[Tutor] extracting numbers from a list

[Danny Yoo]

On Wed, 18 Oct 2006, kumar s wrote:

> I am not sure if I got your opinion about the way I solved.

Hi Kumar,

I was replying to your question:

     http://mail.python.org/pipermail/tutor/2006-October/050064.html

in which you showed a program that had problems.  I'm assuming that you 
haven't figured out what's wrong with the code in your post.  I've 
presented an approach that shows how one might try to approach the problem 
without being brilliant.  Did you understand every step in the program in:

     http://mail.python.org/pipermail/tutor/2006-October/050121.html

or was there something there that you didn't understand?  If so, please 
ask for clarification, because everything there should have been 
approachable to you.



> do you mean that there is something wrong with the way i solved it.

Well... how do I put this gently?  Your program doesn't work, so I can't 
really give a positive or negative opinion about its effectiveness until 
it does.  I can sorta see what the code tries to do, but it's going about 
it in a complicated fashion.

I can give my opinion that:

###################
fx = a[0]
fy = a[1]
b = a[2:]
ai = iter(b)
last = ai.next()
for j in ai:
     print fy+1,last
     last = j
###################

seems to be juggling too many variables, too much state.  I think the 
conceptual bug here is that the code fumbles around, trying to do 
something toward a goal, but it loses intentionality.  I don't understand 
the intent of 'fx', since that variable is never used.  I don't understand 
the expression 'fy+1' in the print statement, since 'fy' never changes. 
There's just a lot of little small things there that seem irrelevant to 
the problem at hand.


I haven't tried to fix your program; rather, I've tried to show you an 
approach that can systematic attack these kinds of problems without much 
trouble.  The problem I see that you've been having is that you're getting 
stuck as soon as the complexity of a problem reaches a threshold.  I'm 
trying to show you techniques to compensate when problems get hard.  I 
could just as easily have just fixed the above code, but how would this 
really help you?  For you, it would be as if I just pulled that program 
out of thin air!



But just to satisfy you, such a program would have looked like this:

#################
x = a[0]
for y in a[1:]:
     print x, y
     x = y + 1
#################

The relationship between this and your original program is somewhat there.

But again, if I just showed this to you, how does that really help you to 
program better?  The program here is written to conserve memory, but at 
the expense of being harder to develop, understand, and explain.  And, if 
I were doing this as a computer scientist, I'd have to develop a loop 
invariant argument to feel satisfied that it was doing the right thing. 
In comparison, the inductive definition has the force of a proof behind 
it.


> I am not sure If I explained the problem correctly in terms of exons, 
> transcripts. If not I would be happy to send you a pdf file with a 
> figure.

That's ok.  I don't think it is necessary as long as you clearly explain 
the problem you're trying to solve.  You should be able to communicate 
what you're trying to do.

(And just as background: I've worked in the past as a bioinformatician at 
arabidopsis.org.  The terminology that you've been using has not been an 
issue.)