Development Bounties

[craigix]

it doesn't seem right target fortheir donation (ie, porter is doing 1/100th the work of original team usually) and against bounties usually for easy stuff. Ie: dropping asm into snes emu isn't really hard and should not be bounties. Better menus for emus? Lame use of bounty.

No one has though have they, so if it's easy maybe you should do it and claim the money?

The Pandora isn't going the way of the Wiz where everything was just left half finished and abandoned, if we have to have bounties to achieve success then we will. That was the idea of the dev fund and now we're putting it in to use.

Once the main things are done they are done and we can all relax and enjoy the Pandora. I don't want to be waiting years for that situation and basing it on hope.

We need action now, we can't stand still.

When it will happen gratis if we wait. that's the point of the community, and opensource, i think.

Alas that way of thinking does not always work.

The push on software needs to happen now, and of course the main devs will be offered money to continue what they have been doing - but sometimes no one wants to do a certain task - those places are where we need to push things and get things done with the dev fund.

I don't want to be sitting here in a year still saying 'it would be really cool if someone did that for free!' and just waiting possibly forever for it to happen.

This push is needed early on if we want to show people how awesome the Pandora is, waiting for months or a year and people will have already made up their mind.

 

[john4p]

- Autozoom (via OpenGL ES hardware-scaling) to use the full height of the Pandora's screen for UAE4All
- working write-to-floppy functionality in UAE4All (currently not possible to save in Amiga games)
- savestate-support for UAE4All

- add Cyclone to PUAE
- Configurable controls for PUAE
- Autozoom and fullscreen mode for PUAE

- CPS-3 emulator (for Street Fighter III: Second Impact and Third Strike)

 

[SomeGuy99]

So long as the existing firmware devs get some money like Craig said, I don't see the problem.

 

[JayFoxRox]

IMHO there should be a fixed value of money (a portion of the available money) which is split between projects which can be voted up. Then there could be a rule that a project without major upgrades could not be voted 2 months in a row and if it is a port the original devs should also get a portion of the money (after the porter has shown how much work the port was).

Useful software is going to be written anyway and with such a system devs would be motivated to target a broader audience.


One could then create an extra bounty list for stuff which needs to be done ASAP - which would be for bugs in the OS, for example.



//Edit: Plus a lot of stuff on here is not even possible atm or can't be done by the community unless lots of work and time is invested. IMO devs should do whatever they can first (rather than wasting their time on something almost impossible), then get paid if they did a good job..

 

[Pleng]

This push is needed early on if we want to show people how awesome the Pandora is, waiting for months or a year and people will have already made up their mind.

In which case send some cash to Tinnus and get him to re-write that accelerated 3D plugin for the PSX emulator that he lost!

Though, of course, that might upset Zodd. Better give him some cash too.

 

[Oscaruzzo]

The push on software needs to happen now [CUT]

This push is needed early on if we want to show people how awesome the Pandora is, waiting for months or a year and people will have already made up their mind.

I've said this elsewhere, but the best "push" you can give to software development is: PUBLISH A DECENT AND WORKING SDK! possibly for both linux and windows. (sorry for the caps, but maybe this time it will be heard).

The wiki is in an awful state. Some info must be gathered from the forums, some other info is on some blog arund (most notably from DJWillis' blog , some is actually on the wiki, but it is outdated and/or wrong.

If you really want to boost development, you really really need some good reference documentation, some good libraries, an SDK (cross compiler, QEmu setup, etc). You could put a bounty on this, if no one wants to do it for free, but I really think this is a major show-stopper for the Pandora.

 

[kuru]

BTW somezthing like ReMoko/BlueMaemo would make me very happy.

Video of BlueMaemo on the n900:

Spoiler

[youtube]http://www.youtube.com/watch?v=gOTpmRKHTLY&feature=related[/youtube]

Edit: Tag-Fu

 

[Natasha]

Is there any chance to be paid for development in a queue bump? I think my coding skills might be good enough to be helpful, but I'd rather develop on the real thing. I don't exactly know my place in the queue, but I'm afraid without higher intervention it will probably be too late to push anymore.

My guess is 'no', because there is no reason to trust me, but still…

 

[tsh]

Maybe the bounties can be seen as a way of prioritising the development of trivial stuff (for trivial $$$) and also to encourage the less glamorous and more fundamental features (but maybe these become more of a community effort, and it gets messy).
I sort of agree with Craig that the software needs a lot of work in order for Pandora to reach it's potential, but I'd be quite wary of offering too much $$$ for any one task since most people who contribute will do so for their own satisfaction (and contributions to the wiki are just as valid as coding if anyone is feeling left out!)

 

[WizardStan]

I've said this elsewhere, but the best "push" you can give to software development is: PUBLISH A DECENT AND WORKING SDK! possibly for both linux and windows. (sorry for the caps, but maybe this time it will be heard).

There are some very major problems with this.
1) SDK for what language? If you are just looking for a language that works, Python is already working quite well, from what I've seen. Your SDK already exists, just write! It does Java (I'm pretty sure), so that's covered. It's only a matter of time until someone figures out .Net support. Or you want a C++ SDK. Well in that case ...
2) ... what libraries? It's trivial to set up a cross compiler for Linux (and slightly less trivial but still easy if you follow the instructions in Windows) for building simple apps. But text adventures get boring after a while. So what gets added? SDL? OpenGLES? LUA? GTK? You start adding every library you think a person might want and it's no longer simple. You need to keep the
Seriously, this is a full blown Linux computer with an seemingly unlimited number of options; an SDK does nothing but restrict you! If you want to build applications, get a cross compiler, and do with it whatever you want. If you have trouble getting some library you want to use, just ask and (if it's possible) I'm sure someone will gladly build it for you.

