@OlafS25

Why are we still discussing the bounty?

The donators that has posted so far seems more than happy to see AmiDark collect NOW (or rather yesterday, so we didnt have to hear about it anymore).

The whole point is accepting review for what it is. Extract the useful info you get from it. If something is unclear OR if the reviewer has misunderstood a function, then communicate that clearly and objectivitly, instead of going into defense mode.

If someone tells me Ive executed a procedure incorrectly, I would accept that input, but at the same time I ask how and why. That way I can evaluate if the input is valid OR learn the reasoning why the procedure is as it is. At the end of such a review I AND the reviewer would have learned something.
This applies both to development and any other line of work.

I realise this is just a hobby, but take a look at http://ada.untergrund.net/ or http://csdb.dk/forums/?roomid=11
Plenty of back and forth, where people appriciate feedback regarding their coding projects.

Last edited by Overflow on 14-Jan-2015 at 12:40 PM.
Last edited by Overflow on 14-Jan-2015 at 12:35 PM.

Im bit afraid of saying my opinion at this point anymore as it seems everyones disagreeing here.

But first of all, although I do agree that only those who put money in bounty (I didnt) should decide wether the bounty is done or not, I do think there is still value even with those opinion who didnt put money to bounty, especially now that it seems that to some parts it is a question of how some wording should have been understood.

It seems to me that at least AmiDarks intention was to word in such a way to describe it as it now is. And I also would say that it is pretty much in state I was expecting it to be. However, the code quality is a matter I cant really say much, since I am not good enough at coding to comment about it, hence especially Daytona is bringing important input on this side of the bounty that might be difficult for average user to understand (at least it is for me).

And what comes to code quality, while I might not agree with Daytona about teh quality standard that the code should be (but once again, i dont really understand how important most of the things are that he says), i do agree that it is a valid point that also the quality of the code should be determined to some extent.

To compare this to a house. It isnt enough that house looks good, it also needs to be able to stand still when normal wind is blowing. For even if point is to reinforce the house more, I still wouldnt buy the one that is almost done where it is easier and cheaper to rather build a new house than fix the old one.

And in this we come to daytona again. I go buy a house. I can look its looks, size, build year etc. but i wouldnt really know if roof is going to collapse tomorrow or not, wether it can stand the wind or not. This is when i call Daytona to come and check on my behalf and describe me the house how he sees it from the more technical point of view that i dont understand about.

The code quality from Daytonas input it seems to me that for almost all parts it have been acceptable (to me - had i put money in bounty) and in condition i would have expected it to be, but there was one function that seemed bit worrying. It was the part where i suppose you were supposed to use some sort of Pis value, and it had instead been using 6+1 or something like that (approximately same value as 2*Pi).

For this sounded like something that might have not got fixed unless it had been poitned out, and something that could have caused a serious problem for the programmers, depending what it actually was affecting (once again, i dont really know what i am even talking about right now), but if it had something to do with drawing 3D stuff, then it sounds to me like a serious problem actually. For the idea is that you write code once, and then you can run it on different platforms, and they all look the same, regardless of platform (except for some minor differences can occur still anyway), but this sounded like something that if this is responsible of drawing walls in quake, then there could be quite big spaces between wall corners, and you could even see the surprise hiding monster behind that corner on the other side of the wall thanks to this wrong value. I have no idea if it was so serious a bug in code or not, but thats how it sounded to me.

Almost all of the rest seemed more like stuff that would have got fixed in the end product anyway, as far as the end product does come out some day, or things that were more of a issue of performance/usability rather than something actually going wrong.


This bountys wording of description arguing also brings up the question of should there perhaps be thought up some standards on how to describe the incomplete states of bounty to make sure that everyone understands them right. For it doesnt seem to me like AmiDark hadnt filled the bounty as he meant to describe it, it is only a question of did he describe hes intentions in such way that those who put money on it also understood it right.

By the way, does bounty2people have some sort of voting system there?

I mean, it is good to have open free talk like this where outsiders can give input to, so that people who put their money in, can determine their opinion about it better, but wouldnt it be best that in the end (rather sooner than later), bounty2people would put up a voting up that is open only to those who put some money in the bounty.

Then you simply calculate the amount of yes or nos, perhaps based upon the amount they put it, and if for example 75 percent say yet, it is fulfilled, then the money woudl be released to Amidark?

All in all, i Hope Amidark gets hes money soon, and engine gets forwad.

ZZD10H : I've read your statements. I'm writing a new thread/subject for the debug with your request too ;) Sorry for zapping to answer you ..
Will check for MAKEFILE naming convention.

Bugala : Do you buy an unfinished house ? ..
Concerning the pi, it's simply a function to draw 2D circles or 2D elipses on screen ... Nothing to do with 3D as 3D is handled by OpenGL.
It's simply an algorithm I've found on the net, that I have adapted and it worked for me .. It is not the "most optimized" but it still work... Of course it can be optimised and it will be ... But I think we lost ourselves in guess-work.

