an item suggestion for the next poll

nitrofurano on 2009/09/01 16:45

i were seeing the polls made here in pyweek website, like http://pyweek.org/p/10/ , http://pyweek.org/p/8/ , http://pyweek.org/p/4/ and http://pyweek.org/p/2/ , with questions like first operating system, second operating system, ammount of ram and ghz of our computer, which version of python and which libraries we use, etc. - another important question is which operating system we don't use.

For example, i don't use ms-Windows, and there are people (pygame.org is a good example) sending .exe files, expecting everybody can run them - and some can't run on Wine - when i think all people here in PyWeek competition has no access (and not planning to have) ms-Windows seems to be around 30% and 60%, but would be nice knowing better how many we are...

The only last poll i used about PyWeek (somewhere at Google) announced here in the PyWeek message board, says each primary operating system were around 30%, but lots of people does dual-boot (i do with MacOS-X, for example)

(log in to comment)

Comments

adam on 2009/09/01 17:22:

Check again and you'll see that the poll for this contest actually provides this information - the first question is "what is your primary operating system?" for which respondents can only provide one answer, but the second question asks you to list all operating systems you have access to - so the results show that 72% of users have access to Windows, 63% to Linux and 35% to OS X.

nitrofurano on 2009/09/01 17:28:

it's the last poll the Google one? i lost its url... :( - anyway, 28% not using ms-Windows is very significant (more than 1/4 of PyWeek participants...)

Cosmologicon on 2009/09/01 17:48:

One of the great things about PyWeek is that you can't distribute it only as an exe. You have to distribute the source. So all the entries should run on all the major platforms. Ideally, it shouldn't matter what OS anyone is running, although in practice it helps to know, eg, for reproducing bugs.

I'm appreciating it more now that I'm trying to judge Ludum Dare entries. It seems like 2/3 of those are Windows only.

Tee on 2009/09/01 17:56:

In Pyweek, sending an .exe file is recommended, but sending the source code is required. So, putting aside portability problems, I don't see how not having Windows is much of an issue. As long as you have Python and whatever libraries are necessary for the game, you should be ok for most of the games.

Tee on 2009/09/01 17:58:

Argh. I started writing that last post before Cosmologicon posted his. I should remember to check the thread before I hit save... (I guess that's something preview was useful for, too.)

nitrofurano on 2009/09/01 18:04:

This is really great! i'm loving even more Python, PyWeek and Pygame because of this! :)

And this about 2/3 of the Ludum Dare entries being were a 'surprise', since i knew a recent statistic study saying, for example, 70% of the game players do dual-boot with Linux, and use msWindows only for playing msWindows-only games.

Sad is most of the indie games i knew from http://indiegames.com/blog/ are also msWindows-only , and the Kenta Cho's games, which i knew all from Ubuntu repository (mostly C++ with OpenGL), now started coding games only using XNA... :/

tartley on 2009/09/01 20:24:

I thought Kenta Cho used D, not C++? I might be wrong.

nitrofurano on 2009/09/01 21:01:

@tartley - no idea, i imagined were C++, since they were ported so 'easily' to Linux and MacOS-X - but i think his OpenGL games are all really compiled (like .cpp code compiled on GCC, not merelly binded) - i must see the sources with better attention, but if there's no difference between C++ and D, i really can tell you - do someone can confirm that better?

gcewing on 2009/09/02 12:12:

While it may be theoretically possible to play any game on any system, in practice the use of certain libraries can make it very difficult for some platforms -- e.g. Panda3d on MacOSX.

tartley on 2009/09/02 15:07:

The results of the poll are useful, but I don't think they reveal precisely what we need to know.

For example, if 72% of users have access to Windows, and 35% to OS X, then if these two populations do not overlap, then just producing Windows and Mac binaries would be sufficient to make sure that all PyWeek participants will be able to run and score your game.

Clearly, however, this isn't correct, because these two populations do overlap, but how much? How many people would be unable to run your game out of the box if you only produced Windows and Mac binaries? Somewhere between 0% and 28%, but beyond that there is no way to know.

Better would be if the results were processed, to show "If I support [list of platforms], what proportion of people would be able to run my game?" If the raw data (or a suitable slice through it) can be made public, then I wouldn't mind processing it to show this.

  Jonathan