Thoughts from Inside the Box

Google Is Still Smarter than You

2009-04-23T18:57:00.002+03:00

Ever tried to use Google when in a foreign country where you don't speak the language? Google helpfully figures out which language you want to see based on the country where you are. So I've gotten Google in Spanish, Japanese, and now in Slovak, just by traveling to different countries. And my foreign coworkers in Finland complained about this behavior all the time.

If only there was some standard way for a browser to indicate which languages its user can understand, removing the need for such an ad-hoc and error-prone mechanism...

The World Is Collapsing

2008-07-21T17:59:00.002+03:00

Only 17 days to go before scientists in Switzerland throw the switch and create a black hole that will destroy the Earth.

What Is Money?

2008-01-19T20:34:00.000+02:00

Watched an interesting video today: Money as Debt. It's not really healthy to know too much about how the economy works...

It also brought to mind an interesting argument. At one point in the video, it is noted that an individual paying out all debts means more money for that individual, but everybody paying out all debts means that there no longer is any money. So if you happen to subscribe to a philosophy that requires only actions that would be acceptable if everyone did them, you need to ensure you remain constantly at debt. And furthermore, since the whole thing is a massive pyramid scheme, you also need to increase your debt exponentially.

Tieteellinen pätevyys

2008-01-13T15:33:00.000+02:00

Tohtoriksi valmistumiseen ei riitä pelkkä väitöskirja, vaan tarvitaan myös julkinen väitöstilaisuus. Suomessa väitöspäivä ei ole mikään pieni asia, vaan suuri tapaus, joka vaatii valmisteluja:

Ota yhteyttä vastaväittäjään ja kustokseen ja sovi väitöspäivästä
Varaa sali väitöstä varten
Tilaa kahvit väitöksen jälkeen tarjottavaksi (varaa myös tila kahvien tarjoiluun)
Hanki kirjalle ISBN ja mahdollinen julkaisusarjan numero
Toimita kirjan lopullinen versio painoon
Ota yhteyttä tiedottajaan ja kerro tulevasta väitöksestä
Syötä tarpeelliset tiedot yliopiston väitöstietoihin
Kirjan tultua painosta lähetä kappaleita annetun listan mukaan
Etsi sopiva ravintola karonkkaa varten
Vuokraa oikeat asusteet
Varaa kaupungin ulkopuolelta tulevalle vastaväittäjälle hotellihuone

Ja mielenkiintoisinta tietysti on, että tässä n. tusina henkilöä käsittävässä prosessissa vain yksi ei ole käynyt koko hommaa läpi useita kertoja, ja tietysti kaiken tekeminen täsmälleen oikein on juuri hänen vastuullaan...

Google Is Smarter than You

2008-01-08T13:02:00.000+02:00

Just now, I happened to be searching for "tree dog programmer" on Google (don't ask why). Of the first 10 hits it gives, only 2 actually contain the word "tree"; the other 8 match the word "three" instead. Is it too much to ask to assume people can actually type common words correctly, or can correct their mistakes after noticing the problem? Apparently so (note that this "intelligent" spelling correction can be disabled by surrounding the individual word in quotes).

Another interesting example (I don't have one handy now) is when I look for some rare term, usually getting only a few hits. So then Google suggests "Did you mean: <something else>", and it turns out that this suggested "correction" doesn't actually match anything.

Funniest Thing Ever

2008-01-03T23:01:00.000+02:00

I really don't remember the last time I laughed so hard, as I did reading DM of the Rings. If you're familiar with Lord of the Rings (preferably both book and movies) and fantasy role playing (especially if you've been DM/GM), don't miss this.

On a similar topic, Order of the Stick is also great fun.

Personality Testing

2007-08-30T21:35:00.000+03:00

So I'm Picard. I'm pretty sure the last time I saw this test I was Data, but no amount of reasonable fiddling this time got me away from Picard. And with this post, the Picard result is immortalized onto the great Internet, never to change again.

Finland, Internet Backwater

2007-08-20T22:44:00.000+03:00

When I'm shopping for goods or services, my preference is physical store (but keep your salesweasels off my back up to when I've decided, please), Web and email, phone (I really dislike phoning people I don't know, it's all the impersonality of email without the ability to be precise). So, since most businesses I want to do business with publish an email address, I usually email them if it's not convenient to drag my backside there.

