|
CodeCraft |
|
This is a long thread. Click here to view the threaded list. |
|
Jeffrey Lee |
Message #26278, posted by Phlamethrower at 15:50, 5/12/2002 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
/me misses it |
|
[ Log in to reply ] |
|
Phil Mellor |
Message #26279, posted by monkeyson2 at 15:53, 5/12/2002, in reply to message #26278 |
Please don't let them make me be a monkey butler
Posts: 12380
|
Nothing to stop TIB setting a competition up.
What categories would we have? |
|
[ Log in to reply ] |
|
John Hoare |
Message #26280, posted by moss at 15:54, 5/12/2002, in reply to message #26279 |
Posts: 9348
|
What happened to CodeCraft? |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26283, posted by Phlamethrower at 15:58, 5/12/2002, in reply to message #26280 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
Dunno; it just didn't happen this year
I've gotta run to a lecture at the mo though |
|
[ Log in to reply ] |
|
John Hoare |
Message #26284, posted by moss at 15:59, 5/12/2002, in reply to message #26283 |
Posts: 9348
|
I've gotta run to a lecture at the mo though
I've got to run home and rewrite the Asylum site for tomorrow |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26312, posted by Phlamethrower at 17:17, 5/12/2002, in reply to message #26284 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
I've got to run home and rewrite the Asylum site for tomorrow Yes you have, you fat twat.
Hmm, codecraftiness...
Well the obvious categories would be minimal demo coding, where you program what you want inside a certain size limit and the 'best' one wins, as decided by the voters.
Or you could have a more goal based one, where there might not necessarily be voting - e.g. a challenge to write the fastest possible sprite plotter for a certain colour depth, using a given speed test program, etc. Of course with that idea the entries would have to be tested on a range of hardware (e.g. kinetic & plain SA) in order to find out which one is the fastest overall.
Could also be interesting to have a competition in some esoteric language, such as Befunge |
|
[ Log in to reply ] |
|
John Hoare |
Message #26313, posted by moss at 17:19, 5/12/2002, in reply to message #26312 |
Posts: 9348
|
Yes you have, you fat twat. I'm dust. Could also be interesting to have a competition in some esoteric language, such as Befunge So you can win one |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26316, posted by Phlamethrower at 17:26, 5/12/2002, in reply to message #26313 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
Could also be interesting to have a competition in some esoteric language, such as Befunge So you can win one Tsk! I got joint 1st place in one of the Code Craft 3 categories, and almost 2nd place in another! Now I know more about the kind of thing people are after I could probably do better, as well.
But yes, I probably would win a Befunge one, simply because I'm so obsessed with dodgy languages |
|
[ Log in to reply ] |
|
John Hoare |
Message #26318, posted by moss at 17:27, 5/12/2002, in reply to message #26316 |
Posts: 9348
|
Tsk! I got joint 1st place in one of the Code Craft 3 categories, and almost 2nd place in another! Oooh! Well done! |
|
[ Log in to reply ] |
|
Phil Mellor |
Message #26323, posted by monkeyson2 at 17:33, 5/12/2002, in reply to message #26312 |
Please don't let them make me be a monkey butler
Posts: 12380
|
Well the obvious categories would be minimal demo coding, where you program what you want inside a certain size limit and the 'best' one wins, as decided by the voters. Yup.
Or you could have a more goal based one, where there might not necessarily be voting - e.g. a challenge to write the fastest possible sprite plotter for a certain colour depth, using a given speed test program, etc. Of course with that idea the entries would have to be tested on a range of hardware (e.g. kinetic & plain SA) in order to find out which one is the fastest overall. I'm not sure about a "doing it fastest" challenge - if it's too difficult it might not attract too many coders. How about something like the *INFO themed challenges - Cloud, Circle, Spooky, Maze solving, etc? Bonus points can be given to entertaining/novel approaches, even if they're not so technically competent. And it would give people a chance against of beating the likes of you!
Could also be interesting to have a competition in some esoteric language, such as Befunge One line BASIC please! |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26325, posted by Phlamethrower at 17:36, 5/12/2002, in reply to message #26323 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
I'm not sure about a "doing it fastest" challenge - if it's too difficult it might not attract too many coders. How about something like the *INFO themed challenges - Cloud, Circle, Spooky, Maze solving, etc? Bonus points can be given to entertaining/novel approaches, even if they're not so technically competent. And it would give people a chance against of beating the likes of you! *INFOCraft?
