Re: Hungarian notation

[cwr@cts.com (Will Rose)]

Richard A. O'Keefe (ok@goanna.cs.rmit.EDU.AU) wrote:
: bobduff@world.std.com (Robert A Duff) writes:

: >P.P.S. My take on comments is: Comment (only) where necessary.  Whenever
: >you can express some fact in the programming language, it's better to do
: >that than to write a comment.

: I like reducing things to slogans.
: I haven't been able to come up with a good one for this.
: What I say is
: 	"Your comments should tell a human reader precisely the
: 	 things s/he needs to understand the purpose and operation
: 	 of the module but are not obvious from the code."
: This is the right stuff, but it needs to be said better.
: Can anyone come up with a really memorable way of expressing this?

"Comment at one level above the code" - trouble is, this is a bit
cryptic and needs a comment to explain it...

: >Any time you feel the need to write a
: >comment, that's a failure of the programming language -- it can't
: >express what you need to say.

: I don't know that I agree with this.  How can the code make it clear
: why you chose *not* to do something?  How does the code describe
: the space of possible tradeoffs from which you selected this particular
: point and why?

I strongly disagree with the 'failure of the language' argument; I tend
to write comments at an intermediate level between the code and the
reader.  Obviously the code is written as clearly as possible, since
I have to debug it, but above that level it's handy to have reminders
of just what in the world is going on in a broader sense, and why.

The next level above comments is probably the manual, but that seldom
gets written.

Will
cwr@crash.cts.com