Re: software engineering and the notion of authorship

[ok@goanna.cs.rmit.edu.au (Richard A. O'Keefe)]

fjh@murlibobo.cs.mu.OZ.AU (Fergus Henderson) writes:
>>Yeah, but the real problem comes 10 years down the road, when the experts
>>have quit or died or moved to other projects and plain do not remember the
>>wonderfully complicated code they created back in '96... 

>If the code is clear, well-written and well-documented -- you should
>use code reviews to ensure this -- then I don't see why attaching
>authorship labels or assigning clear areas of responsibility is going
>to cause you problems 10 years down the road.


Fergus and I see eye-to-eye on many points.
However, he is working in a project with a small number of brilliant
programmers who are good at writing comments.

I had to maintain a large chunk of code which had only one comment (well,
I tell a lie, there were two: a commented out chunk of code -- with no
reason -- and the copyright notice) and a lot of one or two letter 
identifiers.  And code reviews were not a political practicality with
that piece of code:  it had been written by the _founder_ of the
company before I joined.  And he had left.

The "ego-less culture" that allows programmers to adopt a common style
is the same culture that is needed to allow the founder to submit his
precious work to other people's judgement.  I'm not arguing against
personal responsibility, only that the responsibility is responsibility
to *communicate*, and that adopting a common style in a project is one
of the things that *makes* the code clear.

-- 
Fifty years of programming language research, and we end up with C++ ???
Richard A. O'Keefe; http://www.cs.rmit.edu.au/~ok; RMIT Comp.Sci.