Tuesday, January 27, 2009


Long time ago, people used to believe that the Earth has been the center of the world, and everything else has been revolving around our little planet.

To the naked eye, it seemed reasonable and didn't contradict the everyday / everynight experience.

Then the telescope has been invented. It allowed (at least, to those who have taken the trouble to get one) to see not only the Sun, Venus and Mars, but also some more distant planets. Apart from that, people were becoming more and more skilled in math, and started trying to use their math skills to predict the way the planets move.

So began the trouble.

It just so happened that the planets moved in quite tricky ways. Instead of just rolling in giant circles along the sky, it seemed to the Earthlings that the planets were revolving in circles around the circles (it was the only possibility to bring at least some harmony into these so unexpectedly stochastic movements). For these weird paths, the special name has been coined: epicycles. But why the planets behaved this way? Probably there existed some theories on that behalf. For the reason explained just below, only the people who are interested with history of science, and of astronomy in particular, might still know a bit about them.

The reason is now well known: somebody happened to get another idea, namely that the planets have been indeed revolving, but not around the Earth. The Sun appeared to be the real center. And if you would accept this idea then there were no need for these complex epicycles, because the planets were simply moving along nice circular orbits (which is way better and simpler to calculate). Later on, it appeared that these orbits were actually a bit elliptical, but this is another story.

There has been some struggle around this idea, and other subsequent ideas which shifted the center farther and farther away from our little habitat, but finally, the people had accepted that the Earth is not the center, the Sun is not the center and even the center of our Galaxy is not the center of everything. Because it helped to simplify the representation of the world, and the simplicity is the key to beauty, and the beauty is appealing to the humankind way more than complexity. (We can go on and on from here).

The point of this is very simple, but has to be stressed now and then: if your world starts getting way too complicated, it might be time to shift your point of view.

Monday, January 19, 2009

Open-source seem to be a good candidate for the next round of tech-related hype (just after the dotcoms and web2.0). Ideas which have been implemented more or less quietly by a bunch of enthusiasts (like Wikipedia, for example) are already cloned (Google's Knols) - open-source is a philosophy which allows to grab the ideas of others and reimplement them, as long as you don't forget to mention the original authors (if they can still be traceable)...

Several things which I can't help noticing:

1. I can't understand how the old patent system and open-source approach to the copyright are going to coexist together. For example, Google is now actively promoting their new mobile phone, Android, and encourages young, enthusiastic and (often) not-well-paid-yet guys and girls to write new, exiting, and no doubt, open-source applications to promote Android even further (the best ones are getting the blue ribbons indeed). What about the patents for these ideas? Who is going to hold them at the end? These guys (I have asked one of these winners at the Android Developer's camp in Amsterdam) seem to never think about this.

2. What about the data? Who owns the data collected by Wikipedia? Who owns the data collected by e.g. TomTom (my previous employer) via MapShare technology? Who owns the data collected by OCLC (my current employer) via WorldCat? (This one is actually the flamebait right now, the other two have not been discussed, or I don't know about it). Who owns the data collected by Google via their system of blogs (like this one :) ), webmail, etc? Is it possible to talk about the data ownership in this case, or only about the ownership of the data maintenance structures and the right to use the data because of that?

2a. Don't forget, "data maintenance" should include data verification and systematizing, which is not the work for novices and requires quite a strong discipline, like every necessary routine . It also implies guaranteed all-time access to the data (which means hardware issues), and these parts are not that easy to open-source. Distributed computing is one of the answers to that, but the payoff is performance. If the responce time is supposed to be critical, then you need to have some dedicated hardware and it has to be supported, therefore somebody is supposed to get paid for it.

3. Open-source, in itself, is not a silver bullet. If you compare an open-source solution and a closed-source one from the point of actual costs you spend to get what you want, the result might be quite surprising: fixed costs for closed-source solution (price of the software + maintenance costs, which is usually agreed upon beforehand) versus open costs (nothing + salary of the guy or guys who are going to support this piece of software for you).

4. Software is not good or bad because it is open- or closed-source. It is good when:
- it is clearly documented;
- it is well tested;
- it is well supported;
- it is written well.

