Shlomi Fish’s Advice for the Young


There used to be a common meme to offer advice to high school and college students. Here is my advice, based on what I know, and practice:

“Live like you were going to die tomorrow; Learn like you were going to live forever”

This is the single best piece of advice one could give, but you’d amazed how many people defy it. (It’s attributed to Mahatma Gandhi.) Most people I know don’t seem to have time for anything. They are crowded with responsibilities. “We’re too busy”, they say.

I’m not too busy, and never will be. When I worked or studied I also was never too busy to talk to people I liked or met at that time; to work on open source software and web sites; to relax; to help people with their computer problems. I became relatively popular. More people knew me by name than I could recall them. (including some very pretty girls, albeit with them I usually recalled their names eventually). Everyone knew I was one of the resident UNIX gurus. It was fun.

I treasure the few people who always answer my E-mails on time, always have time to chat on the phone, unless they have a previous engagement. Know well enough that one doesn’t become productive by working more than 40 hours a week, and always say “I’ll do it when I have some spare cycles” instead of “I don’t have time to do it.”

You must be able to dedicate some time for your top priorities because you are going to die tomorrow.

And there are people who refuse to learn anything. They don’t have time and think that by not learning and acquiring new skills, they’ll still be successful. Java people think Java is the best thing since sliced bread and don’t learn Perl, Python or Ruby. C++ people don’t know how straightforward and painless Object-Oriented programming is in high-level languages. Programmers who are used to Perl 4 writing Perl 5 scripts the Perl 4 way, despite knowing Perl 5 pretty well. The list goes on.

Such persons are the worst. They may be pretty good at what they do, but they’ll never get better, and will maintain their code in a language that will become deprecated and then have to learn something new after not practising their learning skills for a long time.

Learn and Integrate the following things. By integrate, I meat take the time to digest and don’t reject as absurd on the spot. Note that you have to read them from cover to cover and not to skip in between, otherwise you’ll get a lot of false impressions.

Windows will be dead; Microsoft may or may not Die with it.

I put my money that practically the majority of people would switch to UNIX-based OSes (most notably the various distributions of Linux and Mac OS X) a few years from now. Linux nay-sayers are having less and less good arguments to support Linux that “just works”, is easier to use, requires less constant maintenance, and have so many technical advantages over Windows that enumerating all of them here would be not possible.

Just search for “Linux vs. Windows” or “Open Source vs. Windows.”. Windows is trying to catch up but usually gives solutions that are either buggy at first or not as integrated, or not as good or worse, or simply too little too late. In Microsoft conferences, the Microsofties are notorious for getting excited about new Windows features that Linux and other UNIXes has had for years. The Win32 API is horrible and very programmer unfriendly, while the various Unix APIs are in very good shape.

Windows Vista will come without any of the features that were most publicized. Development on Windows is slowing to a halt. Meanwhile development on the various components that make a Linux system is progressing at a greater and greater speed. There seems to be a huge shortage in clueful Linux workers at least here in Israel. No-one can find any. Linux is hot.

Microsoft will probably survive. It has .NET which is very portable. (either the proprietary version by Microsoft which runs fine on FreeBSD, or the open-source Mono.) It also has many other good products and some decent APIs that can be ported to Linux. And as Joel Spolsky notes it still has a lot of space to manuever, having a lot of money in the bank and being very profitable, and can still end up re-inventing itself into something else.

But Windows is dead. It does not mean it will not be used for years to come in various forms (just like Novell NetWare, DOS, Mac OS Classic, and other dead OSes can still be encountered in various places), but it will still be dead.

I’m not as paranoid as ESR is about what Microsoft will do next. I think they will simply embrace Linux, and port their software to run on it.

Why is all that important? Because you’d better learn Unix now rather than later, knowing that it is going to be the future of computing.

