Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Programming > Java > Programming Contest: BoxifyMe

Reply
Thread Tools

Programming Contest: BoxifyMe

 
 
Skybuck Flying
Guest
Posts: n/a
 
      12-13-2010

"Brett Davis" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> In article <6b49b$4d03777e$54190f09$(E-Mail Removed)1.nb.ho me.nl>,
> "Skybuck Flying" <(E-Mail Removed)> wrote:
>
>> The contest has turned into an intrigueing development.
>>
>> A contester has presented himself and stepped forward into the dragon's
>> layer !
>> =D
>>
>> And what an unforgettable entry it was ! =D
>>
>> It was shocking at first and marvoulous at second glance.
>>
>> It has provided a new insight into a possible direction of solutions.

>
> This sort of technology is used for green screen masking of video.
> I worked on such code. You should know that long thin polys can
> have horrible performance on some graphics hardware.
> A bias toward squarish shapes will give better throughput despite
> sometimes more boxes to render.
>
> Entry 5 has a bug, in the center left is a large brown box, that
> box should have grown down one pixel into the pink box.


It's not really a bug, it's probably an algorithm inefficiency... from the
high number of boxes it can be seen that wasn't ofcourse efficient and I
tried to point that out... meanwhile that algorithm has been made more
efficient and entry6 proves that.

> A truly greedy algorithm should have given up some width to steal
> half of the blueGreen boxes space. Bias toward most pixels grabbed.
> A 2x10 is way better than a 1x12, etc.


Could be a heuristic.. but doubtfull... area doesn't really play a role that
much... sometimes it might be better to make a line to get rid of those
pesky little pixels/boxes which happen to fall on the same line but just one
pixel different... none-the-less area or ratio might still be a nice
heuristic.

> You can probably cut the item count of your render list by a third
> by including polys. An area greedy algorithm will leave you with
> lots of 3 and 6 pixel right triangles, and similar stuff that
> explodes into lots of little boxes.


This is a novel idea, for opengl and physics this could make sense.

It's a bit risky though... if the polygon/triangle because to large and it
has a flat angle like so:

.
.............

Then the diagonal that would result from that would describe too many
pixels...

The diagonal would look something like:

-------
-----.......

So it starts to describe non-existing pixels.


But for very small pixels like
.
...

it might work.

Perhaps also something like:
.
..
...
.....


As long as the pixel describe a perfect triangle/diagonal such an idea might
be possible.

So I will consider it when I start the physics testing !

Thanks for this idea, most appreciated !

Bye,
Skybuck.


 
Reply With Quote
 
 
 
 
Skybuck Flying
Guest
Posts: n/a
 
      12-13-2010

"Skybuck Flying" <(E-Mail Removed)> wrote in message
news:23c6c$4d064626$54190f09$(E-Mail Removed)1.nb .home.nl...
>
> "Brett Davis" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> In article <6b49b$4d03777e$54190f09$(E-Mail Removed)1.nb.ho me.nl>,
>> "Skybuck Flying" <(E-Mail Removed)> wrote:
>>
>>> The contest has turned into an intrigueing development.
>>>
>>> A contester has presented himself and stepped forward into the dragon's
>>> layer !
>>> =D
>>>
>>> And what an unforgettable entry it was ! =D
>>>
>>> It was shocking at first and marvoulous at second glance.
>>>
>>> It has provided a new insight into a possible direction of solutions.

>>
>> This sort of technology is used for green screen masking of video.
>> I worked on such code. You should know that long thin polys can
>> have horrible performance on some graphics hardware.
>> A bias toward squarish shapes will give better throughput despite
>> sometimes more boxes to render.
>>
>> Entry 5 has a bug, in the center left is a large brown box, that
>> box should have grown down one pixel into the pink box.

>
> It's not really a bug, it's probably an algorithm inefficiency... from the
> high number of boxes it can be seen that wasn't ofcourse efficient and I
> tried to point that out... meanwhile that algorithm has been made more
> efficient and entry6 proves that.


Actually that's not true... I just remembered entry6's algorithm was changed
somewhat... so it's not considered boxy...

So your points about the "boxy" being inefficient is true... and it wasn't
corrected yet.. since entry6 works a bit different...

Perhaps I will consult opengl newsgroup in future about your statement:

"You should know that long thin polys can have horrible performance on some
graphics hardware"

Perhaps you mean very strange cases where edges lie closely to each other ?

Or perhaps long in the sense of bad memory lookups or so... or perhaps just
inefficient hardware... hmmm..

Bye,
Skybuck.


 
Reply With Quote
 
 
 
 
Skybuck Flying
Guest
Posts: n/a
 
      12-13-2010
Maybe floating point issue's...

Perhaps hardware detects that 12 bit or 16 bit or 24 bit is not enough for
floating point numbers to describe the slopes... so then it might change to
64 bit or something which might be slower.

Bye,
Skybuck.


 
Reply With Quote
 
Skybuck Flying
Guest
Posts: n/a
 
      12-13-2010
A new contester has entered the contest !

And what an entry it is !

The new contester rises to the top spot ! And is now in the lead with only
217 boxes !

He goes by the name of "Rob Pratt"

No further details where provided... I would at least like to know what
programming language was used... this is after all a programming contest


The website had been updated to include Rob Pratt's First Entry.

Will he submit more ?!? Time will tell For now he can sit comfortably
on the throne ! LOL.

http://www.skybuck.org/BoxifyMe/

Bye,
Skybuck.



 
Reply With Quote
 
Skybuck Flying
Guest
Posts: n/a
 
      12-13-2010

"

"John Metcalf" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
Hi Skybuck,

I didn't see a deadline mentioned. Is there any chance you can make
it after the deadline for the International Corewar Programming
Contest?

http://www.corewar.info/tournament/icpc1

Thanks,

John
"

I think so.. the contest will probably be running for a while !

