re: Learning to Read!

anandeep — Tue, 21 Nov 2006 18:33:33 GMT

Wesley,

Thanks for your comments.

I have actually read both the books you mention and have the OS book on my shelf in the office! The other one is a prized possession in my home library.

It is interesting that Tanenbaum's Minix was in the tradition of Linux. I remember reading the C code - but dont think I could have waded through it without the commentary in the book.

There hasnt been much CS support for reducing complexity in reading code - techniques and tools that work havent seen the kind of attention that they deserve IMHO. That's why I was so excited to see Kothari's work.

My theory is that the community aspect of open source reduces the complexity of code reading, since the forums and individuals (not to mention O'Reilly!)coach you on what to read. Reading code where there isn't this community around you would be a very onerous and complex task.

re: Learning to Read!

Wesley Parish — Tue, 21 Nov 2006 02:36:55 GMT

This topic - code reading - is mentioned in at least two of the major computer science textbooks in my personal library - Tanenbaum's "Operating Systems: Design and Implementation" and Aho, Sethi, and Ullman's "Compilers: Principles, Techniques, and Tools".

Both writers were writing at a stage when the commercial culture of source code secrecy had almost succeeded in wiping out code reading in tertiary institutions. And both of the books regret this.

As far as training, etc, goes, a little bit of code reading goes quite a long way. For a newbie, seeing what such-and-such a snippet of code does, gets him or her in the frame of mind where it is possible, therefore achievable to write usable programs.

Code reading to start maintaining a source tree that has been unused for a while - now that is a major challenge, particularly if it isn't well documented.

Since that was the case with the Y2K bug, it should be more of a priority among programmers and other software professionals.