All of this can go wrong for open-source project as well as for closed-source one, but in the case of open-source project, if you happen to be the client and the project dies, you are left with nothing because nobody has been responsible. The documentation is not the thing programmers like to do; you won't find much documentation even for Mozilla (one of the longest and extremely popular open-source projects). The wider scope the open-source project has, the more important becomes the question of organizing the "crowdforce"; it is more difficult than to order around guys who are reporting to you because, after all, these enthusiastic guys and gals work for fun!.. If you spoil the fun, they'll go!

5. The term "crowdsourcing", to be honest, makes me remember the work of José Ortega y Gasset called, in English, "The Revolt of the Masses". The point of this work is that unorganized masses are actually the reactionist force. The crowd, if not organized, is less smart than any single member of it. What is needed for open-source project to succeed is the organized crowd, so that it really becomes the whole where various parts play various roles for the common good. If the project leaders manage to achieve and keep this state, good for them. If they also know what they want to achieve with their project and can share their vision with their team and get their support in return, excellent. That is how it should be. But there is no guarantee that it will always be like that.

6. "The human factor" becomes extremely important. People have emotions. You never know what makes the other one explode until you are swooshed away by the blast. If the project becomes big enough (on the order of several dozens of active participants) the conflicts will be inevitable. The project leader has to be prepared to solve them, sometimes by force, and to have enough self-confidence, so that the others would not question every decision he makes. This is a tough part.

7. Finally, the disclaimer: I am quite enthusiastic about the open-source approach, but that is why I can't stop thinking about the ways how it can fail. It would be a real pity if open source, as a result of the coming hype, becomes an "anti-buzzword", like these dotcoms of the past century.

Update: Open-source has its charms, there is no doubt about it. It allows the developers to become experts in the areas they like, and to stay experts no matter what their official affiliation is. It makes the workforce more mobile, because your expertise with (the development of) open-source products is often easier to take along to the next assignment than the expertise with (the development of) closed-source products. Therefore, it makes the work for the developers look more like hobby, which is good for the morale. That is why it is so important to make sure that open-source initiative, now that it came into open light, will stay healthy.

Thursday, January 15, 2009

The gist of Goedel's theorem seems to be found in the fact that our language becomes a pitfall when it is used for self-reflection. Any formal system which we could have thought of (till now) has the same flaw.

No human-made formal system appears to be capable of self-reflection; yet any human can do it.

Why are our brains "made" in such peculiar way? As if there is a blind spot which prevents us from seeing the source of light. Will we be able to develop a workaround this spot? If the answer is yes, what sort of "language" will be used for such purpose (i.e. the formal system for which Goedel proof does not hold)? Will it be a "language" at all? What is, after all, the definition of the concept of "language"?

Tuesday, January 06, 2009

Continuing the translation-related stream of thoughts...What is, ultimately, the purpose of translation? According to Hofstadter, a thought could be seen as the specific itinerary via the (sub)network of active symbols (with all traffic scrupulously recorded). Along with this, we have to realize:
- there is no other network like this one;
- the author of the thought (be in in the form of a novel, a piece of poetry or even a phrase) does not necessarily perceive in all detail what was the purpose of all this and what the implications might be;
- ultimately, if to take this utterly seriously one will find himself / herself in the situation of S.B.Odin from Strugatsky's book 'Monday begins on Saturday', who, as we remember, knew how to perform every possible miracle, but was not able to really do anything, because it was never possible to satisfy all conditions!

Therefore, we can never expect the translation to be ultimately perfect. It simply cannot be, because no two human beings possess the same symbol network. The interesting question therefore is, what is the threshold of the information loss. How much information can be preserved when transmitted from one human being to another? How much information should really be preserved? (What if the noise is part of any thought, like virtual particles cloud is part of any "real" particle - sure we don't have to translate the noise?..)

Monday, January 05, 2009

Reading Hofstadter's GEB, I've come to the point, where he cites "Jabberwocky" in English, German and French and compares the beginnings of the three translations of "Crime and Punishment" into English. This, and subsequent episodes, have provoked some response in my symbol network, which I am going to try to pin down.

