I'm sorry, I just received a late update (at great cost) from Anthony Baxter.
The last keynote from Linux.conf.au was about the upcoming Python "We'll break all your code" 3k. Perhaps not quite keynote material, but it covered a gamut of issues which will break old code when the move happens - and he did work for a colorful company. The set & dict comprehensions, function annotations and dictionary views are probably worth the terrible loss of reduce(), my favourite companion to map(). And obviously old style classes & string exceptions were excess fat to be trimmed anyway. But there was more interesting code to test.
>> from __future__ import braces File "<stdin>", line 1 SyntaxError: not a chance >> import this
And check the output twice.--
I believe a little incompatibility is the spice of life, particularly if he has income and she is pattable.
-- Ogden Nash
Occasionally, I do wonder why I do what I do, especially with software. The answers I've got with might not make sense to everyone - actually, it no longer makes too much sense to *me* today. But one thing's remained the same - it wasn't the money then and it isn't money now.
Everyone admits that more money helps (talks, walks you home and even buys you dinner). But the question at hand was whether money, more accurately, the presence thereof, hurts open source projects. And Stormy Peters began with that question, into her LCA '08 keynote.
The internal motivations to write F/OSS usually are weakened when there's greenbacks flying. Not to mention money brings in the people who herd it, from the clan of MBA (not to be confused with N.W.A). The processes and rest of the beasts of communication, reports and roadmaps which follow aren't necessarily bad, but they aren't "fun". But even when hacking hasn't been instutionalized, money is a selective reward, which splits the community into those who are paid and the rest. Dunc-Tanc is a classic example of this divide in action. And I'd personally lost that rush the moment I start equating efforts to money - there's a stress level which you can't pay me to take up, but in which I used to live day in & day out.
Stormy Peters takes on the topic from the point of view of someone actively involved in funding open source projects & developers. Occasionally, I run into clearer insights than mine, but conclusions I gathered from the talk are something to a closer look at.
Once a f/oss developer, always a f/oss dev. Even when you've gotten off the rush for some project, pushing it along as a good engineer, the hacker in you doesn't sit around. So, even if you don't do the same for free, it's very likely you'll do something else for the same reasons you did it first - look at the ex-Eazel folks. Anyway, history might just rhyme or pun, or maybe people just go do the same again ( *cough* Gnome Online Desktop), but whatever they already did - is not lost.
Anyway, to the objective reader, I'd suggest that you ignore my blatant bias in the above post and go watch the video.--
It's classic percolate-up economics, recognizing that money is like manure: It works best if you spread it around.
After all the dust from the miniconfs settled down, the real conference began with a real treat - Bruce Schneier talking on Reconceptualizing Security [video (~150 MB)]. The keynote was about dealing with security as an intangible feeling as well as the reality of security.
Lemons: But before I head off into what that talk was about, let me introduce a new concept - Market for Lemons. The idea is that when the quality of a product (like a used car) cannot be assessed easily, the bad car (i.e a lemon) would sell for a profit, while the good car wouldn't have any customers because it is expensive. The scenario is exactly tilted towards the dishonest seller, in the absence of proper information and/or blacklists for lemon peddlers.
The security software market right now is such a one (I'd say that the job market is too). Products which are pure snakeoil are completely indistinguishable from real products from a brochure or presentation. As was clearly illustrated today. The company which invests more efforts in creating the illusion of security would potentially win over someone who has been actually spending time & effort to secure their product. The price war cannot be fought between good & bad without enough in-depth information about stuff under-the-hood.
There is probably a sweet spot between real security and the illusion of security, which would be profitable. It wouldn't be bullet-proof security by any stretch of imagination, but the sense of security wouldn't necessarily be misplaced. Which is what the profit-oriented capitalist economy would gravitate to, provided there was enough awareness & information about what is actually in the box.
Q&A: I actually got to ask Bruce Schneier a question - about the meta-stable state of this balance, where every new development/marketing tactic would again upset the market off this point. He admitted that it was an interesting question, but suggested that it would probably dampen into insignificance (think about soaps & detergent brands).
Anyway, there I was, communicating to Bruce Schneier without encyrption.--
There's small choice in rotten apples.