Some six years back I needed my TV repaired. I emailed one company, no response for a few days. So I pick another one. They call me back, leaving a message on my answering machine that indicates they print out all email that comes in and then someone goes over the printouts before the day ends. Funny method, but they did get my money since they replied at least. Well, the first company replied too, after a couple of weeks.

But surely, all this was years ago, and things are better now. Not really. Recently I ordered some bookshelves over the Web, and of course their preference was to call me even though I provided an email address as well (the form had a mandatory phone number and an optional email address, it figures). Not only that, the call came from an unlisted number, so I rejected it, as is my custom if I'm even the slightest bit busy. So they had to relent and send me email asking how I'd like to pay. As is my custom with business email, I replied immediately, expecting that they'd handle things the next day. Hah! It's now been two weeks and not a beep from them (I wasn't really able to contact them last week, but will tomorrow).

I don't really have high hopes of anything requiring actual service to ever work in Finland over the Internet. My experience is that businesses provide email addresses, since nowadays it's expected, but they don't really bother to read them or to provide the same level of service through email that you'd get if you just show up at the store (of course, then you'll get pre-empted by someone who calls in, but usually that's not such a big deal).

First Crossword Completed

2007-08-15T00:07:00.000+03:00

One of my hobbies is solving crossword puzzles, which I've been doing for nearly 15 years now. But I never tried making them myself before I met another person who had just started doing that and pestered me to try as well. So now I finally got my first crossword finished, and tomorrow I'll send it to my test solver.

I started by first putting in all the words into a free-form grid so I could move the black squares when necessary. This phase was very much like solving a crossword. It's easier because the grid is not static so bad words can be replaced by longer or shorter ones as well, but it's harder because there is no certainty that even with good words it's possible to complete.

Figuring out the clues I had been doing a bit during the grid making as well. I started with a list of 50 clues or so, and expanded on it by about 20 when new good words appeared in the grid. Eventually I managed to use about 40 clues and had to invent 70 new ones after the grid was complete. I don't think I did a really good job with this, but I'll just wait for the solver's opinion.

The final part is then writing the clues. In a Finnish crossword the clues are usually embedded inside the grid, so getting them there legibly takes some serious penmanship. I mostly managed, but I can also understand why the usual method is to draw the crossword on a larger sheet and then shrink it for printing (in fact, I had also prepared for this alternative).

Upgrading Internet Connectivity

2007-08-13T22:44:00.000+03:00

I finally did what I had been thinking for at least two years: I got myself a cable modem and gave up on pay-per-minute dial-up. Yes, it seems kind of weird that a computer geek uses dial-up, but it has worked pretty well so far (and would still have, I just got off my lazy ass and did something now).

I figure the main restriction from using dial-up was that I haven't been able to participate in Web forums. But even that's not so bad, Web forums suck anyway, Usenet is so much better technology-wise. I'll have to watch myself now to avoid picking up new time sinks with the new opportunities.

The major pain came with reconfiguring my mail and news program. I use Gnus, which I have configured to work perfectly for me, and this of course means that changing servers is a chore I'd prefer to do without. Surprisingly, this time things went just fine from the Gnus end (I guess it helped that the mail server also changed its name in addition to everything else). For news I decided to look further ahead and went with NIN, also in part because all of their customers seem very happy with it, and a well-administered news server is important to me.

Mailing lists are still insane, though. I'm subscribed to a dozen Debian lists, and even though they have a handy Web form for subscribing and unsubscribing, each list still generates one confirmation email for unsubscription and another for subscription, which all have to be replied to.

This did bring me to a decision: If I ever write a mail program, I'm going to include support for mailing lists and to make it easy to switch addresses and servers. This is an interesting point: a complex operation of a program that is rarely used but needs to be perfect when used, must have a single button to do it, but it's really hard to test such functionality.

And of course there's still the main reason for switching to a faster always-on connectivity: now I can upgrade my Debian system every day and don't have to use my weekly sneakernet connection.

Out of Pre-ground Coffee

2007-08-11T18:20:00.000+03:00

A few months back I gave up using a drip filter coffee maker, for a variety of reasons. Unfortunately, in Finland it's really hard to find ground coffee coarser than filter coffee, especially since I prefer to drink Fair trade coffee. In fact, I'm limited to pretty much one brand, which I don't really like that much.

So, of course I had to get my own grinder, since the selection in beans is slightly better (though nowhere near the state of filter coffee). And today I finally ran out of the ground coffee I had previously bought (incidentally, both at work and at home) and will have to switch completely to self-ground coffee. And it does taste better than pre-ground, even when it's the same kind.

