Opinions from the developers point of view?


Bosbeetle

Terminally lost
Joined
Sep 7, 2008
Messages
4,287
Age
43
Location
The Netherlands
Website
Visit site
Hello everybody, we have been seeing a lot of reviews from new owners of the pandora, but I was wondering about people that are actually making software. I posted some guide questions for you devvers, Im just very curious about your opinions.


1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?
 
Last edited by a moderator:
1. What device did you mostly dev for before the pandora came on the scene?

GP2X/Wiz.

2. How does the pandora compare to this device devving wise?

The Pandora is the easiest handheld platform to develop for, no questions about it. Nothing compares.

3. What is the biggest downside to the pandora devving wise?

Nobody has one!

4. What is the biggest pro about pandora devving?

Its a real computer with its own compiler onboard.

5. What would you change or need, to make the pandora a better machine to dev for?

Customers. Integrated PND-App-Store download.

6. Anything else you like to share about this subject?

Its wonderful and I wish I could spend 20 hours a day writing code for the thing. Anyone want to hire a Pandora developer? :)
 
1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?
1. Playstation Portable - I still do. I'm one out of two Authors of the 6.35 PRO Custom Firmware.


2. It's way easier to debug on the Pandora, for PSP you need dirty tricks and a lot of selfcoded tracing tools to get the job done.


3. I've yet to find one.


4. See #2. So much easier if you can just use a backtrace to find your crash source, instead of having a blackscreen that just powers the device off.


5. Hm... a more lightweight QEMU based linux virtual machine available online + toolchain preinstalled on there, would save newcomers a lot of time, the current 64bit based Ubuntu VMs are slow as hell.


6. While getting into it was a bit rough for me (like with any new device you get honestly), once you are in its pretty smooth and I doubt there are any devices out that make developing easier than this one.
 
3. PYgame is pretty slow on it. If that was working at decent speeds(60fps for eg) then a lot of beginner coders like me could be knocking out games and Pnding them up. I've pretty much completed devving one platformer on pygame (Robo hell) but on the more complex levels I was getting 4FPS :(


So I've now gone from the beautiful world of python and having one PY script to the ugly world of c++ and SDL. Should make me a better progger though so not too fussed.
 
1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?

1: dingoo/x86 linux


2: glorius, much better in every way (though a bit of a different scope tbh)


3: lack of full opengl - just means we've all got to learn how gl and gles work, and the differences in order to get things working right


4: it's near enough a full desktop system, no silly restrictions/cut down version of linux


4: full gl support :)


6: we need more people willing(and capable) to test programs properly, as most devs don't get much time, and tend to miss bugs :p
 
6: we need more people willing(and capable) to test programs properly, as most devs don't get much time, and tend to miss bugs

Hmm I would like to say I can help but I am terribly busy and ive noticed that bug testing is more than just using it until it breaks :D
 
1. What device did you mostly dev for before the pandora came on the scene?

Windows PC.

2. How does the pandora compare to this device devving wise?

I don't develop on the pandora. As for how it compares for writing stuff for it... well, the win32 API is better than linux, is all. I have to resort to SDL, and that's just not nice.

3. What is the biggest downside to the pandora devving wise?

Sub-optimal SDL for the most part. I can't decide if it's SDL that's slowing my app down, or the CPU. I have a 366MHz x86 laptop which runs my app faster than the Pandora does.

4. What is the biggest pro about pandora devving?

One hardware set to target. If something runs well on mine, it will run well on others.

5. What would you change or need, to make the pandora a better machine to dev for?

Better SDL, if I'm going to be forced to use it. I don't know enough about linux yet, but I just can't seem to find an equivalent to the Windows DIB. Under windows, my app updates every pixel of the display at 50fps without a hitch. On the Pandora/SDL I can see it doing it, and framerates are on the floor.


When the windows version is feature-complete, I'll look into optimising the Pandora version.


D.
 
I don't develop on the pandora. As for how it compares for writing stuff for i

That's what I ment however for people who do develop on the pandora it also could be interpreted the other way :D


Btw zxdunny did you try the newer version of SDL that notaz 'made'? http://www.gp32x.de/board/index.php?/topic/57506-improved-sdl/ (if you are the dunny on gp32x you definetely have sorry for this comment)
 
Last edited by a moderator:
Interesting thread. I'll add my answers after I can put down my Pandora long enough to actually set up a build environment. (And don't suggest on-device developing. That's just painful.)

Better SDL, if I'm going to be forced to use it. I don't know enough about linux yet, but I just can't seem to find an equivalent to the Windows DIB. Under windows, my app updates every pixel of the display at 50fps without a hitch. On the Pandora/SDL I can see it doing it, and framerates are on the floor.
That doesn't sound right. Are you rendering the entire screen for every pixel update?
 
Last edited by a moderator:
Hello everybody, we have been seeing a lot of reviews from new owners of the pandora, but I was wondering about people that are actually making software. I posted some guide questions for you devvers, Im just very curious about your opinions.


1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?

1. GP2X/Wiz/Caanoo


2. Its like a PC so its much easier to port and debug applications. Having full X11 and GLES opens up many more apps. There are lot less specifics and quirks to work around.


3. Developing GLES apps can be a pain, the driver is not forgiving any misuse results in a full lockup.


4. Onboard debugging, pc like environment so pc applications just work.


5. More RAM (have to use swap instead), better GLES debugging (using PC gles libs can help)


6. Having the pandora has been a good learning tool, its helped me get the foot in the door in learning opengl and opengles. It has so many options available to development. Ive been able to work on things I never thought I would have or been capable of. There nothing more rewarding than thinking I wish someone could that and then it turns out your the one thats ends up doing it.
 
