progress

Hey, does anyone mind sharing their progress?

How many words (maximum) have you been able to display from the given list? And how long is your script execution time? Mine just zips by in less then a second, and I have yet to add in the code to find the execution time. I'm on an amd 1800+, 512mb ddr266, win2k pro, apache1.3.26+php4.2.1.

I'm trying to sort it in a special way so that the words with the most possible connections will be added to my matrix first...

I still have to experiment with different sorting algorithms since in my program choosing the ideal sorting program will display a LOT more words than crappy ones (at one point i only got 6 words to display !@#!).

Oh and Matt, I assume you are the jugde? How do you plan to calculate the # of words displayed and the connections? Would it be easier if we made our programs calculate this? Of course there is always the "cheating" involved if we had our program print out # of words and connections :-). Oh yeah: and what computer will you be using to run our programs? Is it the crappy computer you mentioned in the contest rules (266mhz, 128 ram)?

-sridhar

Add that total number of squares used to you score. I am able to score avarage of 30 words, total score of about 700 with the given wordlist. Random and a 200MHz linux is used with total running time of 10 seconds.

Hi,

I got 38 words when I optimized my starting word position. I used a different sorting algorithm, which I believe works quite when.

I tried not sorting it and I get about 10 words less.

Here are my details:
# of words: 38 * 5 = 190
# of spaces taken: 197 * 1 = 197
# of connector: 45 * 3 = 135
My total score: 522
Total processing time: 1.63 seconds (on average).
I'm on a amd1800+, 512 ddr266, apache1.3.26+php4.2.1, windows 2k pro with latest service packs and all updates (which i doubt should really make a difference :-)

I've only worked on this script yesterday (you probably saw all those crazy posts!). I also did a leftovers collection to collect all the words not added and then rerun the algorithm to see if they'll fit anywhere... unfortunately all this did was take up twice the time (and more memory for the array).

numbah, how were you able to get 700 with just 30 words? Did you get all the really long words on there? Or you probably had tons of connectors :-)

I'm in the process of putting up my script. I used a firewall on my home system (where i'm running this), so I can't give you the URL. My web server is getting a new HD, so my web host has taken it down for about 2 days. I'll certainly put up my script then.

Also, have any of you tried increasing the grid size (just for the heck of it)? I got all 154 words to display with a grid size of 53x52 (min) :-)

good luck to all those who're entering,
-sridhar

Hehe, made a mistake:

change "I believe it works quite when" to "... quite well".

-sridhar

How can you optimize the first word position without knowing all the words that will be attached to it?
I iterate the "place a word if it isn't used yet" loop 4 times. If I try any more, it will go beyond the 30s time limit.
I gotta try that grid size increase and see how it works!
BTW.. I switched to Apache 2.0.3.9 from IIS and my proctime is now 3x better.. good advert for the Apache folks.

I was wondering if anyone would be kind enough to maybe explain the steps their algorythms are doing?
I'm not enterin the contest, I just think it's cool too learn more php by the codewalkers site in general.
Regards.

I'd ask that no one answer the previous question. The scripts submitted will all be available after the contest is over...

Devlsh: I optimize it according to how long the word is, the # of connectors it has, and depending on how long the second word is....

as matt suggested, i don't want to delve deeper into explaining my code until the contest is over :-)....

but i doubt I'll even come to close to winning seeing numbah has claimed a score of 700 while the most I get is 522 (average around 460).


OH: have any of your tried using different word lists? How about someone create a word list and we all use the same one to see how well we do. That way we can get a sense of who has the highest score and what we need to meet up to.

-sridhar

I would urge people not to pay attention to the scores posted here.

Last contest some people didn't enter because of scores posted here and come to find out, they would have done better than the scores posted here with the input I used to judge.

I am close to not allowing people to post scores here because of that reason. I don't mind general discussion or rule clarification, but when posts dicourage others from even submitting it is a bad thing.

Remember, it's not whether you win or lose, coding in PHP is just fun. And that is what this is all about!

I'm done (unless it turns out that the wordlist.txt file is in a different format than the one I downloaded from here). I've got my runtime down to 13 seconds on an 1800+ XP running WinXP with 768M DDR. Since I am running a pseudo-random algorithm, my words used, etc. change on each execution. I average 25 words and 30 connections. I'm not using Apache, I'm running PHP 4.2.1 under IIS. Which brings up a good question, would I be better off running Apache when testing execution times? If you want to see my output, check out this link:

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Link removed to avoid negtive effects on the contest.

Maybe Matt could supply us with one more wordlist.txt file for further testing?

Okay, sorry if I discouraged anyone...

yes, one more wordfile.txt will be really helpful.

Also Matt: you DO mean 30 seconds for each run right? Not all three runs combined.

-sridhar

Hey buddy!!
Believe me if you show the same performance you would win the contest...
Yesterday I started developing my algorithm and
(around 5 o'clock a.m) it was working (however
unoptimized)..Today I was able to get not more then
519 pts. I didn't apply any particular sorting yet..
nevertheless 700 seems to be quite a large score to
me.

Yours Kuulest.

Sorry 'bout that 700 mess.... I had a little misunderstanding with the problems scoring specifications and after fixing the code avarage score dropped to ~500... darn

numbah, sorry to use you as an example here

This is exactly why you shouldn't pay attention to scores posted here...