Breaking My "I Watch TV" Misconception

2007-08-09T12:23:00.000+03:00

I own a TV. I watch it occasionally. So, since Finland is going all-digital soon, I made plans to get a digital set-top box, so I could continue watching.

But yesterday, when reading the paper, I finally realized that I really don't watch TV at all. Practically all I use the TV for is watching DVDs, either movies or series. This year I've been following maybe two or three series, but really, I could have just got the DVDs for those as well. I don't watch the news either, I prefer newspapers for that.

So, with that realization, I concluded I don't need to "upgrade" to digital at all. And thus, when the analog transmissions finally end, I can also look forward to periodic visits from the TV license inspectors.

Thesis Submitted

2007-08-07T14:26:00.000+03:00

I've submitted my PhD thesis for review, and it's now halfway through the allowed three-month review period. I seem to be chronically unable to work without deadlines: during the last week before submitting, I added over 60 pages, about the same amount as in the preceding two months.

There's nothing actually groundbreaking there, but I did run many experiments and wrote many pages of text, and that's what counts, right?

What's the Point of Blogging, Anyway?

2007-08-05T06:34:00.000+03:00

So I have this blog that hasn't really seen much activity. Originally it seemed to me that every now and then I'd get some idea in my head that would be useful to communicate and that a blog would work for that. Well, that didn't happen; either I didn't actually get any ideas, or didn't think they were worth writing down, or just plain forgot I have a blog.

Anyway, this post marks my first resolution to actually start writing something here a bit more often, and let's see what happens. Comments on things are welcome, but I doubt anyone even knows of this blog's existence, so I'm expecting this to be mostly a therapeutic exercise for me in releasing my world pain (as a secondary school teacher of mine once called it).

EXI Documents Published

2007-08-04T17:06:00.000+03:00

Well, the EXI group finally managed to publish the first draft of the format specification and the second draft of the measurements note. The work is taking more time than we originally planned, at the moment we are about half a year behind schedule, and it doesn't look like there's much chance to catch up significantly.

But I really like the format itself. It's a clean approach and it's very clear that it works well. It could maybe use a bit more formal way of describing things, but there's stuff there that isn't very amenable to being treated in current formal language theory (as far as I know), so that might not happen in time.

The Evil Write-Compile-Test Cycle

2006-03-26T23:45:00.000+03:00

Computers are now fast and interactive. When writing code, you can just do a compile for every function you write to make sure you got the syntax right. You can write simple test programs to find out how some construct behaves.

Except that this kind of development is extremely bad (yes, I do it too). When you just have the computer tell you if you got the syntax correct, you will not actually learn the language properly. When you use test programs to figure out library use, you don't get to understand the library.

The main problem with this comes when designers of languages and libraries start to expect this kind of development. They become less likely to document their creations properly, relying on users to apply this trial-and-error method. Potentially there will be a few simple examples that do not tell anything useful, just how to use the most basic functionality.

For libraries, having the source available is the worst thing that you could have (I mean, when you're a user of the library). Designers too easily get the mindset that if their user has the source available, there is little need to provide documentation, since the source says how the library behaves. Of course, the source typically never says how the library is intended to behave, so a user can never tell whether some property of the library is accidental or designed.

My belief is that I should be able to write at least a few thousand lines of code only relying on published documentation on the language and any libraries I use, and have this code work correctly on the first compilation, assuming I have taken reasonable care in writing. My current experience is that this seems to be an explicit non-goal for most language and library designers.

Ingredients of a Successful Research Project

2006-03-12T22:38:00.000+02:00

I have been working for four years in a research project in the mobile computing area. From observation during the project and discussions with people I have come to understand some of the things that make for a useful research project. Note that, in full, this probably applies only to the kind of applied computer science that we
do.

First of all, it is important to actually implement something. Developing algorithms and simulating them only takes you to a certain point. Having your code run on actual devices and using real networks is pretty essential to understanding all the various issues.

Of course, it's much easier to just run simulations. There is no need to spend months in essentially boilerplate code, no need to care about limitations and bugs in actual platforms. On the other hand, having something that really works is the only way to demonstrate that your ideas are really implementable and not just abstract nonsense.

Probably the most important thing is to have a common vision. The project needs to know the big picture: what is actually being built. This allows everyone to consider how each individual piece fits into
the whole. In contrast, without the big picture, there is the danger that everyone just does their own thing, making the project more of a collection of people all doing their own thing.

