In 2015 I had the very fortunate opportunity to invest some time in my skill as a programmer. At this time I had an undergraduate degree (Bachelor of Science) double majoring in software and computer networks. I had also completed the first two thirds of a masters degree, also in computer science. I chose not to complete the thesis because I felt I did not have anything to add to the field, graduating with a diploma of Computer Science (pass with high distinction). It was during this decision that I became acutely aware of my lack of programming ability.
I picked up Kernighan and Ritchie [KnR12] and started from scratch. Herein started a two year apprenticeship of software development.
Why an ‘apprenticeship’? I did not start out with much of an idea of what I was trying to achieve. I just knew that if I was going to get any good at this programming business I needed more experience. My preferred style of learning is book reading, and I had done plenty of this at uni. But it was also self evident that you don’t learn to code by reading books. Well why not actually do the exercises. Then as long as I picked up the right books I would be assured success, right?
Over the next two years I would read the best part of 53 computer science text books and complete the majority of the exercises in any that had them. During this time I read a number of books and blog posts about software craftsmanship. Say what you will about the movement the idea of an apprenticeship struck a chord with me and I started to view what I was doing as such. Text book authors were my masters and I was there to learn.
This post is intended for three people.
- Software developers wishing to understand the path I chose.
- Potential employers, wishing to understand my experience.
- Family and friends, wishing to know what I was up to all that time locked away.
There are many ways to achieve ones goals mine is but one. Each individual is different, we learn differently and we engage in different ways. I do not in any way suggest that this is the best path. It is simply the path I chose.
The Beginning
As stated above the first text I pick up was KnR. To this day, in my opinion, this is one of the best technical books ever written. If there is perfection in printed text then KnR is it. My first obstacle was to be able to concentrate for long periods of time doing mentally taxing work. At first if I remember correctly I could only concentrate for a few hours each day. It was at least three months before I could complete a full work day and I still remember with glee my first 12 hour day.
During this stage I committed to organize my time as if I was a remote employee. I would endevour to work a normal 40 work week and I would log work session times and topics. (Though during some of the more obsessive phases I was coding much more than this.) If you are interested the logs are all online. At first I started logging by hand. The logs from the start are incomplete. I believe that I worked on KnR from May 2015 through July. During this time I also separated from my wife. Be careful kids too much coding can effect your social life :) During this time I also settled on a development environment. For the record the stack I settled on is:
Emacs, Zsh, Terminator, Mutt, Firefox, Git, and GCC.
For the record I have used in the past Ubuntu, Debian, Slackware, and Arch. I now use Ubuntu.
UNIX Programming
The week before I started undergrad, in a fit of enthusiasm, I went to the store and bought a copy of Advance Programming in the UNIX Environment (APUE). This was the 2nd edition. Prior to starting university my total computer experience consisted of: A term writing logo in year 9. A month or two writing a text based game in DOS and two months playing around on the internet (primarily on ‘hack this site’). Oh and there was the time I deleted my Mums C: drive when I was a kid. Suffice to say I didn’t clock up any hours in front of the keyboard after that. To say that APUE was a little advanced for me is an understatement. I however had a number of attempts to read it over the years with varying levels of success.
In August I commenced work on APUE for real (I even bought the new 3rd edition [SnR13]). In parallel I decided to work on another classic UNIX text, UNIX Systems Programming (USP) [SFR04]. By this stage I was routinely working 12 hours a day stopping only for food breaks. I did this 4 days a week and played with my kids the other 3 days. It appears, from the logs, that I also started to read Knuth during this time. I included all reading in logged time as time worked.
On September 20 I picked up Learning Perl [RS11]. I clearly remember wanting to have something else to work on to break up the C systems programming. At this stage there were some low moments when things got rather arduous (SysV IPC I’m talking to you). On the 2nd of October I completed APUE [SnR13]. Every chapter and every exercise. I didn’t make it through the last section of USP [SFR04], I can’t remember why. As soon as those two were finished and while still working on Perl stuff (moved onto Intermediate Perl [RS12]) I started UNIX Network Programming (UNP) [SFR04]. This is anther classic UNIX text, Stevens really is a heavy weight in the field.
October saw the end of Knuth volume 1 [Kn97v1]. It seems I started volume 2 [Kn97v2]. The logs don’t look like I finished it all in one go but I did finish it. When I say finish I mean my eyes passed over all the words, how much of it went in I cannot say :) Eventually I completed volume 2 and started volume 3. I never did make it far into volume 3 though.
