Norbert is deeply involved in the TeX Live and Debian TeX development.
Dave Walden, interviewer: Please tell me a bit about yourself independent of TeX.
Norbert Preining, interviewee: I was born in Vienna, Austria, where I also grew up, went to school and to university. Although my main study was Mathematics at the Vienna University of Technology, I took a lot of courses in other areas, from Classical Philology (how it is called here: Latin and Ancient Greek) to Tibetology and Buddhism.
After some years working at the Vienna University of Technology, I moved to Siena, Italy, for 2.5 years, after which I returned again to Vienna for two years. From there I moved to my current location in Kanazawa, Japan, working at the Japan Advanced Institute for Science and Technology.
I consider myself somehow torn into three parts: one is my profession, logician, mathematical logic, in particular many-valued logics, proof theory, geometry. This is what I am doing as my main job.
Another part is TeX related, I think we will get more questions on this topic later on.
The last part is mountaineering. I started not really early, as far as I remember around the age of 12 with some simple hiking and mountaineering with my uncles, and one course at a mountaineering school for children. During high school time I did a bit of rock climbing, and during that time, i.e., high school time, my weekends were torn between playing in handball competitions and chess tournaments, and going to the mountains. I really got serious with mountaineering only at university, where I slipped into the Alpine Club in Vienna. From that time on it became more and more an important and integral part of my life. The logical consequence of that being that some years ago I finally managed to obtain the highest qualification, the International Mountain Guide. After that I worked for some years part time as mountain guide. This is a very good addition to the purely brain-related research at the university. Mountain guiding is much more multi-dimensional: it is physical labour, you need your brain (or you and your clients will soon be dead), and you need a lot of social skills to manage clients even in complicated and dangerous situations. I love that job as well as my job as logician; for me the perfect combination.
That led so far that about 2 years ago I had already settled on quitting university work — or better, not to apply another 10 times for different grants — and to work full time as mountain guide. At that very point I got an offer to go to Japan. Probably the hardest decision for me at that time, throwing everything overboard and starting a complete new life far from Europe.
So for the time being (two more years and hopefully more), I am here in Japan; and, I have to say, it was a very good decision. Although I cannot work as mountain guide here (theoretically possible, but not practically), I can do loads of mountaineering myself here, the working conditions are very good and inspiring, and I live for the first time in my life close to the sea.
DW: Let's talk about each of the three components of your life in turn. Since this is a TUG interview, let's start with TeX. How did you first come to be involved with TeX?
NP: That was for the second project during the math studies. I remember very well that for the first we used Microsoft Word, and my colleagues and I were so disappointed about the math typesetting. Coincidentally at that time, we had a very short introduction to LaTeX, which we immediately fully embraced. Since that time there was no way around or away from TeX.
DW: When was that that you took the LaTeX course, and what distribution of LaTeX, editor, etc., did you first use?
NP: That must have been in 1991 or 1992, at the beginning of my studies. At that time I did not even have my own computer (I bought my first computer for writing my Master's thesis), but was using some big Unix machine in the department on the students' terminals. I guess it was teTeX that was installed on these machines — not really sure. At that time, and continuing until now, for serious editing I use Emacs and for quick and dirty things I use Vim. The introduction to LaTeX was based on the introduction to LaTeX by Hubert Partl et al. from Vienna, which was the starting point for the now well known “The Not So Short Introduction to LaTeX2e”.
DW: Please tell me your evolution as a TeX user and then as a developer. (Can I assume that you write your logic papers or books using TeX?) Also, already several years ago you were involved with TeX Live and with the Debian distribution of TeX (cf., http://tug.org/TUGboat/Articles/tb26-3/tb84preining.pdf and http://tug.org/TUGboat/Articles/tb29-1/tb91preining.pdf), and my understanding is that in the years since you have become even more deeply involved with TeX Live.
NP: As a user I am pretty normal I would say. I use LaTeX for my day-in-day-out routine, write the scientific works as well as normal letters with it. Of course I used it for many other things, like automated generation of invitation letters for conferences and that kind of stuff — nothing spectacular. My situation is still now like that, I have never programmed any fancy macro package — only small ones for in-house usage, and most of them I wouldn't dare to publish.