As with all activities involving groups of people, it is essential that the project group can communicate among themselves. When doing research, discussions during the idea stages are necessary to refine the ideas properly.

In general, an informal atmosphere in a research project works best. It is no use to try to communicate during formal meetings or the like. A much better alternative is when people are able to just pop in on each other and start talking. Of course, this means that the whole group should be located in the same place.

You cannot really run a research project in the same manner as you would a software project. Research project workers are pretty much equals, so it is not possible to just impose rules from above. Instead, any structure needs to come from just being sensible, figured out and refined among the group.

Finally, the main enabler for success is the infrastructure. The project needs to be able to share their work and to bring the various pieces together. The hardware and most of the tools to do this should come from the organization's IT department. However, a research project needs to have much more freedom in using their systems than normal workers, so the IT infrastructure needs to recognize this.

To take the most advantage of the infrastructure, it is necessary to have someone (or more than one person) in the group who understands all the collaboration systems that are being used. This person can then design collaboration patterns appropriate to the tools that are being used. But like in all things, this person cannot just go and implement according to his or her whims, but must take into account what the rest of the group feels comfortable with.

I will refrain from evaluating our project according to these
criteria. Suffice it to say that I still enjoy working there.

Avoiding Automatic Recognition

2005-12-10T19:30:00.000+02:00

A while ago I was attending a presentation session related to context-aware computing. One of the presentations was about identifying people by their walk. The floor has sensor that measure the pressure that your steps make, and then they figure out from the pressure curve who you are.

This got me thinking about all the research that is being done on automatic recognition of people. Things like this, face recognition, typing patterns, whatnot. They may not be perfect, but there's a lot of money being spent on them, and they may get good enough.

So why couldn't someone start research on thwarting these systems? For instance, for the walking pattern you could possibly switch shoes frequently and try to vary your speed and step length. Other ideas might work for other technologies. And of course you need some computer assistance to keep from having patterns in your avoidance techniques. I think it might make for interesting and useful research, though I don't know who would fund it.

But one general point to make is that in the past you had privacy because there was no way to violate it on a large scale. In the future, there will be a way, so having privacy will require some active attempt to preserve it.

Text Editors and Scripting for Presenting Measurements

2005-11-02T13:41:00.000+02:00

I write academic computer science papers for a living (well, partly, at least). Usually that means I need to run measurements of some program and then massage the raw numbers I get into an understandable form. By now, I've pretty much automated everything, and the workflow might be interesting to read.

First of all, all my papers are written in LaTeX using GNU Emacs. BibTeX is used for bibliographies and I have a single file (in my own FubML format) containing all the references I (or the project) have collected.

The starting point for formatting the measurements is the program that I'm measuring. I set the thing up so it prints some form of header information for every set of measurements and each individual measurement gets printed something like "Input: 2532 (100)", with a name, measured value, and replication count. This is typically even human readable.

Time for processing. I have some generic Perl to read the kind of data I output and then to generate a large array of MetaPost definitions, which includes all the measurements.

Again, after some generic MetaPost code I can get all sorts of aggregate numbers out of the data, like averages and deviations. The graph package of MetaPost is then used for making neat graphs out of this data. Again, most of the code is generic and reusable from one paper to the next. The only-paper specific part is typically the setting of a few variables that control which measurements to include.

Sometimes the data are best presented in a tabular manner. If I need this, the Perl script that reads the raw data also prints out the tables. These get printed in LaTeX tabular format and included as such in the LaTeX document. In fact, pretty much any numbers that I present from the measurements have been automatically generated from the raw data.

Of course, all of this gets tied with Make rules so that a single command first processes the data and then compiles the paper into PDF. I have a generic Makefile for LaTeX processing so I only need to set the dependencies right; these ensure that only the parts that need processing due to changes get processed.

This system has several good sides (well, to me anyway). First of all, everything I use in it has a text-based command language, so I can prepare everything with Emacs (well, nearly so; I haven't yet automated getting information from network packet dumps). Second, whenever I take new measurements (say, if I forgot something or just need more replications) the system rebuilds all the figures without any additional work on my part. Third, since all numbers are generated from the same data, they are consistent. And finally, it's pretty easy to write the Perl scripts so that they don't care about how much data I have, so changing what I'm presenting is typically easy.

There's still one problem: when I take new measurements and update all the figures and tables, the scripts do not yet update any conclusions I've made based on the data...