-- William Shakespeare, "The Taming of the Shrew"
Monday: I had this illusion that I had no jet lag after reaching melbourne. So, I headed out to the security miniconf. The WiFi wasn't up yet (it never is for the first day). The couple of talks I was really focussing on was the OpenSSH one and the 'aggressive' Network defense one.
The latter talk had some interesting 'facts' about ICMP. The ICMP host redirect used in the sense of "talk to the hand" packet dumping. Also suggestions for an ICMP quench message to be un-deprecated as well as a lot of ire directed at sysadmins who just drop ICMP at fire-walls.
The OpenSSH talk had a fair amount of detail about writing secure code and more importantly retro-fitting security into existing code. Ironically, it began by quoting a Microsoft research paper as inspiration for a lot of work. Moving on, it shows how different is sshd compared to the standard unix daemon. Rather than move ahead accept-fork model followed very successfully by apache, sshd uses accept-fork-exec to run a master and slave process. The master handles all the higher privilege functions while the latter handles the protocol/crypto/book-keeping functionality.
Tuesday: I slept off till late afternoon. Woke up in time to peek into the gaming miniconf and walk around between Lygon st. & Swanston st.
Still not adjusted to local time yet, but the extra coffee seems to be helping today.--
So little time, so little to do.
-- Oscar Levant
LCA '08 beckons and so does a fairly long vacation down under. The blanks left behind have been filled up and I'm sitting in Singapore right now and as usual this trip was also run quite close to the wire in terms of scheduling. But I'm here and on my way.
I'll be reaching MEL by Sunday night (yes, sleepover at the SG transit hotel) and then the week goes into the conference. After which, I plan to hop onto a kangaroo and head off along the coast to Brisbane (that's about as clear my plans are).
If any of you feel like putting up this low-budget traveller somewhere on that coastal route, ping net dot php at gopalv (in reverse).--
Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell.
-- Doctor Who, "Androids of Tara
Unlike last year, this year I had a pl_n.
It's a little sketchy now, but looks like I've got cash for the tickets, without doing anything too extreme .
Post conference, I've got about a week to get from Melbourne to Brisbane via some sort of ground transport, along the gold coast. Hopefully, I'll be able to find some company to drive and share gas money with. Otherwise, I need to figure out how public transport can take me to point B from point A possibly passing through point Sydney.
But before all that, foss.in beckons.--
Life is what happens to you while you're busy making other plans.
-- John Lennon, "Beautiful Boy"
As I sat around in the keynote today, listening to Kathy Sierra go on and on about users, bits of Jono Bacon's and Zonker's talks popped up. You could classify projects as belonging to three categories - scratch your own itch, benign neglect and user centric. Most library and system developers tend to belong to the middle category, where their users are actually developers and the project waddles along according to the core dev's plans.
What that keynote actually did for me was to rationalize the use of some emotional play into a proper strategy targetting users rather than consider that an insult to their intelligence (or rationality). Well, I'd rather term it more generally as acceptance of their sensitivity and essential humanity, now. The three talks combined have given me a bit of things to think about, about the mistakes I've made over the past couple of years in handling dotgnu. Until we threw up the Southern Summer of Code, the project was literally stuck in a rut.
The talk dealt with the suck zone of doing anything, which makes a lot of sense to me. The hardest part to do is the first few bits, which everybody knows. But what was hammered in was the message of hope and of course, pride. The concept of levels and a set path upward for someone who comes in is pretty good, especially for a slightly large community - sort of mirroring Jono's breaking and entering a community comments (well, he didn't say that - but that's sort of what he meant, I think).
Happiness is a compile away !: The other thing that she talked about was about flow and meaning. I can pretty much relate to the first, which I prefer to call deep hack. And to conclude, a bunch of jokes - ending it with a high note after the heart pullers (like the puppy and the baby photos) in the middle. I absolutely loved Chris Blizzard as Volverine - a good movie that was ... oh, wait.
"This isn't right. This isn't even wrong." -- Wolfgang Pauli
I've said this before and let me say this again - we need to let people play around with stuff. The freedom, stability and all the other advantages of F/OSS do not appeal to the layman - you've got to get him interested by other means, if only to start with. And IMHO that's exactly what LCA Open Day did.
