Karl Berry

(completed 2005-09-27);

Karl Berry has been a long-time board member of TUG, became TUG president in 2003, and was subsequently elected for further terms. Among other projects, he is co-administrator of the tug.org server, co-editor of TeX Live, and a member of the TUGboat production team.



Dave Walden, interviewer:     You and I have been working together on various TUG activities for the better part of a year, so I won't pretend I don't know a good bit about you already. In fact, I'll assume that readers of this interview are already familiar with your biographical information that is already on the web:

I'll start with a hopefully straightforward question. I presume you are not just a TeX developer and also use TeX or its friends from time to time. Which of TeX and its friends do you use and for what types of writing, and what distribution(s) and editing system(s) do you use and why?

Karl Berry, interviewee:     I use pretty much all of the major TeX variations for one thing or another:

Regarding distributions, I almost always use TeX Live on GNU/Linux, sometimes even the bleeding edge development version. I occasionally use the teTeX that came with the operating system for testing, etc.

As for editing, I've lived in (tty-mode) GNU Emacs for 20 years and don't plan to change any time soon, if ever. I started using it because (among other reasons) its incredible redisplay algorithm actually made it feasible to edit over a 110-baud modem line. It's not an editor, it's a way of life.

DW:     You mentioned Eplain which you developed and Texinfo which you maintain. Can you estimate for me how many different TeX-related development and maintenance projects, large or small, you have worked on and perhaps list the three or four that you feel are most significant or you in some other way got the biggest kick out of?

KB:     Counting all the projects I've worked on, not just been a principal developer of, it's probably several dozen.

The most significant ones that come to mind are developing Kpathsea (a library for path searching), and integrating it with dvips, xdvi, and dviljk, and with Web2c TeX (a system for converting Knuth's original sources into C for compilation). In more recent years, I've become heavily involved in TeX Live and editing TUGboat.

The project I enjoyed researching and developing the most was probably the GNU Font Utilities, which was a tool chain to go from scanned images to outline fonts. Few fonts have ever actually been produced with it. I was amused to learn of one through a TUGboat paper — the authors had not contacted me while they were working on the project; I found it very impressive that they could actually get it to work. I was also happy to learn that the widely-used potrace originated from that code. I have always had a strong interest in fonts and typeface design.

DW:     Do I presume correctly that Kpathsea stands for “Karl's path searching?”

KB:     Yes, I'm afraid so. I didn't want to call it anything general like just “pathsearch”, and couldn't think of a better name. If I had it to do over again, I'd probably call it tpathsea (TeX Path Searching) or something else, but it's too late. Olaf Weber has done a fine job maintaining Kpathsea and Web2c for quite a few years now, so it should really be okpathsea. Which is a good description, too, come to think of it — it's not perfect, it certainly has its problems, but overall, I think it's ok.

The other circumstance in which my name sometimes appears is the “Karl Berry naming scheme” for fonts, which is those cryptic font names like ptmr8r which everyone (including me) hates. I try to just call it Fontname, but I guess that is too generic to catch on. It is strange to see my name in book indexes for this painful 8.3 naming scheme — not exactly what I would choose to be most visible for.

So it seems I'm not satisfied with either specific or generic project names. Naming is tough.

DW:     What was the first TeX development project (as opposed to use) you were involved with, and what caused you to become involved?

KB:     The first TeX development project I did was in 1984 or so, writing a DVI driver for a 300 dpi QMS laser printer; I started from the .web source for another driver (dvi2lgp from Lawrence Livermore, as I recall), so it was not terribly difficult. Doing this was essentially part of the very first TeX installation I did as a student at Dartmouth, since there was no other decent output device available, so it also counts as my first use. It was a lot of fun to figure out how to download fonts, etc.; this was pre-PostScript. Routinely available bitmap displays were still in the future; the standard interface to the VAXen we had was Zenith 19 terminals. I liked those terminals a lot.

At the time, all the TeX files were stored in a single directory by file type. It was a few years later, when I got (even more) interested in fonts, that I started updating the drivers I used, Dvips and Xdvi, as well as the Web2c-based TeX, to do subdirectory path searching, which eventually led to Kpathsea, the TeX Directory Structure, etc. Not sure it was the best approach to take in retrospect, but it was natural at the time.

DW:     How were you drawn into involvement in the TUG board and then as president?

KB:     There was no one big reason which impelled me to run initially for the TUG board in the mid-90's. I knew Sebastian Rahtz and others who were on the board at the time, which probably brought it to mind. I'd been a member of TUG for a long time already at that point, although it had always seemed a distant organization run by people I had no knowledge of, since I didn't go to conferences. So I wanted to try to contribute something organizationally, as well as technically.

About being president, I decided to volunteer for the job in 2003 because, first, no one else had expressed interest in taking it up, and second, in my personal life a little time had opened up, and I was curious to try something new. In 2005 I ran again, this time because, as you might guess, I had very much enjoyed the last two years working with the TUG board, other user group leaders, and most especially with Robin Laakso, TUG's executive director. She handles all aspects of the administrative side of TUG in most exemplary fashion, essentially single-handed.

DW:     In an earlier answer you mentioned that most of your communication is by email and in your last answer you noted that you didn't go to TUG conferences. These statements bring two questions to mind. First (and this is just out of curiosity), how many people in how many countries do you communicate with in the average week or month regarding TeX and TUG? Second (a more serious question), since most TeX users probably seldom if ever communicate in person with other TeX users except at their own sites, what are your approaches as TUG president to reducing the feeling of TeX users that TUG is a distant organization run by people they have no knowledge of?

