As I said I would, I installed OpenSol­aris after I fin­ished my FOPARA paper. It cer­tainly did not go smoothly. (more…)

I sub­mit­ted the paper to FOPARA this after­noon. It’s not a fant­astic paper, but it’s not too bad. I prob­ably say that about every paper. The scheme we came up with for bounds infer­ence was slightly hair­ier than I’d hoped which means there were a lot of loose ends to track down while stay­ing within the page limit.

In any case I’m pretty excited about that being done. I can relax a bit and do more imple­ment­a­tion work before LCC.

On a note of annoy­ance, I put a link in the paper to the Pola pro­ject home only to find that Red­mine stopped work­ing about two days ago. Dream­host may have done an upgrade of Rails or some­thing, I’m not sure, but it’s a prob­lem I didn’t fore­see. I’ll have to try to fix it tonight or tomorrow.

My sister-​​in-​​law just made an awe­somely cute new entry to her pho­to­graphy blog which led to:

Jasna: can you fin­ish your PhD and get a real job ASAP so I can start pop­ping out babies?

With the recent announce­ment from Google regard­ing the release of Google Wave—slated for late Septem­ber 2009 — I finally decided to do some read­ing on it. I recom­mend watch­ing the abridged video of it to get an idea, though the abridged ver­sion leaves out some cool stuff about the protocol.

I love the pro­tocol. Both Google’s actions and words point to Wave being some­thing which is not con­trolled by Google in any way. The pro­tocol will be open and Google’s ref­er­ence imple­ment­a­tion will be open source. To start with, Google will be the world’s largest Wave pro­vider since they’re the ones who have developed the pro­tocol, ref­er­ence imple­ment­a­tion and have done the test­ing on it, but ideally organ­iz­a­tions will start offer­ing Wave accounts in exactly the same way they offer email accounts today.

If you watch the video, you get a very ambi­tious view of Wave: it will not only obsol­ete email, but also instant mes­saging, polling, blog­ging, pic­ture shar­ing and file shar­ing, wikis and even video games. Being a little bit famil­iar with XMPP — the tech­no­logy Wave is built upon — I have to say it was only a mat­ter of time before someone tried to do some­thing this ambi­tious. XMPP is a beau­ti­ful pro­tocol and open and extens­ible to boot. Wave seems like some­thing that had to hap­pen eventually.

Even if Google Wave sup­plants noth­ing other than email, it would be well worth it. Email has a lot of short­com­ings, as any­one who took my net­work­ing course should appre­ci­ate. Cryp­to­graph­ic­ally secure authen­tic­a­tion and encryp­tion end-​​to-​​end is required by the pro­tocol — it is impossible to send any­thing unen­cryp­ted with Wave — and that alone is some­thing to be pleased about.

The down­side to this all is I’m get­ting all excited now about it and I want to start work­ing on devel­op­ing a Wave cli­ent. I’m all adult and bor­ing these days with my “respons­ib­il­it­ies” and so it has to get thrown to the bot­tom of my to-​​do list along with all the other cool non – research-​​related pro­jects I’d love to do.

And I should say I have a hor­rible his­tory of pre­dict­ing which tech­no­lo­gies will catch on and which won’t. In my world, MSN and Face­book should be fail­ures; PGP and XMPP should have taken the world by storm. With that track record, I sup­pose Google Wave is doomed to be a spec­tac­u­lar fail­ure, which would be a real shame.

From a chat with my friend Marla, edited for brevity:

Marla: did you work on the paper?
Mike: haha
Mike: a little bit
Mike: I came across a rather excit­ing prob­lem with it
Mike: I hope I can still get it done in time
Marla: :-O
Marla: excit­ing and prob­lem don’t often go together
Mike: of course they do!
Mike: that’s what research is all about, silly :P
Marla: >_​<
Marla: this is why I’m not a grad student

Yes, I’m hard at work on the paper for FOPARA, still, as the deadline’s been exten­ded. And yes, I have come across a rather excit­ing prob­lem, deal­ing with infer­ring size bounds from safe recur­sions (folds). I think if I can get a solu­tion to the prob­lem, even a rel­at­ively naïve one, for the paper I’ll be happy, though I think the prob­lem as a whole has a lot of poten­tial for clev­erness down the road.

Is there any good reason to refer to large amounts of data in terms of powers of 2? E.g., where 1kB is 1024 bytes (now prop­erly called a kib­i­b­yte) instead of 1000 bytes? People rag on hard drive man­u­fac­tur­ers for deceit­fully list­ing a 320 mil­lion byte capa­city drive as 320GB instead of 300GiB — and I’ll admit there’s undoubtedly some wil­ful deceit going on there — but to be hon­est, using powers of 10 makes a lot more sense to me.

The only instance I can see where using powers of 2 could pos­sibly be bene­fi­cial is when refer­ring exactly to the num­ber of MMU pages or the num­ber of file sys­tem blocks alloc­ated for some­thing, but in those instances I think I would rather say some­thing like “7 pages” than “28k”.