There were people of all ages, sizes and interests around. And quite a lot of interesting stuff, even for seen'em all folks like me. For instance, there was this cool motion detection and tracking tool, projected onto a huge screen. I realized what it was when I walked in and saw a few stars move exactly like I did. As I stepped back, those exploded and were replaced by others tracking the other folks coming in. Perfect eye-candy which can keep people interested in the show (which it really was). But what really caught my eye was the MythTV demos, running on some sw33t spankin' hardware.
I think it can be called a total success in terms of attendance and crowd pulling ability. But it sort of pokes fun of my previous post about LCA not being an evangelism event - I don't mind, really.--
The problem of leadership is inevitably: Who will play God?
-- Maud'Dib, Dune
I've been asked this question a couple of times today. Whenever I mention that I've flown in from Bangalore to attend this conference, someone does ask again - What brings you here ?. Considering the loopy hoops I've had to jump through to get here - in fact, not having an apartment to go back to in Bangalore - I've asked myself that question more than a few times.
Why LCA ?: There are a lot of conferences around the world. Some of them are commercial, or rather of a more corporate flavour - which basically covers a fair bit of them. Then there are conferences focussed on evangelism, which I'm not really into, can't sit around teaching people stuff. And then there are those organized for users, like our grand old install fests. But in between all these, right there in the sweet spot of developer heaven is LCA - developer stuff and literally nothing else.
Why this year ?: Now, that's a really interesting question. I had to think a lot about that. First up, last year had been awesome - I was totally transformed when I went back to India. I was sort of in a pit, in a lot of ways, somewhat like this year. But that whole week at Dunedin sort of put the spirit back in me and I came back raring to code out all 8"sorts of things - which I did, even slayed that bug from hell.
But three days into the conference, I'm still not even at level zero. It is the little things isn't it ? The laptop I had borrowed from yahoo has a crap hdd and is randomly destroying bits of my data - photos mostly. Basically a b0rked bit of hardware is an irresistable challenge and wasted a fair bit of my time.
And the other fun thing was meeting new people. Except this year, half the people I met were people I already knew from last year. Being split up into multiple hostels doesn't help that and neither does the lack of a common room in New College.
I've still got two days left to go and probably the cricket match on saturday. All of which and the rest of the week of not coding should push me out of this coding lull (no code since mid-nov, zOMG).
*fingers crossed* ...--
When in doubt, mumble;
When in trouble, delegate;
When in charge, ponder.
-- James H. Boren
First up, this blog has been delayed. The wifi network at the conference hostels weren't up yesterday and I had better things to do than write blog entries in the pavilion.
GOODMORNING FREEDOM LOVERS: Every morning, for all of last week, sometime around morning sydney time (which is 3:30 AM for me), jdub arrives on #linux.conf.au with this message of freedom, love and well, morning.
Only Speakers can heckle: Jdub went into the introduction full on. Starting from the photos of the team - who have managed to juggle with a colossal budget of 500,000 AUD. And with some embarassing pictures, we headed on to tackle the other special things about this LCA - which have nothing to do with each other.
What if distros were sofas ?: And then it got into a bit of distro bashing - with debian and gentoo taking the brunt of it. Debian got a couch without a back, which doesn't give (*wait for it*) much support, but if we as a community all sit together and lean back, we can support each other. And well, gentoo got an IKEA chair.
Embedded miniconf: I attended the Qtopia greephone talk, firstly because Rhysw works on that and partly because I wanted to see how QCop stuff works. The talk was interesting and they actually managed to get the demo off coding live.
Slacked off most of afternoon and headed off downtown in the evening - more on that later.--
The clash of ideas is the sound of freedom.
I got my tickets to Sydney confirmed on Thursday evening, for an extra 200 USD, making it a round 1400 USD. Right now I'm sitting in the lobby of Changi airport, literally chained to the power plug by an office laptop with a b0rked battery. It must've been the lack of sleep, jet lag and the location in combination, but I found myself reminscing about my trip last year (without the aid of my blog).