KB:     Off the top of my head, I'd guess I talk TeX to maybe 50 people in an average week, in a dozen countries. I've never answered a TeX question from an Antarctica resident (though I have heard from a couple about the continent in general, as it's been another interest of mine), but TeX people from all other continents have been heard from.

The “distant organization” problem is one that is in my mind in nearly everything TeXish that I do, since I felt it so strongly myself. I don't have any magic solution. I just try to write semi-regular announcements, editorials, web pages, etc., in the most direct way I can, trying to get across that TUG exists because of and for TeX users, not as some organization apart from them. Especially in these days of megacorporations who apparently control nearly all aspects of our everyday life, there's a tendency to think every company is rich, huge and either merging or getting merged. None of these are the case with TUG!

TUG has no hierarchy, no bureaucracy; everyone on the board is a volunteer, as are essentially all the developers and everyone else in our TeX world, outside of the commercial implementation. TUG per se is a framework: an office to coordinate activities, finances, and outreach, and a bit of infrastructure, like the web server this interview is posted on. It always comes down to individual(s) putting in their time to make a given project actually happen. So I try to encourage and welcome involvement at any level.

DW:     As you said, TUG is a volunteer organization with few exceptions. Please tell me about any system, principles, or practices you have for getting new volunteers?

KB:     I wish I had anything systematic enough to be labeled a “principle”. All I really know to do is send email on whatever lists are appropriate and post notices on web pages. Mostly that reaches the same group of people (who are no doubt rather tired of my pleading by now), but occasionally someone new jumps in, which is always cause for rejoicing.

Also, especially through my editing so many articles in TUGboat, but also through the couple of recent TeX conferences I attended, sometimes I know an individual I can ask specifically who I think might be interested. Even if they don't have time or inclination, I've found everyone to be unfailingly gracious about these requests out of the blue. One of the things that keeps me going.

DW:     In addition to your development work and your administrative work as TUG president, you also put lots of effort into maintaining the TUG web site and also all sorts of systems related to other aspects of TUG's activities. Yet, there are good arguments and indicators that TeX's future has passed, and clearly its time as a leading edge system is perceived by most of the world as passed, if they ever heard of it at all. What is it about TeX and TUG that keeps you working so hard on their behalf?

KB:     Much of the answer lies with Don Knuth. When I first read tex.web, The TeXbook, and his other articles, programs, and books (TeX-related or otherwise), something in me was very responsive to his goal of making a beautiful program that produces beautiful output — and then giving it to the world. I never met him (until very recently), but that didn't matter. Then when I came across Richard Stallman's writings a couple of years later, I realized again that I felt, personally, that free software like TeX simply has an ethical basis which proprietary software does not. I also appreciate and enjoy typesetting, book design, and letterforms, so working on TeX seemed a natural place where I could contribute to the world of free software.

Although those are reasons why I started with TeX, they are also reasons why I continue with it. The computer world (and the non-computer world, for that matter) has become increasingly dominated by a very few large companies, as I mentioned. TeX resists that; it is simultaneously very idiosyncratic and very useful, and the standards mavens hate that. I love it.

As well as such philosophical reasons, TeX continues to have a large user community; interacting with users, and being one myself, are certainly big positive factors. I would be uninterested in TeX if there weren't always new problems being posed and new users, and developers, arriving.

Ultimately, typesetting combines artistry and practicality, esthetics and communication. For me, it's been a fine way to spend the last 20 years or so. Maybe it'll remain viable for 20 more.

DW:     Your GNU Friends interview tells us a good bit about your personal life up to three or four years ago. Will you please give a bit of an update? Also, are there other organizations you use your technical skills on behalf of and, if so, can you give us an example or two?

KB:     The other primary group I spend time volunteering for is the GNU Project. Besides maintaining GNU Texinfo, GNU Hello, and contributing to other packages, I work a fair amount with rms, evaluating new software packages offered to GNU, keeping old projects alive when maintainers move on, administering lots of GNU mailing lists and the GNU coding standards and maintainer documents, etc.

Locally, I do a little volunteer work for the library, currently including hosting a tiny web page about our audio book club. (Especially as audio books, some of my favorites have been The Memory of Running by Ron McLarty, Prodigal Summer by Barbara Kingsolver, and Hearts in Atlantis by Stephen King, all read by the authors.) I also try to help out the great local architect and contractor who built our house, Rick Howard, by running a small web site for his family's business.

As for my non-technical life, the biggest change was moving to that house on the Oregon coast full time last spring — here are a sampling of recent pictures, mostly taken by my wife Mare. However, soon we'll be coastal part-timers again; we're making plans to spend a good chunk of time each year in Seattle, to be closer to our grandson Benjamin as he grows up.

DW:     I've spent a lot of time on the Oregon coast and in Seattle — they are both beautiful, and Seattle is not exactly without coast line. In any case, thank you very much for taking the time to do this interview. You've filled in a lot I personally did not know, and I trust for other readers as well.

KB:     Thanks for giving me the opportunity — and for all the work (including this interview series) you yourself do for TeX and TUG.

Web site regeneration of January 5, 2010; TUG home page; search; contact webmaster.