Maybe even forever... it's like a King of the Hill contest.

The best programmer/algoritm/entry sits on the throne !

Bye,
Skybuck.


On Dec 11, 5:00 am, "Skybuck Flying" <(E-Mail Removed)>
wrote:
> Skybuck's Boxify Me Programming Contest:
>
> http://www.skybuck.org/BoxifyMe/
>
> Your mission is to "Boxify" the following picture with a minimum ammount
> of
> boxes (=rectangles).



 
Reply With Quote
 
Skybuck Flying
Guest
Posts: n/a
 
      12-14-2010
An interesting development has taken place in the competition.

After carefull analysis by looking at the new visualizations I suspect Rob
Pratt of either:

1. Using the input from James Dow Allen to base his own entry/solution on.
or
2. Using the same kind of algorithm and/or software as James Dow Allen.
or
3. It's simply James Dow Allen pretending to be Rob Pratt.

It's very suspicious to say the least... you be the judge...

The visualizations are at the bottom, the CoordinateMap is a dead give away
me thinks ! Rob's has many similarities/patterns as that of James'es.

http://www.skybuck.org/BoxifyMe/

Bye,
Skybuck.



 
Reply With Quote
 
James Dow Allen
Guest
Posts: n/a
 
      12-14-2010
On Dec 14, 2:33*pm, "Skybuck Flying" <(E-Mail Removed)>
wrote:
> An interesting development has taken place in the competition.
>
> After carefull analysis by looking at the new visualizations I suspect Rob
> Pratt of either:
>
> 1. Using the input from James Dow Allen to base his own entry/solution on..
> or
> 2. Using the same kind of algorithm and/or software as James Dow Allen.
> or
> 3. It's simply James Dow Allen pretending to be Rob Pratt.


Well, I suppose it's my word against yours, but I'm confident that
neither of us is impersonating the other, and confident Rob
Pratt would agree!

That the solutions would be quite similar surprises me not in
the least: mine was developed in a very straightforward and obvious
way. I *am* surprised that Mr. Pratt has improved on my score
so dramatically and am anxious to see his solution, but ...

I cannot access your BoxifyMe page today. (Does anyone else have
a similar problem?) Skybuck, Please e-mail me Mr. Pratt's
solution if you have time.

Best wishes,
Rob Pratt^H^H^H^H^H^H^H^H^HJames Dow Allen
 
Reply With Quote
 
Skybuck Flying
Guest
Posts: n/a
 
      12-14-2010

"James Dow Allen" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
On Dec 14, 2:33 pm, "Skybuck Flying" <(E-Mail Removed)>
wrote:
> An interesting development has taken place in the competition.
>
> After carefull analysis by looking at the new visualizations I suspect Rob
> Pratt of either:
>
> 1. Using the input from James Dow Allen to base his own entry/solution on.
> or
> 2. Using the same kind of algorithm and/or software as James Dow Allen.
> or
> 3. It's simply James Dow Allen pretending to be Rob Pratt.


"
Well, I suppose it's my word against yours, but I'm confident that
neither of us is impersonating the other, and confident Rob
Pratt would agree!

That the solutions would be quite similar surprises me not in
the least: mine was developed in a very straightforward and obvious
way. I *am* surprised that Mr. Pratt has improved on my score
so dramatically and am anxious to see his solution, but ...

I cannot access your BoxifyMe page today. (Does anyone else have
a similar problem?) Skybuck, Please e-mail me Mr. Pratt's
solution if you have time.

Best wishes,
Rob Pratt^H^H^H^H^H^H^H^H^HJames Dow Allen
"

I'd much rather have you access my site properly.

I did introduce some "style sheets" maybe that's screwing up your browser ?

What operating system and browser are you using ?

And how much GHZ and Memory do you have ?

Bye,
Skybuck.


 
Reply With Quote
 
Lew
Guest
Posts: n/a
 
      12-14-2010
Skybuck Flying wrote:
> The best programmer/algoritm/entry sits on the throne !


Not coincidentally, that's where the best programmers come up with the best
algorithms.

--
Lew
Ceci n'est pas une pipe.
 
Reply With Quote
 
James Dow Allen
Guest
Posts: n/a
 
      12-14-2010
On Dec 14, 6:45*pm, "Skybuck Flying" <(E-Mail Removed)>
wrote:
> "James Dow Allen" <(E-Mail Removed)> wrote in messagenews:(E-Mail Removed)...
> I cannot access your BoxifyMe page today. *(Does anyone else have
> a similar problem?) *Skybuck, Please e-mail me Mr. Pratt's
> solution if you have time.


It seems no one else has reported the problem here.
But would thread participants please confirm if they *can* access
BoxifyMe?

> What operating system and browser are you using ?


Windows XP. Both IE and Firefox fail to load, just
"waiting" for a long time.
http://www.skybuck.org/BoxifyMe/Boxi...llenEntry1.txt
also fails to load, though http://www.skybuck.org/ itself
loads promptly.

> And how much GHZ and Memory do you have ?


Uhhh ... enough so that if that's the problem
I don't *want* to load your page!

Best ever,
James
 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Programming Contest: BoxifyMe Skybuck Flying C Programming 33 12-15-2010 04:18 PM
C (functional programming) VS C++ (object oriented programming) Joe Mayo C Programming 168 10-22-2007 01:00 AM
Can Your Programming Language Do This? Joel on functional programming and briefly on anonymous functions! Casey Hawthorne Python 4 08-04-2006 05:23 AM
Wireless PEAP/MSCHAPV2 client programming question Jim Howard Wireless Networking 6 07-02-2005 11:53 AM
systems programming versus application programming Matt Java 35 07-22-2004 08:10 AM



Advertisments