Learn Many Different and Orthogonal Technologies

You should learn the following languages and technologies:

These are languages I recommend you to learn. They were the languages that provided me with the most inspiration and insights. I don’t like all of them as much as I do others, but they are nonetheless truly enlightening. I believe Perl or Python (etc.) are the most useful of them, and also the most suitable for learning as a first programming language.

Another point is to learn about the design of digital circuits (out of logical gates and Flip-flops, not the transistor composition of them) and about computer architectures. Too many programmers don’t have a good understanding of a how a computer really works, which ends up showing in sub-optimal code.

Always reverse your growth death.

Growth death means you’ve stopped to grow, and become more cynical as time goes by and less capable. It is unnecessary because some people have never had growth death in their entire life time, while one of my friends already had it in his mid-to-late twenties. I never experienced growth death, and still grow in my skills, capability and productivity. You should too.

There are no caveats. Either you grow or you don’t. And nothing necessitates that you don’t grow. People in concentration camps sometimes never stopped growing, and often experienced spiritual or intellectual enlightenments.

It doesn’t matter how many responsibilities you have, you never have to stop growing. I have been corresponding with a man in his late 70’s , who is still programming, travels a lot, is very active, and is otherwise very enthusiastic about life. So he also still has a living growth.

Speak and Write Properly in English and in your Mother Tongue

This is something that Eric S. Raymond suggests in “How to become a Hacker”, but many people still ignore. Speaking in a language with many mistakes, will cause people to look down on you, and they’ll immediately notice the problems in what you speak. Whenever I write something in either English or Hebrew, I make sure to phrase myself properly. Whenever I learn of a new Hebrew common mistake, I immediately register it, and notice it from then onward. I never consciously write it, and try to correct myself if I utter it improperly. Other people whom I constantly remark on improper language, still make these mistakes, and I wonder how much conscious effort they put into their writing.

Note that all-lowercase writing, while acceptable on the IRC (albeit I do not write this way either) is completely frowned upon in E-mails, and leaves an impression of lack of professionalism, no time and care to press the shift key, and general disinterest. Don’t do it. Putting such improperly-written language on the Web is even more harmful. (And often mailing lists are archived on the Web for posterity).

Learn History in General and the History of Computing in Particular

Paul Graham recommends that people learn History but not so much political history, but cultural, technological and artistic history. He’s right in a way. Knowing who conquered who at what time is quite useless. Knowing how things developed and how people thought or behaved at the time, or what happened is actually useful and enlightening.

Many people graduate from Computer Science degrees, hardly knowing anything about Computer history. They are familiar with Knuth, Dijkstra and other great names, only throughout the algorithms they have formulated, and are not familiar with when things came to being, or why they are this way at all. There are many good resources for learning about computer history on the Net, and you should probably read them. Chatting on the IRC in the appropriate channels can also help you learn more about it.


Thanks to the Freenode people Darien, crimson_penguin, GXTi and quamaretto for commenting on earlier drafts of this essay.

Other Good Sources of Advice

  1. Eric Raymond’s How to Become a Hacker - an excellent document about the hacker’s way of life, attitude and practices.
  2. What You’ll Wish You’d Known - Paul Graham’s Advice for High School Students
  3. Paul Graham’s Advice for Undergraduate College Students
  4. Joel Spolsky’s Advice for Computer Science College Students
  5. Sriram Krishnan’s response to Spolsky’s advice
  6. “If” by Rudyard Kipling
  7. Everybody’s Free (to Wear Sunscreen). Just don’t take it too seriously…
  8. “Ecclesiastes” or “Qoheleth” - the scroll from the Jewish Bible - very old (~300 B.C.) but also extremely good.


Creative Commons License

This document is Copyright by Shlomi Fish, 2005, and is available under the terms of the Creative Commons Attribution License 3.0 Unported (or at your option any later version of that licence).

For securing additional rights, please contact Shlomi Fish and see the explicit requirements that are being spelt from abiding by that licence.

Meta Information

Written By:
Shlomi Fish
Publication Date:
Last Updated:
Creative Commons Attribution 2.5 License.