More interesting is probably my early involvement with Omega. The background of that was my studies of Tibetology and Buddhism in Vienna. At some point I decided I want to be able to write Tibetan right into my editor and get decent Tibetan output. The standard for Tibetan is ctib by Oliver Corff, which I first used for quite some time. But I was not happy with the need for a preprocessor, and in addition I found some missing characters in the font. So I wrote support for Tibetan for Omega/Lambda, that is, some input OTP (Omega Translation Process) and style files, and called that otibet. I think besides myself nobody ever used it, but one never knows. Working on that I got to know the Unicode standard, and learned to fight my way through character encoding problems (in the early days of Emacs multilingualization). It also allowed me visit Japan for the first time, to present my work at M17N2000, the Fourth International Symposium on Multilingual Information Processing in Tsukuba. Since then my studies of Tibetan have unfortunately been put on ice. Later on I did some similar things, mostly to help friends publishing books in Devanagari.
Concerning my genesis as developer I just checked my mail archives, and I subscribed to the TeX Live mailing list in December 2001, I guess around that time I was starting to use TeX Live. Maybe around the same time I also started to contribute to the TeX Live project by building binaries for Linux on the alpha architecture.
The next step was getting involved with packaging TeX Live for Debian. I was at that time already using Debian a lot, and thought it might be nice to have TeX Live in Debian. The first proposal I could find was already a revised version of how to package TeX Live for Debian, and it was from January 2005. The preparations for that took quite some time, during which I also adopted the Debian Texinfo packages from Josip Rodin. The process took exactly one year, until in January 2006 when the first packages of TeX Live were accepted into Debian. At that time Debian was the first to provide packages of TeX Live, which turned out to be the way to go for all of them when Thomas Esser stopped working on teTeX. Since then Debian has included TeX Live 2005, 2007, and 2009.
My more profound involvement into TeX Live itself came at the EuroBachoTeX 2007. We, the TeX Live development team, got together to discuss changes of the TeX Live infrastructure — the scripts that drive the distribution. Up to that time most things were based on Fabrice Popineau's Perl code, which worked very nicely, but it was very hard to adapt and extend. But we wanted to get network installation and running updates working, so something had to be changed. After many hours of discussion at EuroBachoTeX and returning to Siena where I lived at that time, I — more as a proof of concept — implemented some minimal system in Perl based on the same ideas as the Package files of Debian (simple, easily parsable text files that are line based).
Well, since then much time has passed, and the proof-of-concept-code was never rewritten; but it was extended and is still mostly in use today. The code base has grown: the Perl modules alone are about 12,000 lines of code and comments, the installer another few thousands, and the always growing tlmgr amounts to close to 10,000 more lines of code. By now many different people have contributed ideas, suggestions, and code. The code base is now relatively stable. There are some ideas I want to implement, or have already in development versions, but that need long-time testing.
DW: Did you ever have any formal training in computer programming, or did you just pick it up on your own?
NP: In high school we had computer science, where we programmed a bit a kind of playground assembler, then turtle graphics, and Pascal. During my studies I had to program in various languages, including Fortran (don't remind me about that misconception) and C. During an internship I learned and programmed Scheme (a Lisp dialect) for a real-time simulation. Also during studies I did some operating system programming — low level communication, etc. But nothing that you wold count as formal training. Most of the languages I learned by myself by writing little programs. My first C++ program was a string library, followed by an implementation of string automata (something from logic/automata theory).
I can't remember if I ever programmed much Perl before starting with the TeX Live for Debian programming. There it was necessary due to the original modules being written in Perl. So I got the Learning Perl book, then the Programming Perl book, and at some point the Perl/Tk book and some others. So I learned Perl programming during the packaging of TeX Live for Debian (which is the reason the code is still such a mess there). That helped that the code base on the TeX Live internals rewrite became a bit better, hopefully ;-)
DW: Your website says that you work in many-valued logics (especially Gödel logics), linear orderings, Kripke Frames, intermediate logics, geometry and logic proof theory. Can you sketch for a lay person what that is about roughly? Also, is your career in that field mostly as a teacher or mostly as a researcher (or a combination). Also, my impression is that historically two kinds of people have studied logic: mathematicians and philosophers. Is that statement true and, if true historically, is it still a valid distinction today?