No, it isn't the talks that I remember about the conference - well, with the possible exception of the Damian Conway keynote. Other than the thrill of my first foreign excursion, the conference didn't quite have the sterility of the average technical conference - wasn't a single lecture in what I attended.
But the people were more interesting than the talks. And I think the 24x5 nature of the conference, because of the hostels, plays a big part in encouraging interaction. And when such a large number of interesting people are thrown together, stuff's bound to happen.
And then there are the conversations - I remember one which started off about Women & FOSS (remember D00d3tt3z on ILUG-d ?), which turned into a discussion comparing unicycle hockey with unicycle polo.
As someone said - couple more sleeps till lca '07.--
A conclusion is the place where you got tired of thinking.
The LCA submissions have been rated, reviewed and the letters sent out. And it shouldn't come as a very big surprise that my talk hasn't been selected - here's to all those who said "Yeah right, save your talks for real conferences". I can't honestly say that I'm not disappointed, mainly because there is no point in talking about libjit in 2008 - it would have crossed over the bump into old & busted. And I'm definitely not going to submit a talk to conference which I haven't attended - I'm too fragile to go through an audience mismatch again.
But on the other hand, it is a testament to the quality of conference. if my talks aren't good enough, that sure is a conference which I shouldn't miss - which brings me to the real reason why I'm depressed right now.
My LCA 2006 trip in total cost around 1.4 Lakhs (~3200 USD) - of which dotgnu chipped in 1700 USD, Yahoo! chipped in 600 USD and I pulled the rest of the ~1000 USD. But this year, I see no reason for DotGNU to pay anything, because most of my efforts have gone into APC rather than the usual creative energy vented out on dotgnu. Since I have switched teams & manager, I think Yahoo! will wash hands clean, especially after what happened to spo0nman.
Since I'm not getting a raise this year, despite complaints, I don't think I can build a slush fund to fulfil the extra ~2000 USD (7 months' savings) - but even for the 1000 USD part, I'll probably have to give up a few luxuries. There's a bit of me that says I should be more optimistic, but deep inside me I know why I'm not.--
Disappointment leaves a scar which ultimate fulfillment never entirely removes.
-- Thomas Hardy
The conference had practically ended by saturday morning. Mark Shuttleworth was giving the keynote and we all sort of headed back after the keynote. I met up with Andrew Tridgell and just listened to what he was saying. I think we need more people like him who will talk quietly but will not be found wanting when the call goes out to all good men to stand up to things like Bitkeeper. He was saying some interesting about mercurial - which is an up and coming distributed version control system.
During lunchtime was when all the hair promised during the last night's auction was sort of collected. First up was Dave Miller, whose beard was on the chopping block. It was a Win-Win because we got to look at a half-decent looking Dave and Dave got a shave for free.
Rusty showed up with his own razor and sort of lived up to - If you gotta get something done right, do it yourself. The world has never seen Rusty without his mush since he was sixteen - except for that vacation where he destroyed all evidence by dropping the camera into sixteen feet of water. Maddog was keen to capitalize on the photo opportunity.
Rusty does the job and walks around with a new mush. I remember hearing Jeff yell out now you look even more like a pornstar as soon as Rusty put on the falsie.
Jeff Waugh was up next. I don't think I need words to explain the story below. Reminds me too much of cartoon story boards. Well, provided you know the characters in the story here - Jeff & Pia Waugh supported by Maddog.
Biggest thing post lunch was the fact that I dropped my HDD and had to go through some mental anguish. Then attended the panel discussion which essentially had a small flame war about Mac OS X vs FOSS. The future of the desktop seems to have been hijacked mid-way by OS X and well seeded in the geek community by Apple. Tridgell replied with "maybe you shouldn't pick your software by merely quality alone".
Then the next year plans were announced and a few jokes by the organizers - like the Jeff and Pia Smith reference (Pia Smith is now Pia Waugh) and the Scary bald man Jeff.
I then headed out to dinner with the gang. Noticed that Rhys's jacket had the remanents of whatever was Oracle's Power Browser project.
Then came the time to say good bye. Even those who were staying in New Zealand were headed out onto the road. Unicol was feeling rather empty - and the drained feeling that a conference brings when all the adrenaline dies down.