Security
My next focus would be security and cryptography. Here I made my first of two attempts at the security programming challenges (previously called ‘Matasano’) now online at cryptopals. The first attempt was using Perl. Wednesday the 20th October I completed UNP [SFR04]. I picked up Schneier [BS15] and got cracking (no pun intended). By mid way through November I came to the conclusion that Python would be better suited to the security challenges and would also be a useful tool to have in my kit. I ordered Python Essential Reference [DB09] and started work on Python the Hard Way. On the 23rd I restarted the cryptopals challenges from scratch in Python.
Enjoying the crypto challenges, in December, I started working in parallel on the Eudyptula Challenge. I also threw up a few solutions on mathematical programming challenge site Project Euler. In December I read Version Control with Git [LMC12] and Beginning Linux Programming [MnS09].
Open Source Contribution
January saw a loss of direction. I read a book on Web development [RS13], continued cryptopals and made my first attempt at functional programming with The Little Schemer [FnF96] and The Scheme Programming Language [KD09]. January also hosted what turned out to be a pivotal text Clean Code [RM09] by Uncle Bob. This book had a massive effect on me and drives my software development to this day. If I had to pick one book from the whole apprenticeship to define as essential this would be it. January also saw the start of my first ever open source contributions. This was to the Open Bazaar Server project written in Python.
I spent a bit over a month solid working on Open Bazaar. Development was done via GitHub and I learned a lot from the experience.
Mid way through February I stopped working on Open Bazaar and commenced learning Golang via The Go Programming Language [DnK16]. During March I also read up on my shell of choice zsh in From Bash to Z Shell [KPS05]
It was during the months around this time that I thought it a good idea to touch up on my mathematics skill. I can’t say I was extremely successful (I measure success by having completed a large portion of a text) but I did attempt multiple texts from various branches of pure and applied mathematics and even learned a thing or two. With similar brain developmental ideas in mind I attempted to learn the violin. I practiced an hour a day for a few months eventually giving up on the idea along with the math study.
Systems Programming
On the 12th of March something drove me to pick up The Linux Programming Interface (TLPI) by Michael Kerrisk [MK10] and get back into systems programming in C. I should mention here that an undercurrent driving me since January 2013 was a desire to be a Linux Kernel developer. This desire was born out of attendance of the conference LinuxConfAu in Perth 2013. Since then I had hoped to become a competent enough programmer to try my hand at Kernel development. More on this later. This explains my choice to do more C coding even though the Kerrisk text is very similar in content to APUE, amusingly it was SysV IPC that stopped me in this text also. I completed every chapter and exercise up to, and including, chapter 44. Kerrisk is an extremely competent writer and I agree with those who say that the mantle has now passed from Stevens to Kerrisk when it comes to leaning systems programming in the *nix environment.
I continue working on TLPI until the end of April. Then over the next few months I would read the following operating systems texts
- Modern Operating Systems [AT14]
- Linux Device Drivers 3rd Edition [CRK05]
- Essential Linux Device Drivers [SV08]
- Linux Kernel Development [RL10]
In May I would add ARM Assembly Language [WH09] to the list of completed texts (including most but not all of the programming exercises).
Functional Programming
June 2016 would see the start of my second functional programming block. This included re-reading The Little Schemer [FnF96] and re-working all the code. I then started one of the highlights of the apprenticeship The Structure and Interpretation of Computer Programs [ASS96]. This text needs no introduction and for very good reason. I make no excuse when I say that this book was tough going. The text is comprised of 5 chapters. I worked every exercise in the first 4 chapters. Chapter 5 implements an interpreter but by the end of chapter 4 I was getting lost. I did not manage to complete chapter 5. During this phase I used GNU Guile Scheme and got a small patch accepted into the project. The patch was a fix to a known bug and removed an erroneous compiler warning that was emitted every time the REPL started. The patch was in C and Scheme.
August brought more functional programming with Learn You a Haskell For Great Good and Real World Haskell [SGS09] although I did not complete Real World Haskell. Functional programming continued into September with work on Ninety-Nine Haskell Problems and CIS194. It was during CIS194 that I hit a wall with functional programming and decided to move back to world of imperative codes.
