awesome speakers who came here against a whole bunch of odds (thanks Shreyas, Ankita and Piyush!)

Saurabh Nanda made it possible of course, to arrange the event at all.

And of course all the volunteers out here kicked butt. They tirelessly did stuff late into the night if it needed doing — designing and putting up posters, waking up early (or not sleeping) to get to the airport in the morning, and on and on.

A very special thanks goes out to Atul Chitnis — our event would’ve floundered somewhere right near the beginning. He guided, helped, prodded, pushed and made FOSSKriti what it is. And there’s Sankarshan Mukhopadhyay, Pradeepto Bhattacharya, and Runa Bhattacharjee who supported us from the moment we told them what we were planning to do. Thanks, guys!

So things are winding down, and I’m off to pass out in a ditch somewhere. :-)

[Rohit has to get photos up (put ’em up, dammit!), and I really really hope we can get videos of some of the events (update soon)]

We had some trouble starting up the machines, but our awesome participants stuck it out till we got things hammered out, with an amazingly patient dBera helping a lot. Things went on till 4 a.m. (at which point we needed to get just a little shut-eye (a few didn’t sleep at all) to prepare for the next day). At this point, we submitted a Firefox Beagle Search Bar written by Jai Kumar Singh (a.k.a. flukebox). In a day or two, another group should be able to submit a patch for a BibTeX filter. And there’s another project that I hope to talk more about in the days to come.

Yesterday (Friday) morning, we had Shreyas Srinivasan delivering a talk on building sexy UIs with Clutter. We had about 200 people in the hall, and it was a huge success. How do I know this? There were a lot of questions after the talk and they were almost all relevant and insightful. Afterwards, a bunch of people even got in touch with Shreyas about getting started on contributing. ‘Twas good!

Today started with Ankita Garg delivering a workshop on the Linux kernel. The lab was packed, and we had to turn away a huge number of people. A lot of the crowd was too new to Linux to appreciate what was happening, but some people took away what was intended.

This was followed by Piyush Verma’s talk on KDE4 for users. This went well — good turn out, and people learnt about KDE, asked good questions. He’s going to be talking about KDE 4 from a developer’s perspective. Looking forward to that. :-)

Next up was Ankita again, with a talk on LinuxChix. We had expected a large turnout but had about 13 women and 20 men. But the crowd was interested, and we had a good discussion after the talk covering the LinuxChix organisation, and how it can be useful in the context of the audience. Ankita’s got the list of stuff that came out of the discuss so I’ll link to it when it’s out.

Since we ticked off a whole bunch of IITK people by turning them away from the kernel workshop (we decided to keep only the non-IITKians). So Ankita very awesomely agreed to take another session. And we had a packed lab again. This time, the crowd was a lot more savvy, and things went better. If only we’d had more time, we could’ve had a longer session with more questions. But bravo to Ankita for talking for 6 hours today. Anyone who’s delivered even an hour-long talk will know what a feat that is.

As the finale for today, Saurabh Nanda and Chaitanya Gupta of Cleartrip.com are conducting a Ruby on Rails workshop, followed by a hackfest. We’ve got the CSE department’s largest classroom full of interested and engaged people, so we’re happy to go as late into the night as we need to. :-D

So that’s how FOSSKriti’s been going so far. Tomorrow’s the last day. It’s been legendary! (I told you so …)

[photos soon]

The workshops are on Ruby on Rails (a hackfest will follow the workshop) and Common Lisp. I was around when these guys were planning out the workshops, and their energy and passion when it comes to RoR and CL is incredibly contagious.

So not only are these guys making FOSSKriti possible, they’re helping us make it kick even more butt!

(p.s.: more info on the CL workshop and RoR hackfest will be out soon)

What’s more, we’re having a Beagle Hackfest too! We’ve collected a set of easy-to-attack tasks that have cropped up on the mailing list and IRC in recent times. It’s been a long time since I was around a Beagle/Dashboard hackfest. I still remember the first time I got all excited about this stuff, back in the day. :-)

The schedule’s out, with a couple of surprises still to be sprung. If you’re going to be dropping by, drop me a line here!