And yes, this is inspired by repar­ti­tion­ing my hard drive and being annoyed at hav­ing to con­vert a num­ber like 202691383296 bytes into gib­i­b­ytes (base 2 giga­bytes) because the tool stu­pidly won’t work with base 10 giga­bytes. Some­times I think people go out of their way to make things more arcane than they need to be.

For those who know a bit about me, you know one of my favour­ite musi­cians is Gisle Martens Meyer, his most fam­ous musical pro­ject being Ugress. I fol­low the blog, I buy the music, I down­load the music that isn’t sold, I even briefly con­sidered mak­ing a trip to Min­nesota at one point to see his first (and so far only) live show in North Amer­ica. If I’m a “fan” of anything/​anyone in this world, it’s surely Ugress. Still it took me over 2 weeks to build up the need to part with my hard-​​earned £6 (it’s always a fun sur­prise to find out at the end of the month how much my credit card com­pany has charged me for for­eign cur­rency) and buy his latest album. In fair­ness, I am very cheap.

(If you’re inter­ested in his music, most of his albums offer free tracks and the rest are only 79 pence each)

First, a brief review. Remin­is­cience is much bet­ter than Uni­corn (the pre­vi­ously most recent album) in my opin­ion. Uni­corn had some good songs but the dir­ec­tion of the album was a little poppy for me, focus­ing a lot around the vocals of Christine Litle (who I must con­fess is pretty darned cute and a good singer to boot). This album’s more about the beats and the sound aes­thet­ics, which is what got me inter­ested in Ugress in the first place. I’ve only listened to half the album so far and I already I have a good 3 or 4 new favour­ite songs.

Read­ing his blog, you get the impres­sion that GMM is more a sound engin­eer than a com­poser. That’s prob­ably an exag­ger­a­tion (and it’s easier in text to describe engin­eer­ing pro­cesses than a par­tic­u­lar com­pos­i­tion on the go), but it’s pretty clear he takes sound pretty ser­i­ously. I’ve paid spe­cial atten­tion to it lately and noticed he really does a remark­able job of get­ting the most of the dynamic range that the (16-​​bit sample depth/​90dB, if I’m not mis­taken) MP3s will give you. Cool inter­est­ing post: describ­ing how he used a speech syn­thes­izer to gen­er­ate lyr­ics. Okay now I really am sound­ing like a fan.…

Any­way what I really wanted to talk about was GMM’s reac­tion to the album being pir­ated (or copy­right infringed upon to use made-​​up pos­sibly cor­rect ter­min­o­logy). He doesn’t just do the pro­duc­tion him­self, but he does all of the pub­lish­ing and dis­tri­bu­tion and what­not him­self. Tra­di­tional music labels, I under­stand, are as good as dead and worse than use­less. He gives out some free tracks, the rest are cheap and dis­trib­uted largely online, he’s okay with mash-​​ups and pub­lic per­form­ances and deriv­at­ive works and using his music as a back­ground to your stu­dent film. He dis­trib­utes the music in high qual­ity (320kbit/​s) non-​​DRM MP3s and even in FLAC.

What’s inter­est­ing about this is that he’s one of the very few (the only one I’m really closely famil­iar with) that is doing everything Right and makes it his day job (and then some; I worry a little if he has a life besides). Not that it I should be sur­prised that it would get pir­ated. People will pir­ate any­thing. People pir­ate free soft­ware. It’s almost like people put more effort into pir­at­ing than it would take to get some­thing legit­im­ately just for the sake of pir­at­ing. Well I’m not being fair: clearly people are pir­at­ing because a £6 album costs £6 more than a £0 album.

The primary reason I sup­port free soft­ware is to give free­dom to the users to use and repur­pose the soft­ware how­ever they like. This is from a dog­matic stand­point (I treat soft­ware pub­lish­ers con­trolling their soft­ware with great sus­pi­cion), from a qual­ity stand­point (there’s an annoy­ing bug in OS X’s Terminal.app for the past 6 months or so that I could likely fix in lit­er­ally a minute, but Apple has pre­ven­ted me from doing so) and from a human­it­arian stand­point (it’s near impossible to advance human­ity with improv­ing on one another’s work without per­mis­sion). I can see how this car­ries over from soft­ware into other realms of “inform­a­tion”. One thing I’m def­in­itely not is an eco­nom­ist, but I’ve increas­ingly viewed the Free Soft­ware Foundation’s free­dom 2 (that you should be able to share what you like with your neigh­bour) with some hes­it­a­tion. It’s a fant­astic idea in a great many of cases (cf., Linux), but I don’t see this as a strict mat­ter of free­dom, but as a mat­ter of eco­nom­ics and it’s one that I think can be treated on a case-​​by-​​case basis.

With this in mind I see GMM’s way of doing things as the Right way of doing things, or more accur­ately, as a Right way of doing things and would like more people to send him money in exchange for his hard work. Then again, I am just a fan. Ser­i­ously, The Bosporus Incid­ent is awesome.