Algorithms
By the end of September I was writing Golang again with a second reading of The Go Programming Language [DnK16] in unison with work on the very entertaining HackerRank.com. I would spend the next three months doing algorithms problems on this site eventually making it to the 98th percentile (of approx 600 000 coders). During this algorithms block I would read the at times daunting Introduction to Algorithms [CLRS09]. I also made study of The Algorithm Design Manual [SS12] and the online text Open Data Structures.
And on into the Kernel
Mid January 2017 was linux.conf.au in Geelong. And, as was now customary, attendance got me super psyched on Kernel development. On returning from the conference I dropped work on algorithms and started Kernel programming. This included re-reading Linux Device Drivers [CRK05] and making a (short lived) attempt at writing a kernel from scratch.
January though to mid March saw various Kernel development tasks. I mistakenly did some what I later learned were very rude clean patches into the core kernel. PowerPC got to feel my newbie wrath with some clean up patches and also patches into the GNU Assembler for PPC assembly directives. Continued effort was applied to the Eudyptula challenge also.
At some stage I started to focus exclusively on the KS7010 Wi-Fi SDIO driver. During this time I started to share what I had learned about Kernel development with other upcoming newbies via the Kernel Newbies mailing list. Somewhere here I got a tutorial presentation on Kernel hacking accepted into Open Source Summit North America. It was these events and my continued work on the Kernel that eventually lead me to the conclusion that my apprenticeship was at an end. But like it says in Apprenticeship Patterns [HO10] in was not until much later that I realized that these were the closing days.
There were numerous other books I read during this time that didn’t get a mention. Honourable mention goes to The Pragmatic Programmer [HnT99]. For a full reading list see ~/books.
I know spend my days hacking on the Kernel. If you got this far thank you for your time. If this post is in anyway useful to you I am happy. May your days be full of hard tractable problems.
Bibliography:
[ASS96] Structure and Interpretation of Computer Programs Harold Abelson, Gerald Jay Sussman, Julie Sussman
[AT14] Modern Operating Systems Andrew S. Tanenbaum
[BS15] Applied Cryptography Bruce Schneier
[CLRS09] Introduction to Algorithms Thomas H. Cormen, Charles C. Leiserson, Ronald L. Rivest, Clifford Stein
[CRK05] Linux Device Drivers 3rd Edition Jonathan Corbet, Alessandro Rubini, Greg Kroah-Hartman
[DB09] Python Essential Reference David M. Beazley
[DnK16] The Go Programming Language Alan A. A. Donovan, Brian W. Kernig
[FnF96] The Little Schemer Daniel P. Friedman, Matthias Felleisen
[HnT99] The Pragmatic Programmer Andrew Hunt, David Thomas
[HO10] Apprenticeship Patterns Guidance for the Aspiring Software Craftsman - David H. Hoover and Adewale Oshineye
[KD09] The Scheme Programming Language R. Kent Dybvig
[Kn97v1] The Art of Computer Programming (volume 1) Fundamental Algorithms - Donald E. Knuth
[Kn97v2] The Art of Computer Programming (volume 2) Seminumerical Algorithms - Donald E. Knuth
[KnR12] The C Programming Language Brian W. Kernighan and Dennis M. Ritchie
[KPS05] From Bash to Z Shell Conquering the Command - Oliver Kiddle, Jerry Peek, Peter Stephenson
[LMC12] Version Control with Git Jon Loeliger, Matthew McCullough
[MK10] The Linux Programming Interface Michael Kerrisk
[MnS04] Beginning Linux Programming Neil Matthew, Richard Stones
[RL10] Linux Kernel Development Robert Love
[RM09] Clean Code, A Handbook of Agile Software Craftsmanship - Robert C. Martin
[RnR03] UNIX Systems Programming Kay A. Robbins, Steven Robbins
[RS11] Learning Perl Randal L. Schwartz, brian d foy, Tom Phoenix
[RS12] Intermediate Perl Randal L. Schwartz, brian d foy, Tom Phoenix
[RS13] Programming the World Wide Rebert W. Sebesta
[SFR04] UNIX Network Programming W. Richard Stevens, Bill Fenner, Andrew M. Rudoff
[SGS09] Real World Haskell Bryan O’Sullivan, John Goerzen and Don Stewart
[SnR13] Advanced Programming in the UNIX Environment W. Richard Stevens, Stephen A. Rago
[SS12] The Algorithm Design Manual Steven S Skiena
[SV08] Essential Linux Device Drivers Sreekrishnan Venkateswaran
[WH09] ARM Assembly Language William Hohl