 

[Oscaruzzo]

I've said this elsewhere, but the best "push" you can give to software development is: PUBLISH A DECENT AND WORKING SDK! possibly for both linux and windows. (sorry for the caps, but maybe this time it will be heard).

There are some very major problems with this.
1) SDK for what language? If you are just looking for a language that works, Python is already working quite well, from what I've seen. Your SDK already exists, just write! It does Java (I'm pretty sure), so that's covered. It's only a matter of time until someone figures out .Net support. Or you want a C++ SDK. Well in that case ...

Sorry, I assumed C/C++.

2) ... what libraries? It's trivial to set up a cross compiler for Linux (and slightly less trivial but still easy if you follow the instructions in Windows) for building simple apps. But text adventures get boring after a while. So what gets added? SDL? OpenGLES? LUA? GTK? You start adding every library you think a person might want and it's no longer simple. You need to keep the
Seriously, this is a full blown Linux computer with an seemingly unlimited number of options; an SDK does nothing but restrict you! If you want to build applications, get a cross compiler, and do with it whatever you want. If you have trouble getting some library you want to use, just ask and (if it's possible) I'm sure someone will gladly build it for you.

I think there should be a minimal set of libraries, and a decent concept of what a "Pandora OS" is. Extra libraries should be managed via some kind of repository, just like any other distro. Maemo already does that. Meego is on its way. Why can't we? Of course because that's a LOT of work. But if a different solution is suggested, then it should be clearly documented nonetheless.

Setting up a cross compiler is NOT a trivial task, or at least it was not when the gp2x came out (I still have nightmares involving libc and I'm NOT going trough it again...)

 

[skeezix]

Idetailed setting up a cross compiler in 5 mins on my blog . Linked from pandora wiki .. But maybe hard to find in here dunno

djwillis is working on an sdk and several unofficial ones exist

jeffphone

 

[WizardStan]

Setting up a cross compiler is NOT a trivial task

Download, install, TADA!

 

[LeChuck87]

mhm... maybe

-OpenGL Driver (non ES)

 

[XavierBlak]

Saw this a few days ago, seemed relevant. Using money as a motivator isn't always the best idea. Granted I see craig is trying to guide the development of the pandora through the most direct method available to him. Just seems like it needs to be handled carefully.

 

[hdonk]

Setting up a cross compiler is NOT a trivial task

Download, install, TADA!

It's not just setting up the cross compiler. Getting the right version of the libraries that the Pandora is using, confirming compatability of the OpenGL ES code you write, all are example problems with doing cross development. Not only that, but you won't know how well the code you've generated runs on the OMAP!
Personally I'm waiting for my Pandora before finishing off the Amstrad CPC emulator port. As a result part of me says it's a bit early for bounties - until there's enough units in developer hands, we won't know how fast things would happen without them.

 

[GizmoTheGreen]

now that i think about it
SGStair got wifi working on the ds thanks to a bountie as motivation (note he did reverse engineer it)

akkit.org if i remmber

 

[WizardStan]

It's not just setting up the cross compiler. Getting the right version of the libraries that the Pandora is using, confirming compatability of the OpenGL ES code you write, all are example problems with doing cross development. Not only that, but you won't know how well the code you've generated runs on the OMAP!

Getting the right libraries is not hard, but it isn't trivial, you are correct. I never said they were.
For the bold, now that you've got a cross compiler, you can simply build them yourself. For the less knowledgeable that just want ones that work, ask politely and someone will build for you. And no doubt as more people ask for common and less common libraries, they'll be uploaded to some sort of common repository where you can just download the libraries of your choice to be installed in exactly the same way a regular library would be installed into your editor of choice.
So to recap my points: SDK forces you into an editor with a finite and fixed set of libraries; Generic cross compiler, for not much more effort (and certainly not necessarily more effort than a developer writing an application for x86 would have), lets you use (almost) any editor you want, and any library compatible with ARM you want.
The rest of your points are an argument of on board vs cross compiler which is an entirely different argument that I have no interest in going over again.

 

[joseluisjazz]

Saw this a few days ago, seemed relevant.

Thank you so much for the link. Amazing stuff!

 

[craigix]

Saw this a few days ago, seemed relevant. Using money as a motivator isn't always the best idea. Granted I see craig is trying to guide the development of the pandora through the most direct method available to him. Just seems like it needs to be handled carefully.

I know. I understand the careful balance, that's why I'd like the rewards to be for smaller tasks. Those ones which people say are 'easy' but they just get overlooked for weeks. Little fixes, little solutions.

Those little things can make a big difference.

We know from the wifi fiasco and Christian being paid money and DOING NOTHING for it was a disaster. (Yeah Christian, I'm sure you will threaten the mods again to have your name removed, why didn't you put that sort of effort in to working for us - or just give us our money back?).

But if we say, hey, here is $250 to (for example) tidy up the Amiga emulator - to be paid when it's done, we might get a nice gradual series of small fixes done on various bits of software - there is a level we can reach quickly where the Pandora does most things most people want very very well, I want us to push towards that asap using whatever processes/incentives we can.