More updates soon …

That’s right. This year, IIT Kanpur’s techfest, Techkriti, is going to host FOSSKriti — our very own F/OSS mini-conference. We’re having talks, BoF sessions, workshops and maybe a hackfest or two. It is going to be legendary! More updates in the days to come.

Nirbheek and I are going to be trying to arrange a Dashboard/Beagle hackfest too. Let’s see how that goes. :-)

So if you’re coming to Techkriti, do drop in!

With that done, and a little cheating (i.e. not really doing any parsing on the userQuery), Mikkel’s simple-ui and xesam-adaptor now talk to each other. Whoopee! There’s still a bug I can’t seem to pin down which crashes the adaptor after 2-3 searches from the UI. Will look at that soon. Instructions on taking these out for a whirl are also up.

This is what it looks like:

Well, I’ve got some code up and running. It doesn’t install anywhere yet, but does talk Xesam in it’s own limited and special way. The parser is possibly not too efficient, and definitely has trouble parsing more complex queries (ones with multiple or data elements). You can find the code is in the Gnome subversion repository.

There’s a list of outstanding issues in the top-level directory. For example, a major concern is that there isn’t much of a grouping mechanism in Beagle for queries with lots of ANDs and ORs (something like “(a and b) or c” is not possible right now).

In the initial implementation I left a lot of the smaller details hanging, I’m going about addressing these now. The set of supported metadata fields needs to be finalised before I can touch upon the set of fields returned and sorting order. The parser is also could do with some touching up (I’ve kept it as simple as possible now).

All this should push the adaptor closer to completeness.

GetProperty is supposed to return a so-called “Variant” type. Depending on the property requested, the method might return a bool, an int, a string, or even an array of strings. Now I’m not sure how the C# bindings for DBus (dbus-sharp) handle this. Intuitively, returning an ‘object’ (the class from which all types are derived) should work, but it doesn’t seem to. Will need to dig deeper, since this might just be a problem with my version of dbus-sharp (0.63, from Gentoo’s gentopia overlay), and speak to the creator of dbus-sharp (Alp Toker) about it.

It feels good to be writing C# code. The focus is purely on solving the problem at hand, rather than small, nagging, ancillary tasks. For example, setting up the Xesam listener is as simple as:

public static int Main(string[] args)
{
    Application.Init();


Connection conn = Bus.GetSessionBus();
Service xesam = new Service(conn, "org.freedesktop.xesam.searcher");
Xesam.Search search = new Xesam.Search();
xesam.RegisterObject(search, "/org/freedesktop/xesam/searcher/main");

Application.Run();

return 0;


}

Many thanks to jonp and mdx4ever on #mono for their patience with all my n00b questions. :-)

Fun, fun, fun! Good night!

I will be working on Beagle, an uber-cool desktop search tool. The idea behind such a tool is that most of us today work with a massive amount of diverse information everyday. Often we find ourselves trying to remember the answers to questions like “where’d I put that email about …”, “what was that site I saw on …”, “what was that phone number X gave me on IM”, and “where the !@#$ did I put that file …”. Beagle tracks all this and more information, and makes is very simple for you to find. If you haven’t used it, or a similar tool, I highly recommend it, especially if you do a lot of work on your computer.

I’ve posted more details on what I’ll be working on, and why at http://beagle-project.org/BeagleXesam

Phew! Comments, feedback welcome.

Personally, this is great for me, having a focused task to work on and being able to contribute to an open source project. I’m also excited about working with Mono/C#. This might seem strange, given my reservations about a Microsoft-controlled platform. You have to see the code to understand how much cleaner Beagle looks than an equivalent C/C++ implementation would.

Once the semester ends (2 weeks), I will dive into the code and familiarise myself with it.

Update: It seems that I forgot one thing. :D

Time I read Slaughterhouse 5. Currently on Margaret Atwood’s The Handmaid’s Tale. Time, it is short, as the semester end nears.

And it appears that I will be busy this summer.

Not much else that is spectacularly worthy of reporting. Now to read about routing protocols in mobile, wireless, ad-hoc networks and associated security concerns.