I still had a day left to see around Dunedin, but you'll have to wait for the post - I have so much to write about.--
Don't worry about the world coming to an end today. It's already tomorrow in New Zealand.
Bright sunlight pours through the heavy curtains into my room at Unicol. I woke up thinking "Oh, shit ... I'm late ", scrambled over to the desk to check the time. The time on the clock says 06:20 AM. After the usual morning struggles, I get down to the cafeteria by 07:30 to meet Rhys and Ian. Since the embedded miniconf has been postponed to Tuesday, we decide to head into the debain miniconf and see what's cooking in there. Debian has always interested as a social experiment rather than the actual technical results of the project as such. Of course they're a fun bunch of guys as I soon found out. Since the speakers were a little tardy in getting to the conference halls, Jon Oxer started on some parallel activities :D
After making it possible for someone to literally do apt-get install klue, Jon handed over the proceedings to the various speakers. One of the interesting things that came up was about Computer Angels, during Cameron Patrick's talk about building your own custom distro. I'll say that the problem with linux advocacy is follow up work, which these guys seem to have found a way to achieve. It is easy to get someone to install it, the hard part is going back to fix what they need months after the first install.
After the whole thing settled down, we were supposed to head towards Skwashd's house for dinner and dessert. I made the mistake of taking an afternoon nap and ended up missing most of the dinner. Dave picked me up from Castle road and I did get to eat the rest of the sour cream, ice cream mixed with oreos.
Then after a while of chit chat and discussions about religions, skwashd's previous car crash (including photographs of a car on its roof on the other lane) and dotgnu in general, we all headed back to the hotel on foot.
The next day dawned a lot slower on me. I was a bit more sane post breakfast and headed into the embedded miniconf. There were quite some interesting discussions about what kind of robots to build to host your own embedded linux. Here are a couple of excellent suggestions that got the popular vote.
Since the whole embedded dance around lasted only half a day, I headed back into the debian one to listen to Keith Packard talking about the Fastest Debian machine in the world. This one clocks in at a whopping Mach 3 and sits on a solid propellant rocket. There were a few digressions into patch antennas and accuracy of commercial GPS systems on high accelerations. Also the obligatory jokes about the sock monkey that survived a crash that drove the rocket 80 centimetres into the middle of a sage brush (it always lands on one, in the middle of one). And then there was the tracking antennas. Later on there was Bdale talking about the debian project.
At around five, we assembled at the Uni Link to meet up with the guys organizing the cricket match. We didn't have a big turn out in particular, but we got to a decent mat pitch in the university ground nearby. The match turned out to be Kiwis vs Rest of the World, though we had to supplement the kiwis with a couple of aussies to balance out. In hindsight, that wasn't one of our brightest ideas. Rest of the world lost by 4 runs to the kiwis.
Heading back into Unicol, we ran into some interesting things in the lobby. For starters, it was Christmas and there was a pizza from Hell. I headed back up into my room for a shower and freshening up before we all went out to dinner together.
We went to Terrace in the Octagon and I had trouble finding veg food. It is not that I don't like vegetables, but I like them to be not green - I love potatoes, tomatoes, beet roots and carrots, not lettuce or parsley. I was sitting with my back to bar and I caught this line from Kirby which I promptly misunderstood - Hey, how much for that barmaid ?. Of course the Barmaid turned out to be this glass pipe of beer with a tap at the bottom.
Wednesday was when the conference began in earnest and we all headed out to St David to listen to the first keynote. There was some hilarity about the way Russell had reacted to the network downtime, by hitting the registration page with random kiwi insults. Then the slides said not to mention any names and had a photo of him. He hadn't attended the keynote and was quite surprised to see everyone look at him and just smile while on the way to the halls. But the network story has a lot more twists and turns to I came to know later.
The fibre for the network was lit only on friday evening. That happened because the NZNOG guys knew the admins all around the place and managed to perform some miracles for this cable from Wellington (that's the other island). Then the cable couldn't laid in the university premises and had to be terminated at a building outside the uni. They setup a Trango wireless hookup and somehow managed to get the connection pushed into the patchboards by proxy. It is quite amazing that they managed to get something like this working over a weekend - kudos.
