[9fans] XP (was: code complexity)

[9fans] XP (was: code complexity)

[iwanek]

On a related topic, what does people on this list think of
Extreme Programming (XP)?

XP is supposed to encourage

- writing readable code through pair programming,
- avoiding code bloat by refactoring,

but I have a feeling this is just a literature. I, for one, enjoyed
and profited from reading "The Practice of Programming" so
much more than reading any XP documents so far.

Is anyone practicing XP? Does it help at all to reduce code
complexity?

- kazumi

Re: [9fans] XP (was: code complexity)

[Boyd Roberts]

iwanek@nttdata.co.jp wrote:
> 
> On a related topic, what does people on this list think of
> Extreme Programming (XP)?

I think it's the bleeding obvious that's turned into a religion.

> Is anyone practicing XP?

We do, but not to the point of zealotry.

> Does it help at all to reduce code complexity?

Yes, 'cos someone will say:

   Why did you write that complex junk?

Or the complex junk never gets written in the first place [pair programming].

Re: [9fans] XP (was: code complexity)

[Laura Creighton]

One reason that some people do not sit down and make their code
simpler is because nobody has ever told them to do that before.  At
any rate, when I insist on it, I keep getting told that I am the first
person who has.  The ones with bad attitudes are helped when they
notice that other people are also simplifying.  I think that a lot of
people are taught that 'Only the stupid people in class ever have to
write an answer over again.  Smart people get it perfectly correct the
first time.'  So they are badly offended when you tell them to do it
over again, because they think that you are telling them that they are
stupid fools.  (I have heard a variation from students which goes 'if
I had wanted to revise things, I would have become an Humanities
major'.)

I also think that polite people learn to not mention that _other
people_ have poorly written code, until they cannot notice that _they_
have written poor code.  We had student interns here last summer. Very
early on, the first week, I held a code review of some code I had
written in one hell of a hurry. And it was many, many hours before
they felt that it was in some way, _permitted_ for them to point out
flaws in my code.  Finally they got the idea.  And really liked
savaging their elders for carelessness, and ugliness, and complexity,
and all sorts of good stuff.  But first they had to know that we
really wanted to hear this.  Before that it was disrespectful.

Laura Creighton

Re: [9fans] XP (was: code complexity)

[Steve Kilbane]

Laura wrote:
> And it was many, many hours before
> they felt that it was in some way, _permitted_ for them to point out
> flaws in my code.  Finally they got the idea.  And really liked
> savaging their elders for carelessness, and ugliness, and complexity,
> and all sorts of good stuff.  But first they had to know that we
> really wanted to hear this.

I get this problem in prose reviews too. I try to counter it by
explaining to people that I want to hear both what's good and
what's bad: the former, so that I can choose to do it again, and
the latter so that I can choose to avoid it. But in both cases,
it needs to be constructively critical - *why* is it good/bad?

Another tack, which depends on your environment, is to point
out that it's much better for a colleague to point out you've done
something stupid than a (soon-to-be-ex-) customer.

steve

Re: [9fans] XP (was: code complexity)

[Boyd Roberts]

geoff@collyer.net wrote:
> I haven't read enough about XP to comment on most of it, but I dislike
> the idea of pair programming.

Depends of the circumstances.

> There's nothing wrong with someone else
> reading my code, but I need some quiet and calm to program, so having
> someone nattering in my ear, second-guessing me is going to be a
> disaster.  And how is the other member of the pair selected?

Well yeah, that is a problem.  At least here we're a small group and
are going in the same direction.  Pair programming in a large group
[real world] would be a disaster (unless you like being lectured or
yelled at :).

We don't do it that much, but sometimes we do, when we want to, with
who we want to.  Then again, I guess I am pair programming right now
with a guy 30km away [we are discussing protocol implementation issues].

> Go it alone?

I'd go for the 'go it alone' but call in for help when I/you need it.

Re: [9fans] XP (was: code complexity)

[geoff]

I haven't read enough about XP to comment on most of it, but I dislike
the idea of pair programming.  There's nothing wrong with someone else
reading my code, but I need some quiet and calm to program, so having
someone nattering in my ear, second-guessing me is going to be a
disaster.  And how is the other member of the pair selected?  Having
someone you don't like or don't agree with as your partner would be
even more galling.  It would be like being part of a team that isn't
self-selected but imposed by management (actually a ``team'' in that
case, since it's not real).  And how do you program at home or when
travelling?  Pick some random person as your partner?  Go it alone?