[Edit - Of course it doesn't have to be a competition where other people enter - I'm just up for a challenge ]
One line BASIC please! 1-line Befunge is harder
[Edited by Phlamethrower at 17:39, 5/12/2002] |
|
[ Log in to reply ] |
|
Michael Drake |
Message #26342, posted by tlsa at 17:48, 5/12/2002, in reply to message #26318 |
Posts: 1097
|
/me liked HappyRGB from CodeCraft 2. |
|
[ Log in to reply ] |
|
John Hoare |
Message #26348, posted by moss at 17:56, 5/12/2002, in reply to message #26342 |
Posts: 9348
|
I remember one from (was it CC3?) that made your backdrop a moving starfield. I liked that |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26349, posted by Phlamethrower at 17:58, 5/12/2002, in reply to message #26342 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
Haven't seen that one
In fact, I haven't seen any entries from CC 2. Must have a look at them some time.
I've got a 'fun' idea:
An ARM source code compressor/optimiser
As long as it assumes certain things (Or is given hints about them), e.g. what bit is a jump table/relies on non-label addresses, etc. it could work quite well.
[edit] Although stuff produced by C compilers might give it a bit of a headache - e.g. all the stuff C compilers enjoy sticking on the stack, which could feasibly be held in registers instead. In that case it'd need special flags to say that the data doesn't have to be stored where the source says so.
[Edited by Phlamethrower at 18:05, 5/12/2002] |
|
[ Log in to reply ] |
|
richard cheng |
Message #26351, posted by richcheng at 18:12, 5/12/2002, in reply to message #26348 |
Posts: 655
|
I remember one from (was it CC3?) that made your backdrop a moving starfield. I liked that It was indeed CC3, and it was awesome.
and you could do:
*warp 0
to make the stars stop, or
*warp 9
to make them go really fast. |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26360, posted by Phlamethrower at 20:33, 5/12/2002, in reply to message #26349 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
I've got a 'fun' idea:
An ARM source code compressor/optimiser Since I've now got my pthreads stuff out the way, I think I'll have a shot at this |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26396, posted by Phlamethrower at 13:33, 6/12/2002, in reply to message #26360 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
I've written a program to optimise register usage, but it doesn't work yet
Tried running it on a short 5 instruction test sequence, and it managed to use almost 400MB of RAM before crashing
And that's without any recursion yet
/me suspects something is wrong... |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26404, posted by Phlamethrower at 14:04, 6/12/2002, in reply to message #26396 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
w00t! It works
I'm not even sure if I need the recursion
At the moment it doesn't understand the instructions it's optimising though; all it does is reshuffle the registers so the lowest number are used.
Now I need to work out a more complex way of specifying the data flow - e.g. register groups which need to stay in order (e.g. STM/LDM), and a way of specifying branches and stuff...
And make it understand the instructions so it can remove redundant ones. And write the parser so it accepts assembler files instead of a table full of funny letters
e.g.
O # MOV R0,#1 .O # MOV R1,#2 iIO # ADD R2,R0,R1 I .n # MOV R3,R0 n I| # MOV R0,R2 U U # End
is optimised to
n # MOV R2,#1 n | # MOV R0,#2 H H # ADD R0,R0,R2 | H # MOV R2,R2 H | # MOV R0,R0 U U # End
(Where I've changed the instructions on the right manually to show what's going on). Each column on the left represents a register, and the table of them makes 'usage chains' linking instructions together that use the same registers. Then it just looks for chains which can be joined together (e.g. attaching the end of one chain to the start of another) in order to reduce the number of registers
[Edited by Phlamethrower at 14:05, 6/12/2002] |
|
[ Log in to reply ] |
|
Richard Goodwin |
Message #26413, posted by rich at 15:01, 6/12/2002, in reply to message #26279 |
Dictator for life
Posts: 6828
|
Nothing to stop TIB setting a competition up.
What categories would we have? I'm not averse to this idea BTW, if someone more knowledgable of such matters want to start the ball rolling. Start submitting your list of categories...
(BTW, I think a 1K BASIC category would allow for more interesting stuff than a 1 liner, while still having the same feeling of extreme brevity)
Oh, and who would be the judges? ________ Cheers, Rich.
|
|
[ Log in to reply ] |
|
Andrew Poole |
Message #26414, posted by andypoole at 15:13, 6/12/2002, in reply to message #26413 |
Posts: 5558
|
Nothing to stop TIB setting a competition up.
What categories would we have? I'm not averse to this idea BTW, if someone more knowledgable of such matters want to start the ball rolling. Start submitting your list of categories...
