Noulakaz

The blog of Avinash, Christina, Anya and Kyan Meetoo.

Linux

On the necessity to master C

By 13 Comments

Photo from Flickr

C is an important programming language. It was created more than 30 years ago by Brian Kernighan and Dennis Ritchie and it is still used a lot especially to write systems software.

The Linux kernel, for instance, is written in C.

C is the new Assembly and this makes it an important language to master if one wants to be a proficient computer scientist.

In my opinion, there are two reasons why C is the new Assembly.

The first reason is portability. Assembly code is not portable and C is. Remember, C was invented around 1973 to write UNIX in a portable way. It is more sensible to learn C than to learn, say, IA-32 Assembly code especially it your time is limited.

Secondly, programming in Assembly is rarely done today because modern processors are superscalar and therefore reorder instructions. As a result, what you write is not what is being executed (or, more precisely, the order may be different). As far as I know, it is more efficient to use a good optimising C compiler (like gcc) than trying to manually optimise Assembly code on modern processors.

C allows someone to really understand how a Von Neumann computer works. And this is crucial to become a good problem-solver.

<rant>

At the University of Mauritius, the first language a student learns is C++. More precisely, it is a limited version of C++ where all the new features are ignored (STL(!), templates and generic programming for example). During the first year, the students only does procedural programming along with a little bit of object-based programming when doing the data structure module.

In my opinion, C would have been a much better choice. Sure, printf is not as easy to use as cout but C is so much simpler to learn. And, personally, I find it more elegant.

I really think that a first year University student should know all of C (and not a small subset of a more complex language). They should really understand what a pointer is and how recursion works.

Building a strong foundation in programming is important for someone who intends to work as a programmer for years and years.

The C Programming Language should be used as a textbook along with another good book on Data Structures.

One immediate benefit is that students learning graphics or systems software would have fewer problems using the standard libraries which are mostly C libraries (UNIX system calls, the pthread concurrency library, OpenGL, etc.)

Finally, instead of learning the syntax and semantics of a language, it would be much more interesting (not to say beneficial) to learn how to solve problems. One possibility is to teach students to solve the kind of problems given during the ACM Inter Collegiate Programming Contest (the easy ones at least…)

</rant>

Opinion?

(Addendum: While reading this post, I realised that another possibility is to teach C to make the student know how a computer works and, in parallel, use a higher-level programming language like Scheme to learn about how to build data structures. Opinion?)

Agile Web Development with Ruby on Rails

By 5 Comments

At last!

Amazon has just sent me the book Agile Web Development with Rails that I ordered one month ago.

This book focuses on Ruby on Rails, the open source framework to write web applications in the Ruby programming language (which is making a lot of buzz on the Internet right now).

Previously, I had purchased Programming Ruby.

I guess I can start experimenting with Rails a lot more now as I finally have the two books pictured on the right :-)

By the way, both Ruby on Rails and Ruby Gems, the Ruby package manager, have just been updated. I did this on my Linux box:

sudo gem update --system
sudo rm /usr/lib/ruby/gems/1.8/source_cache
sudo gem install rails

By the way, here is a tip I got from the Agile Web Development with Rails book on how to generate the Ruby on Rails API documentation automatically:

rails dummy
cd dummy
rake rails:freeze:gems
echo >vendor/rails/activesupport/README
rake doc:rails

A directory called api will be created into the doc directory. Move it to the desktop and, voilà, you have the complete API documentation.

Have fun!

(Image courtesy of Spuggy, source_cache tip from ImaPenguin and Ruby Gems tip from Gregory Brown)

I’ve moved all my emails to Gmail

By 6 Comments

I have moved all my emails (from 1997!) to Gmail using my Linux box using the following technique:

  • Saved all my emails in mbox format in Apple Mail (just do Save As… and choose the mbox format as export type). I guess all modern email clients can save in the mbox format now.
  • Set up a POP3 server on my PC at home. I used dovecot.
  • Create the user pop3user, give it a sensible password and copy the mbox file to /var/mail/pop3user
  • Configure the firewall (in my case, my Livebox) to allow incoming connections to my PC to the POP3 TCP port (port 110)
  • Configure Gmail to fetch emails from my Linux box (I used DynDNS to solve my dynamic IP issue. PS: Don’t forget that there is a built-in DynDNS updater in the Livebox. It works perfectly!)

And the result is:

Gmail has perfectly imported more than 1 gigabyte of emails. The only gripe is that this is a very lengthy process as Gmail only fetches 200 emails at a time and then pauses for some long minutes. I guess the whole import took around 3-4 days (at least).

I have to stress that the dates, the sender and the recipient(s) of the emails are not changed when Gmail imports them. This is extremely cool!

But the result is worth waiting for: I now have ALL my emails since 1997 online. I’m so happy :-)