Saturday, October 20, 2007

.signature

One man's constant is another man's variable. -- Alan Perlis


Alan Perlis was one of the finest specimens of the Real Programmer breed. Back in the days where Computer Scientists didn't exist, he and his kind were responsible for making many of the decisions that shape our view of computers today. I'm particularly fond of Perlis because of his views on Compuer Science:

I think that it's extraordinarily important that we in computer science keep fun in computing. When it started out, it was an awful lot of fun. Of course, the paying customers got shafted every now and then, and after a while we began to take their complaints seriously. We began to feel as if we really were responsible for the successful, error-free perfect use of these machines. I don't think we are. I think we're responsible for stretching them, setting them off in new directions, and keeping fun in the house. I hope the field of computer science never loses its sense of fun. Above all, I hope we don't become missionaries. Don't feel as if you're Bible salesmen. The world has too many of those already. What you know about computing other people will learn. Don't feel as if the key to successful computing is only in your hands. What's in your hands, I think and hope, is intelligence: the ability to see the machine as more than when you were first led up to it, that you can make it more.
The Structure and Interpretation of Computer Programs by Abelson, Sussman, and Sussman

Unfortunately, things haven't quite turned out like Perlis would have wanted.

Besides of his many contributions to Computer Science, such as his work on ALGOL, Perlis is very well known for his Epigrams on Programming, of which our quote is the first one. I like this quote because it reminds me that there can never be an ultimate truth in programming due to our human condition.

Wednesday, October 03, 2007

Nerd Food: Interview with Federico Mena-Quintero

Pretty much anyone who is involved with Free Software - even just as a lowly user like myself - has heard of Federico. His blog is a source of insightful ideas on Gnome, and lately, on performance - combined with a healthy dose of interest in politics and, more importantly, good food. I decided to send a few questions to Federico, mainly on the topics I was most curious about, and he kindly replied to my questions - and did so in record time! Many thanks to Federico for taking time off his busy hacking schedule for this interview.
(C) Gnome Foundation

1. You are one of the founders of the Gnome project, which is currently celebrating ten years of existence. On a recent interview you gave to Fosdem, you considered the platform to be maturing. However, as we all know, the last 10% normally take 90% of the time, and it's considered to be boring work. What do you think the Gnome project needs to do to get people to focus on those remaining 10%?


Basically, to provide an incentive to get that last 10% of the work done :) Instead of smacking people with a stick for not writing documentation, you could have a web page with a bar chart of "percentage of documentation coverage". Then it becomes a competition: use a carrot instead of a stick.

I'd also like companies to get more involved in this. If they want to ship GNOME as a development platform they support, then they could very well employ people to do those missing bits.

2. You have been one of the champions of performance in Gnome for a while now. As functionality increased, Gnome started suffering more and more from performance problems, particularly when looked at from a low end perspective. You have been trying to explain to the masses that performance work is interesting. What do you think can be done to increase developer focus on this neglected area?

The thing about fixing performance problems is that nobody teaches you how to do it. There is very little documentation out there on how to generically approach an optimization problem (I intend to do something about this, but oh, time, time, time!) :)

Also, sometimes you fix a performance problem, but it reappears in the future. This happens when you don't leave an infrastructure in place to let you run a benchmark periodically. You need to be able to see if there are performance regressions.

Our tools are slowly getting better, but there are really very few people working on optimization and profiling tools. It takes a *ton* of time and skill to write a good tool; maybe that's why there are so few of them.

Finally, profiling and optimizing is really about following the scientific method ("make a hypothesis, change one thing at a time, measure, confirm your hypothesis, etc."). This requires discipline and a lot of patience.

Basically, it's a problem of education :)

3. Earlier on this year, Gnome users and developers met for GUADEC. Did you find the conference as productive as in previous years? How important is GUADEC for the Gnome user and developer community?

Yes, this GUADEC was tremendously productive! I think the venue helped a lot; the Birmingham Conservatoire is rather compact and has nice practice rooms that anyone can use. So, you could grab a couple of hackers and go to a room to hack peacefully.

GUADEC has always been important, even more so now that our community is large and widespread. It is about the only time in the year when most of the GNOME contributors get together in a single place and are able to talk in person. Do not underestimate the productivity of talking over a beer :)

4. From the outside world, it appears Novell is a company who has regained it's soul and direction with Linux. How was the transition from Ximian into Novell?

Like all acquisitions, it was a bit rought at first. It's what you get when you switch from being in a small company where you know all of the employees, to one with several thousands of people. You have to adjust to bigger processes, more layers of management, new locations, new paperwork...

It has been very interesting to see the mindset of the old-time Novell people change over time. At first they seemed reluctant to touch Linux and free software, since they were of course Windows users. Then we had a period with lots of questions, lots of bugs that needed to be fixed, lots of re-training... and now we are in a very nice period, when people have accepted that we must all use our own free software. People seem to be productive with it and happy.

I miss the monkeys, though.

5. You are currently telecommuting from Mexico, a position envied by a most developers out there. Do you find that telecommuting helps improving your productivity? Are there any downsides to it?

It has good things and bad things. Good things: working in your pajamas if you feel like it, not having to commute, taking a pause when you are stuck in a hard problem to do a bit of gardening. Bad things: you can't talk to people in person. You must fix all your networking problems yourself. Sometimes, when you are uninspired, it's nice to be able to look over someone else's shoulder or talk to them.

6. Can you describe your typical day at work?

Well, since I work from home... :)

I wake up. If my wife and I are hungry, we make breakfast while my email gets downloaded. If we are not hungry, I'll just check for super-urgent email and then start programming (fixing bugs, doing new development, reviewing patches, etc.).

I usually try to get some programming done in the morning, while my brain is fresh. Processing your email in the morning is a really bad idea; it will take you up to the afternoon and by then you'll be tired to really write code.

We have lunch at really irregular hours. Sometimes it's more like an early dinner. I have the bad habit of not stopping working until I'm exhausted or my wife is angry that we haven't gone out to the supermarket yet, but I'm trying to fix that :)

In the afternoon I tend to do "light" work... maintaining wikis, answering email, coordinating people. I don't really have a fixed work schedule.

7. Many developing countries are increasingly looking at Free Software as a way to bring down the digital divide. Do you find that Mexico is taking advantage of Free Software - particularly since it has two lead Free Software developers? Are there any lessons to be learned from Mexico's experience?

Mexico is blessed and cursed to be so close to the USA. There is plenty of basic usage of free software by individuals (often enthusiastic students), but relatively little usage in the public and private sectors.

People in Mexico get very impressed by rich people; most Mexicans want to be like the rich people from the USA they see on TV. It's very easy to woo us into accepting their ways.

So, every time there has been some noise about using free software in the public sector, Bill Gates has flown down, organized a big business lunch with government officials, and made sure that they keep using Microsoft products. If you are an ignorant politician, you will love to gloat that you had lunch (imagine, lunch!) with Bill Gates, the richest man in the world --- and whatever he says must be correct, of course. The problem we have is that most of our politicians don't have the faintest idea of the economic and cultural implications of free software, unlike those in the European Union (see the recent report on the economic impact of free software there!).

Thanks for the interview!