And then I headed into Skwashd's talk about CalDav, other calendaring and groupware formats. His laptop failed to work on the projector and here we see him sweating out trying to copy files onto Ian's laptop using my USB hard disk. After all that I headed out to visit the Otago Museum which was just across the road - but that deserves a complete blog entry. The only mistake I made was trying to enter the hackfest without a laptop. I didn't have internet access and Rusty lent me his laptop to search for stuff on.
Thursday was a little bit more hectic, because I hadn't slept all night - kept awake by the nagging disappointment of the hackfest. I took a quite long walk into the gardens nearby, which I'll go into in another post. I ended up listening to Dave Miller whose dressing style can be a bit disconcerting. As Rusty introduced - we had this service guy come look at the projectors and suddenly start talking about linux on sparc64 (at usenix). He made a few points which I'd like all youngsters moving into FOSS to actually note down and follow religiously.
And then he went on to explain this funny anecdote about Linus and software interrupts. And how Linus rejected a patch that had used softints to handle console input more efficently. But after realizing that there is no better way to do it, the patch was obfuscated by changing the names to basehandler instead of softint. Linus accepted the patch - therefore Names Matter.
Since RMS couldn't make it there in person, we had make other arrangements to get his opinions on the future of Open Source. Back at unicol that night we found the hidden pool balls and finally managed to clear the whole channel inside the table with Tom's long hands. And we still didn't have a 15 in the motely collection.
Friday dawned beautifully and I ended up walking toward the railway station at around 5 AM in the morning. I headed back into Unicol at 7:30 to have breakfast and prepare for Damien Conway's mindblowing keynote about Sex: technical sessions - at least that was what I was interested in :).
Damien was playing to the gallery and using all his presentation skills. The slides were awesomely funny and dealt with the coincidences of Larry Wall's birth and how perl was asymptotically converging on features (with the obvious suggestion that the asymptote is lisp). What I really took away from the whole keynote was these few words "People will make fun of you whatever you do, so don't give up". Some other people have taken away a few slides I didn't dare take pics of - like michaelcarden.
After that I listened to Maddog talk about Open Sores. Quite interesting and insightful comments about what to tell what kind of person. In fact tailor your message depending on the target audience. Like companies prefer to hear the word control a lot more than freedom.
I then went to talk to Pia about Women (or the lack thereof) in FOSS contributors. Gender equality is one of the interesting social problems I see in FOSS. I am trying to understand the problem rather than try to actually solve it. Essentially the conversation moved in the direction that FOSS environments are a lot mot diverse than an average office and you are less likely to be discriminated for being different (as in you just a little girl patronizing included). Pia brought up the topic of children and Keith Packard jumped in with I have kids too. Let us just say that it turned into a discussion of unicycle hocky versus unicycle polo and how there are people who will unicycle down hills. We are amazing at holding down conversations, especially before dinner. Went to the PHP BoF and then proceeded to trickle down to the Dunedin town hall for dinner.
The food sucked badly - veggies just get salads and orange juice. But the presentations really rocked and then there was Bob. We had a free seat at our table which got served two course because we ate what was served there as well. But Bob didn't get any dessert. And then there was Miguel on a stick with which we had a bit of fun. Rasmus won the beer drinking competition and Rusty ran the auction for two hours.
The John Lions book was being auctioned - signed by some very famous names. The money would go towards establishing the John Lions chair. The book has a dark past. Jeff Waugh offered to shave his head if the bidding went up to 7,500 AUD (when the going was 2,900). Aaron Seiga jumped up and offered to fix any bug in KDE if someone puts up the money. Rusty's moustache, Dave Miller and Groggy's beard was also added to the mix for 10,000 AUD. Finally the cartel of alumni from UNSW paid up the amount. Also I met someone.
I tried to find my way back to Unicol and ended up going to the octagon with a bunch of guys. We ended up coming back to the unicol at sometime nearer to 2. And that's where I'll stop writing about.
Saturday was mostly dull, I'll probably classify that as the conference aftermath. Mainly because a lot of the crowd had disappeared by then. More about that when I get around to describing Rusty minus a moustache.--
A university is what a college becomes when the faculty loses interest in students.