NP: My expertise as a mathematician or logician is mainly in many-valued logics, especially Gödel logics, and proof theory. The main difference between many-valued logics and what we call classical logic, the one used most of the time, is that the former ones are not restricted to only two truth values, “true” and “false”, like classical logic, but allow for different truth values. For example every number between 0 and 1 could function as truth value. Usage scenarios are many: think about modelling expressions of natural language, imprecise information, etc. This is an area that has a mathematical foundation just like classical logic; and the questions we are dealing with are quite similar to the ones of classical logic, like axiomatizability, decidability of various classes, semantics, calculi, completeness, etc.
A recent example by a colleague of mine is using a specific family of many-valued logics, which we call Gödel logics in honor of Kurt Gödel, in a medical expert system. I myself do not understand what is actually going on in this expert system, but together with colleagues we are developing the theoretical foundation and practical decision procedures for this system.
Kurt Gödel [http://kgs.logic.at/index.php?id=23, http://www.ias.edu/people/godel], after whom the logics we are working on are named, although not a very well known researcher, was one of the 100 most influential people of the last century as chosen by the Time magazine. Another nice story about him is that after his exile to the States he was in Princeton, and Einstein once said he is only going to the Institute to meet up with Gödel. Of course this is only a very small view onto the huge field of many-valued logics, but I will not write a text book in this interview ;-)
Concerning your other questions, my career is mostly research related, although I do some teaching (and am enjoying it). All my jobs have always been research jobs, first in Vienna as research assistant, than in Siena as Marie Curie Fellow, then back in Vienna as project leader, and now in Japan in a dedicated research center.
What you mentioned concerning philosophical and mathematical logics I think is still true. There is some interaction, especially if someone from the one field shows interest in the other field; but generally the development is completely independent, mostly because the questions are of fundamentally different types.
DW: Can you give me a pointer to an on-line version of a logic paper you have written?
NP: First Order Gödel Logics. M. Baaz, N. Preining, and R. Zach. APAL, 147:23-47, 2007. Preprint PDF: http://www.logic.at/people/preining/pubs/apal07.pdf
Continuous Fraïssé Conjecture. A. Beckmann, M. Goldstern, and N. Preining. Order, 25(4):281-298, 2008. Preprint PDF: http://www.logic.at/people/preining/pubs/order05-preprint.pdf
DW: Given your facility with computer programming, I wonder if computers play a role in logic these days and if you get to use your programming abilities in your work. Are computers used in logic proofs at all, or to represent complex models which give insights into what's going on so a proof can be found in the traditional (mental) method?
NP: This depends on the area of logic you are working on. I myself use the computer in my scientific work exclusively for writing up the results; the research itself is done on blackboards, papers, discussions. Other areas of mathematical logic make much more use of computers. Prime examples would be automatic theorem provers, proof assistant technologies, and rewriting. In these areas computers are also used for experimentation, proof search, etc. Besides that, there is a famous example from mathematics, the 4-colour-theorem, which was proven in 1976 with the help of computers [http://en.wikipedia.org/wiki/Four_color_theorem].
DW: I am sure few non-mathematicians have actually read and understood Gödel's incompleteness theorem(s). But at least in the United States as lot of people have heard about Gödel from Douglas Hofstadter's book, Gödel, Escher, Bach, which continues to sell well more than three decades after it was written.
NP: Yes, that was also one of my first contacts to logic. I know actually many people who have started to read this book, but most didn't finish it. It contains the core points of Gödel's incompleteness theorem, and although it is not a scientific work, I consider it a very valuable and well written book. Well, I guess I have to, since it was one of the reasons I started doing mathematics ;-)
DW: Let's turn back to your mountaineering. Can you tell me a story or two about difficult climbs, difficult clients, most fun climb, different geographies in which you have climbed, etc.?
