[clug] Re: ACM programming competition
andrew.over at cs.anu.edu.au
Wed Sep 8 06:47:52 GMT 2004
On Wed, Sep 08, 2004 at 02:55:56PM +1000, James Ring wrote:
> On Wed, 8 Sep 2004 01:36 pm, Michael James Stevens wrote:
> > Yes. I have Warren. (u3950692). This gives us a 3 person team. I'm busy
> > with COMP3500 right now, so can you email Eric to get our team in
> All done!
Free (unsolicited! woo!) advice (for the actual competition):
- Avoid computational geometry questions unless you are a comp geom guru
, or you've solved all the others. Yes I know it looks easy, but
there are 3 different corner cases you haven't thought of (excluding the
15 you have).
- There will be several graph theory problems. Know about breadth-first vs
depth-first traversal, along with things like transitive closure
algorithms (floyd-warshall? it's been a while).
- Dynamic programming pops up occasionally and yields nice short solutions
to otherwise intractable problems.
- Language familiarity is more important than support for nice data
structures (you can do very nasty things in C if you have to).
- With the exception of the first problem, don't waste your time typing
until you've coded a solution on paper. You'd be amazed how much less
buggy things are when you've forced yourself to code it up on paper (at
- Style? What's style? (An ideal solution will qualify as an IOCCC entry
when you're done).
- Caffeine is your friend.
And now back to procrastination.
 You're probably not a comp geom guru. 
 I say this from bitter experience. There's almost ALWAYS a corner case
you overlook, and when your only feedback is "wrong answer" it's not so
much fun. Convex hull problems are not too bad, but the rest are
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 185 bytes
Desc: not available
Url : http://lists.samba.org/archive/linux/attachments/20040908/60083e08/attachment.bin
More information about the linux