It is currently Thu Jul 19, 2018 6:02 am Advanced search

code stats, length, brevity

Random stuff about the contest, posts that don't fit in the other forums.

code stats, length, brevity

Postby jcothran » Tue Oct 12, 2010 2:30 pm

It might be interesting to have sub-contests limiting the amount of code involved for a contest. Not only what are the best solutions, but which do it in the least amount of code(that is 'readable')? Could see where that might get into unreadable code as well just for line compaction - maybe just simple stats like the total code amount in the submitted package.
jcothran
Captain
 
Posts: 24
Joined: Mon Oct 11, 2010 6:03 am

Re: code stats, length, brevity

Postby Terr » Tue Oct 12, 2010 7:34 pm

The problem is that'll be heavily dependent on the language used as well as what libraries are available for it.
Terr
Brigadier-General
 
Posts: 155
Joined: Mon Sep 13, 2010 9:19 pm

Re: code stats, length, brevity

Postby jcothran » Tue Oct 12, 2010 7:52 pm

break out by language and whatever is in the submitted zipped package counts towards size including support libraries - the interesting thing to me would be if I see a competitor that is very good with a very small amount of code comparatively - maybe they are using a better recursive approach or other techniques that code size might be of interest versus a larger code base
jcothran
Captain
 
Posts: 24
Joined: Mon Oct 11, 2010 6:03 am

Re: code stats, length, brevity

Postby krokkrok » Wed Oct 13, 2010 1:00 am

Why? What virtue are you highlighting? We could also have a sub-contest for who can use the most 'K's, but that would serve no purpose.

I think this is one of those ideas that have long plagued software: that number of lines of code is a meaningful metric.

In the past, large corporations such as IBM use the KLOC measurement: programmers were actually rated by how many kilo-lines-of-code they wrote: the more your wrote, the more you got paid! It took a surprisingly long time time the industry to notice that all this does was encourage programmers to waste time adding useless code.

The idea of rewarding programmers for the shortest source code is just as ridiculous. Programmers will waste time picking one letter variables, creating arcane macros and reducing overall maintainability for no objective purpose.

I suppose it is not ironic at all that this AI contest will be won by the programmers who can pick the best objective metrics for success of their bot, instead of the ones following arbitrary aesthetic ideals.
krokkrok
Major
 
Posts: 38
Joined: Mon Sep 13, 2010 2:48 am

Re: code stats, length, brevity

Postby jcothran » Wed Oct 13, 2010 3:46 am

The robocode contest has sub-contests with code size limits
http://en.wikipedia.org/wiki/Robocode#C ... restraints

>I think this is one of those ideas that have long plagued software: that number of lines of code is a meaningful metric.

This is where you are off on your own thread. I agree with you that in day-to-day work code shouldn't be short just for the sake of being short and shouldn't be cryptic. Think of it more like boxing contests or weight classes in various sports and the boundaries people try to expand within the artificial limitations of sport to begin with.
jcothran
Captain
 
Posts: 24
Joined: Mon Oct 11, 2010 6:03 am


Return to Misc

Who is online

Users browsing this forum: No registered users and 2 guests

cron