1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?

Let's see...


1. I bounce about devices a lot.. I have done some stuff for most things though, some even got released and can still be found :D


I do wish I did more on the GP32 and got something finished enough to release though.. I liked the GP32..


2. As the Pandora is essentially a mini-computer, it wipes the floor with most other handhelds and consoles just by being able to have GDB et al actually on the device to figure out what's gone wrong. Granted some of the tools that have been developed on other machines are very nice, nothing really beats the "I've crashed, this is why" on the actual device running it, should you have forgotten to link up your debugger at the time.


3. I'm going to echo mostly what's already been said for this - lack of feedback as there's not so many Pandoras out there that seem to come to the forums for some reason and test stuff. Though I do also think that perhaps we should have a stringent "Official" dev environment as although it's pretty awesome that you can set it up any which way you want, I bet it's confusing as sin for people new to the platform! And I'm well aware I'm one that contributes to the confusion!


4. Kind of already answered this, but I'll repeat anyway - being able to install dev tools on the device to figure out why it's crashed, and it being open enough to create your toolchain any way you like.


5. Other than a "this is the Official supported dev environment" on something for new developers, I don't really see much else.. most things are just down to lack of developers tackling it at the moment, so things like SDL can be a bit unoptimised still.. but will get better as more Pandoras get into developer hands, or developers find a fountain of free time!


6. It's actually quite exciting to be in the dev community just now, as more adventurous ports and things are coming out. More GLES stuff is appearing, for example. Fun times ahead! Here's hoping I can help contribute something to it all this time :)
 
I don't develop on the pandora. As for how it compares for writing stuff for i

That's what I ment however for people who do develop on the pandora it also could be interpreted the other way :D


Btw zxdunny did you try the newer version of SDL that notaz 'made'? http://www.gp32x.de/board/index.php?/topic/57506-improved-sdl/ (if you are the dunny on gp32x you definetely have sorry for this comment)

Yes, that's me :) And yes, I've tried it. It's a tad faster, but not much.

Interesting thread. I'll add my answers after I can put down my Pandora long enough to actually set up a build environment. (And don't suggest on-device developing. That's just painful.)

Better SDL, if I'm going to be forced to use it. I don't know enough about linux yet, but I just can't seem to find an equivalent to the Windows DIB. Under windows, my app updates every pixel of the display at 50fps without a hitch. On the Pandora/SDL I can see it doing it, and framerates are on the floor.
That doesn't sound right. Are you rendering the entire screen for every pixel update?

I am at present - it's not an issue on the x86 linux/windows machines that people have been testing on, but on the Pandora it's definitely a problem. I've begun rewriting the display routines to only updated the screen with changed regions. If I were just rendering one surface to the display it would be much easier, but there's quite a lot of processing to do on multiple surfaces every frame. I don't want to have to start using OpenGL as my software renderer is more than capable on PCs (yes, even my 366MHz Vaio laptop manages 50fps under a full load - not so the Pandora).


As I said, I shall continue to optimise my code when I'm done with the PC version. The other problem is that I'm coding in Delphi, and the ARM compiler output is possibly not as well optimised.


D.
 
Last edited by a moderator:
Hello everybody, we have been seeing a lot of reviews from new owners of the pandora, but I was wondering about people that are actually making software. I posted some guide questions for you devvers, Im just very curious about your opinions.


1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?

1. GP2X


2. I still cross compile, so it's the same in that respect, but at least we now have WiFi access to the target device! :)


3. The very low userbase due to delays - this is quite bad when trying to get feedback from your apps. The good news is finally this is improving!


4. It is a full linux computer with very nice controls - it opens up a lot of possibilities and I already mentioned the WiFi.


5. An Officially supported toolchain... there's about 3-4 kicking around now.


6. Simply put... I have been waiting for something like this for ages... well and then I had to wait some more because of delays, but it's very simple to dev for! I also agree with Pickle that GLES should be more fault tolerant, but those problems lie with the closed source drivers...
 
Reading these posts, I plan on actually submitting bug reports to devs rather than just shrugging them off. :p
 
Cool thread. I can say as far as an end user goes, I do try to report bugs when I find them. I've mostly found them in the firmware (and reported what I've found). I don't have all that much time to mess with a lot of different programs on my Pandora. That will probably change once I've finished replaying EarthBound (been waiting for 2 years to play it again, wanted to replay it on my Pandora next).


Since the SNES emulator is the program I use most at the moment, that would be the one that I'd report bugs for, except for the fact that the bugs are basically known and accounted for. Skeezix told me that he was happy to port the SNES9x4X emulator, but had no real interest in optimizing it, so I don't see a lot of point in reporting random slowdowns, etc.


I do really appreciate all of the hard work by the devs! You guys rock for supporting a device that has less than 3k in existence.
 
Hello everybody, we have been seeing a lot of reviews from new owners of the pandora, but I was wondering about people that are actually making software. I posted some guide questions for you devvers, Im just very curious about your opinions.


1. What device did you mostly dev for before the pandora came on the scene?


2. How does the pandora compare to this device devving wise?


3. What is the biggest downside to the pandora devving wise?


4. What is the biggest pro about pandora devving?


5. What would you change or need, to make the pandora a better machine to dev for?


6. Anything else you like to share about this subject?

1. x86 crossplatform linux/windows


2. xfce is a full linux desktop, so it's essentially the same


3. its to hard to setup a development environment


4. because answer 2, its very easy to be source compatible with x86 software


5. an official toolchain, a set of "preferred" libraries, example crossplatform code
 
Back
Top