(BTW, I think a 1K BASIC category would allow for more interesting stuff than a 1 liner, while still having the same feeling of extreme brevity)
Oh, and who would be the judges? Public vote? ________ |
|
[ Log in to reply ] |
|
Richard Goodwin |
Message #26417, posted by rich at 15:33, 6/12/2002, in reply to message #26414 |
Dictator for life
Posts: 6828
|
Public vote? I think it'd be better to restrict it to a set number of people, but they don't all have to be ninja programmers. ________ Cheers, Rich.
|
|
[ Log in to reply ] |
|
Michael Drake |
Message #26423, posted by tlsa at 16:03, 6/12/2002, in reply to message #26417 |
Posts: 1097
|
I think it'd be better to restrict it to a set number of people, but they don't all have to be ninja programmers. Maybe there could be two votes - one for the proper "judges" and one that anyone could vote in. I think a pubic vote would get more people interested and the previous CodeCraft competitions had pubice votes, iirc.
Cheers,
Mike |
|
[ Log in to reply ] |
|
Richard Goodwin |
Message #26430, posted by rich at 16:50, 6/12/2002, in reply to message #26423 |
Dictator for life
Posts: 6828
|
Maybe there could be two votes - one for the proper "judges" and one that anyone could vote in. I think a pubic vote would get more people interested and the previous CodeCraft competitions had pubice votes, iirc. Fair point.
Right, that's voting sorted, what next? ________ Cheers, Rich.
|
|
[ Log in to reply ] |
|
Andrew Poole |
Message #26433, posted by andypoole at 17:03, 6/12/2002, in reply to message #26430 |
Posts: 5558
|
Maybe there could be two votes - one for the proper "judges" and one that anyone could vote in. I think a pubic vote would get more people interested and the previous CodeCraft competitions had pubice votes, iirc. Fair point.
Right, that's voting sorted, what next? prizes? ________ |
|
[ Log in to reply ] |
|
James Shaw |
Message #26434, posted by Hertzsprung at 17:03, 6/12/2002, in reply to message #26433 |
Ghost-like
Posts: 1746
|
Fair point.
Right, that's voting sorted, what next? prizes? mean points |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26437, posted by Phlamethrower at 17:16, 6/12/2002, in reply to message #26413 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
Start submitting your list of categories... *wonders whether he's inadvertently entered himself into something again*
Um, well first of all they should all be 32bit compatible
As you say there's the 1K category. I think that restricting the categories to specific languages could be a bit restrictive; e.g. if we have different categories for each language then we might only get one or two entries in each category. We could avoid that by having a multi-language category, or just restricting people to one language.
The only categories I can think of are the kind of things from previous competitions - 1k demo, 2k tool, 4k demo, etc. Plus a category for doing amazing things in a suitably inappropriate/convoluted langauge
And like I said we could have goal based ones. Anyone up for a competition to program a Java VM/web browser/new version of Printers? |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26439, posted by Phlamethrower at 17:22, 6/12/2002, in reply to message #26433 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
prizes? Iyonix!!!
Devise a competition that'll really help RISC OS in some area, then get Castle to back us
Then find people insane enough to take up the competition ... |
|
[ Log in to reply ] |
|
Tony Haines |
Message #26445, posted by Loris at 17:59, 6/12/2002, in reply to message #26439 |
Ha ha, me mine, mwahahahaha
Posts: 1025
|
As far as I know, the reason there have been no codecrafts recently has been the lack of organisers. If you do want to do it, I suggest contacting Alain Brobecker and/or Eli-Jean Leyssens in the first instance.
I have previously entered (in the first two competitions IIRC) and it was great fun. I've tried to help out a bit once or twice - although I probably ended up hindering things as much as helping.
I certainly would like to see more codecrafts occuring, but I have a few recomendations:
Declare the deadline *far* in advance, then stick to it. Coding takes a long time.
Declare all the classes at the start, and then don't change the specifications at all (except to clarify - which should only be done judiciously).
Make the program sizes as small as possible. 1k is probably large enough for all classes. If you have larger sizes, people pack more in to one entry so you get less entries. I think this effect is more than linear - because instead of producing several small demos, more time is invested in just the one demo to make sure it has the best chance of winning.
Chocolate is at least a part of the traditional prize.
*Advertise* in the press, and do so *early*. Try to get every acorn mag and website to run a small article and plug your website, if not a series.
I suggest the following classes: 1k demo 1k utility 1k game
(I'd particularly like to see the latter)
If you want to grow the group of applicants, you might also try a fourth prize: 'best first entry' for people who have never entered a codecraft program before. |
|
[ Log in to reply ] |
|
Tony Haines |
Message #26448, posted by Loris at 18:04, 6/12/2002, in reply to message #26445 |
Ha ha, me mine, mwahahahaha
Posts: 1025
|
Oh, and phlamethrower, the obvious entry for you is a 1k befunge interpreter. Go on try it I dare you. |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #26449, posted by Phlamethrower at 18:06, 6/12/2002, in reply to message #26445 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
Sounds good to me |
|
[ Log in to reply ] |
|
Pages (2): 1
> >|
|