Re: [pygame] Re: GSoC Proposal: Basic gui system

["RB[0]" <roebros@xxxxxxxxx>]

I don't think incorporating an existing gui into pygame would work.
Whatever it is you decide to do - it will need to be close enough to pygame to be worth it (ie, integrated into event loop and such, to make it easy for new users), abstract enough to allow extendability into a wide range of needs, and powerful enough to look good and scale to full size applications.
From my experience, none of the current pgame guis come close to having all those attributes.
Either they are so abstract that the user has to do almost everything, they are too powerful to easily extend, or they are just funky in their usage.

But that is yet another decision.
How will events be handled? WIll they be returned like pygame events, or handled with callbacks? Or both and confuse people/be un-pythonic?
How will it be rendered? Will it simply be an object you send the screen to at the end of each loop and render then update the screen?
Will it use dirty rects to make the gui fast, or not, so you can use flip?
How many/what widgets do you want to include?
You pointed out how you would take the best 2 or 3 - well that is it, there is so much variance, and no gui I have seen that would be abstract enough to be useable for this claims to be everything you might/will need.
I think most likely you would have a ton of different people chiming in, and in the end you might get 2 or 3 highest rated, but none would get a majority and it would probably fall apart before it even started...

I agree with Sky, it is simple to write basic widgets, but once you get into more advanced stuff it takes time, and lots of it, there are just so many things to do. It isn't particulartly hard, but time consuming, and difficult to do "right". I think though updating the pygame wiki with some of the newer gui's, with a good comparison would probably work better - but if you do take one and make it work, more pwoer to you, and I am sure a lot of people will be grateful...

Good luck :)

On Tue, Mar 24, 2009 at 3:11 PM, Jug <jug@xxxxxxxxxxxxxx> wrote:

If you all think, a lot of people would use a built-in gui but it's
not worth the work, is there a chance to put one of the existing guis
into pygame? We could poll the users on the website and/or the mailing
list about the best pygame gui. Then maybe have a look on the best 2
or 3, discuss some pros and cons and include one into pygame? Because
as I said before, one big problem of all of the existing guis is that
they are at least one more dependency. In addition, new users don't
have to test all guis and find the best but can have a look at
pygame.gui and invest more time to understand it. I'm really tired
from reading up on guis (used 2 and began to read about Albow). A
pygame.gui module would get much more support (bug fixes, new ideas/
widgets, etc.) than now, too.

Regards,
Jug