NP: Uhh, there I could tell you sooo many stories that it would fill a small booklet by itself. I think the most difficult climb was in Chamonix, on the Aig. Verte, a climb called Gabbaroux-Silvy. I would never ever climb such monster myself, but it happened that over two years I was participating in an “International Mountain Leader Course” at the ENSA (Ecole Nationale de Ski et d'Alpinisme). I could go there because I was for many many years very active in the Austrian Alpine Club, with my own group of children, and also as trainer for new youth guides within the Alpine Club.
We climbed this route during a two week course called “Severe Ascent Qualification”, which I could participate in after having completed the initial course and the Winter Qualification. I think the name speaks for itself. My partners were a French mountain guide participating in ice climbing competitions, a Latvian semi-pro working as the head of the anti-terror unit of the Latvian National Bank (as far as I remember), and a Hungarian pro climber. I felt a bit out of place, well, a lot out of place between those crazy climbers. Anyway, after a relaxed start in the morning we went up to the Grand Montet and descended down to the Pic Sans Noms side. Here the fun started, 12 pitches of artificial rock climbing (means using technical equipment for movement, not only for protection). After that we reached a buttress on which we made a bivy. The following day we started at 4am, and climbed more or less without a break till midnight, all on ice and mixed ground, to the top of Aig. Verte. I remember reaching the final ridge, I was so exhausted, around me lightnings and thunder, I devoured an apple as fast as possible, but not really convinced that we would survive. Another seven hours and a long long descent later we reached the hut on the other side. Twenty-seven hours of strenuous exertion made us immediately collapse there.
I am very grateful to my partners; alone I would have never managed to climb up there. With them I had a very crazy experience.
DW: To what extent are you learning Japanese while in Japan?
NP: To an extent I consider slightly crazy. The last year I was about 3–4 times a week in Japanese course, plus a daily diet of 2-3 hours of studying. It occupied my free time in the last year to an extent that besides one small book (btw, very nice piece of Japanese modern literature, Banana Yoshimoto — Kitchen, a very calm story that does not let you stop reading), I didn't read anything but Japanese text books, Kanji books.
The difficulties with Japanese are manyfold, especially for a speaker of an Indoeuropean language. First of all one cannot read or write in the beginning. When I moved to Italy I didn't speak any Italian, but in daily life you also are permanently immersed into the language by seeing written words and immediately grasping them. In Japan in the beginning hardly anything is readable; you have to cling to the sounds. But the sounds are again very difficult for Indoeuropean speakers, as the distinction between long and short vowels is very important (the meaning changes completely). This is only to a very small extent the case in Indoeuropean languages. Most of the times mixing a short with a long vowel does not change the meaning there, but in Japanese people will hardly understand you.
Another problem is that the Japanese language is incredible rich on words. Here I am on very thin ice, since this is my personal experience with languages and I cannot back it up much with statistical data. But I believe that although English is a language with an incredible number of words, only a small amount of them are actually used in daily life. Counting how many words are used in newspapers and books I would guess the number goes down radically. In Japanese this is a bit different; there are many words and many of them are in use, making it hard to understand spoken language.
Then, of course, I am horrible in languages. I am good in structured and logical content, but things like vocabularies are a pain, and I have to hammer them into my brain on a daily basis using a flashcard system.
Fortunately my hard study has paid off. Now, after about one and a half years, I have reached a level where I can communicate with most people, often the whole day only in Japanese. Of course a discussion of the philosophical impact of Kant on the late literature in France (or whatever stupid example comes to your mind) is not something I can follow; but on the level of normal daily life communication, I am definitely able. Furthermore, I dedicate lots of time to the study of Kanjis, to the effect that by now I can take a newspaper and try to read some articles. I am still far from literacy, but reading is possible.
Of course all this work is not necessary. Many people live here for many countless years without ever reaching a decent level of Japanese. But here, as it is everywhere, language is the key to communication. And it gives me real pain if I cannot communicate with people. In Italy I loved to chat with old Tuscan farmers about olive trees or vineyards. They are so rich in experience and knowledge and have so many funny stories to tell. Here in Japan I want to reach the same level — still far away — to enable me to listen to their stories. At the end it is about understanding culture and the people, and without language you cannot do that.
DW: Thank you very much for taking the time to participate in our interview series.
Web site regeneration of April 24, 2011; TUG home page; search; contact webmaster.