First, regarding Hofstadter's remark of the French translation of the word "slithy" ("lubricilleux"), he wonders if using the Latin-based word where its analogue in English is non Latin-based would trigger the additional sense of "alienness" of this word in the French reader which was not the intention in the original. I am not the ultimate expert of this subject but I have the impression that the French perceive the words with Latin roots simply as their own (French, after all, is a Roman language while English is not). Therefore, the translator's choice was probably quite fitting one. It is the fact that French has so many words in common with English (first, thanks to the fact that the English originally also used to be a Roman colony, and second, because the English have been conquered by the French later on) which obfuscates the difference between these two languages, but I wonder why Hofstadter did not mention it (could it be that he did not meet many French people at that time? ;) )

Second, regarding the Dostoevsky's translation. In the first phrase, there is a name of the street (S.Pereulok, abbreviated from Stolyarny pereulok). One of the translators just left S.Pereulok as it was, the other one changed it to S.Lane, and the third one called the little street "Carpenter's Street". Interesting that Hofstadter is against the translation No.3, because to him, it makes all thing sound like one of Dickens' works which could take place in London and in this case, why read Dostoevsky, asks Hofstadter, instead of reading Dikkens who is the ultimate translator of the same ideas into English?..

This is an interesting point. Could it be that different persons would require different flavors of translations for their minds to accept them as most suitable? The one example which immediately surfaced to my mind is the Russian translation of "The Wizard of Oz", which, as almost any Russian speaker knows, is called "The Wizard of Emerald City", has six follow-ups which wander far, far away from the original story (which already has been altered "to get accepted by the Russian children better"), and which is, as far as I know, far, far more popular than the more accurate translations of the original.

From the other side, the most popular Russian translation of Alice in Wonderland is not the one of Nabokov, who tried to do similar trick, substituting English realities with the Russian ones, but either the one of Boris Zakhoder, who preserves the "Englishness" of the book but does makes all the word plays and puzzling paradoxes comprehensive, so to say, or the one of Nina Demurova (who is assumed to be, in any case, the best translator for the little "Jabberwocky" piece)...

Thinking about the connotations and associations which exist in one language and do not exist in another looks per se both puzzing and endlessly interesting, but there are lots of traps where it is so easy to fall. For example, a simple English phrase "one another", at the first glance, seems to be literally translated into Russian as something like"friend-a-friend" (друг друга). But then, the word "друг" here can be an abbreviation of the word "другой", which literally means "other, another". But then again, isn't the word "другой" related to the word "дорогой" (dear), by the rule of making longer Russian words from the shorter Ancient Russian / Church Slavic words? Anyway, the words "friend", "another" and "dear" seem to be related in Russian, if only by the power of alliteration and pseudo-ethimology. What a nice thought can be drawn from this fact - that in the language itself, it is encoded that all the other persons are, by default, our friends, and they are, by default, very dear to us! (Incidentally, the word for "enemy" (враг, ворог) seems to be close with the word for "doing (evil) magic" (ворожить) - which can, if one is in the proper mood, be explained as that our only enemies are those who are doing the magic. It actually fits, if one takes the definition of magic as the attempts to overturn by the force of individual will the laws of nature - such persons could be indeed potentially dangerous!)

Enough for today.

Sunday, January 04, 2009

В кои-то веки я наконец узнала, что лопух - это то же самое, что репейник. Для меня всегда это были два разных растения!

Википедия - это сила :)

Friday, January 02, 2009

Текст, переведенный с другого языка (если только переводчик - не Переводчик с большой буквы), отличается от текста, созданного на этом же языке, как посмертная маска - от работы, сделанной скульптором (даже если он - разгильдяй-недоучка) с живой натуры.

Поэтому я многих современных англоязычных фильмов и книг в русском переводе не воспринимаю. Хорошие переводчики, конечно, никуда не делись. Просто плохих стало чересчур много.

Кстати, случайно попался на глаза неплохой (англоязычный) пост с рассказом о переводе и переводчике. Там даже упоминается соответствующая книжка Хофстадтера, которую, кстати, сам автор считает гораздо более важной, чем ГЭБ.

Когда мы наконец додумаемся до технологии более-менее адекватного перевода, мир здорово изменится. Возможно, это даже произойдет в ближайшие лет 10 - если, конечно, человечество не бросит свои силы на очередную глупость вроде мировой войны. С другой стороны, предыдущей мировой войне мы обязаны возникновением компьютеров... Но лучше бы все-таки без катаклизмов. А то и разговаривать будет не с кем.