Overflow : I really appreciate feedbacks .. But as I am someone really "hypersensitive", I am also really sensible to the way peoples are speaking to me and then, how they do their report... As a simple comparizon. I Prefer how BSZili do his reports on aros-exec.org . he is a guy that know what "respect" and "interactions with others" mean.

I've added a new Thread/Subject especially for the debug that can be done :
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=39807&forum=15&0

Thank you to all people that find bugs in the engine to report them to this thread especially added for that.

Regards,
AmiDARK

Last edited by AmiDARK on 14-Jan-2015 at 01:33 PM.
Last edited by AmiDARK on 14-Jan-2015 at 01:30 PM.

omg.. as if it wasnt expected. this engine has to my knowledge not seen any application save one demo, except apparently examples written by the author himself. i highly doubt that it will change in the future, especially that os4 users, this engine is apparently focusing on, seem to usually lack infrastructure to run these kind of software at reasonable efficiency, that means hardware 3d drivers. it was a caprice of a bounty, and the donors accustomed to throw money in bounties for similar circumstances are fine with it. so the only thing that consequently must be going to happen is that the bounty will be paid, the code released and the whole thing forgotten. lets get over with this.

@AmiDARK
Quote:

Concerning the pi, it's simply a function to draw 2D circles or 2D elipses on screen ... Nothing to do with 3D as 3D is handled by OpenGL. It's simply an algorithm I've found on the net, that I have adapted and it worked for me .. It is not the "most optimized" but it still work... Of course it can be optimised and it will be ...

Not that it really matters since this is probably the least important / most uncritical bug I found so far, I'd like to say it once again: the ellipse-function is not just sub-optimal, it is also broken.

Quote:

But I think we lost ourselves in guess-work.

Well, probably you got lost in guessing, me not. Since it is so easy for this ellipsis function to prove it graphically, here is what you get. Essentially the shape overpaints itself (of course, since it loops > than 2*pi). But due to the 0.1 stepping it also overpaints itself visibly

Large ellipsis drawn using PasteB2DShapeElipse:


Zoomed into that part marked red:


This here is reality.
Although it can be easily proven AmiDark continues to tell us that there is no bug.
And it is the very same with every other bug I reported and he keeps denying. Only that the others are severe.

Last edited by Daytona675x on 14-Jan-2015 at 03:14 PM.

_________________
AmigaOS 4.1 FE (sam460ex Radeon 9200 / RadeonHD), MorphOS 3.8 (PowerMac G4 733MHz Radeon 9000), AROS (x86), A1200 (060 80MHz Indivision MK2), A500, A600, CDTV
Wings Remastered Development Diary

@Overflow

Quote:

Why is it even a discussion?

Discussing things is a very good and mature thing to do - wouldn't it be nice if the Timberwolf bounty was discussed like this before the money was handed over? Let's not blame 'discussion' (or Daytona) for the shortcomings of this bounty project.

I do agree that the bounty requirements have been fullfilled - because, as usual, they're way to vague. But that's not AmiDARK's fault, so he should get his money now.

Daytona seems to have some very valid complaints though - so perhaps this might serve as a good lesson for future bounties: maybe things should be discussed before people start throwing money at something?

If somebody wants to have thousands of dollars for the sources of an unfinished project, or a 15 year old file manager - make a code review before you start a bounty.

If it's an unfinished game development library, have the author compile some example code and run some benchmarks. And, most importantly, take a step back and try to think of people (as in: actual, real-life programmers) who could actually have a use for this. If you can't come up with a single name - walk away.

Has a single coder voiced interest in this library? if not, why is there a bounty for it?

I'm not criticising AmiDARK - he made a proposal and people accepted it. I'm just wondering where the need to throw money at things comes from. And please, don't ever stop to discuss. This might not be the most friendly thread recently, but it's definitely one of the most sane ones.

@Daytona675x

What is your problem?

AmiDark set up a thread for debugging.

Do you understand what people tried to tell you about what and how you are writing?
(@Moderators: Please help...)

I am really happy about every bug wich will be reported. But it is hard to understand what your goal is.

Just my opinion: I don't like the way you are going on and on with your posts in this thread.


@AmiDark

I can really understand you if you lose your nerves with Daytona675x.

_________________
Sometimes we give people a lot of credit just because they’re writing nice sentences even if it isn’t adding up to much.

@Daytona675x :
So, with all your arguments. It's maybe time to "contribute" with your true knowledge ? Do you have a better algorithm for elipse/circles drawing ? It will be really appreciated.
I have already asked you that in the subject concerning debug :

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=39807&forum=15

Thank you.
Regards,
AmiDARK

@Birbo

+1

_________________
Tygre
Scientific Progress Goes Boing!

@zzd10h

Quote:

zzd10h wrote: Compilation OK for few samples under AmigaOS4. Unfortunately, without Warp3D, I'm unable to start them. I was aware of this fact. If somebody could test this executable under AOS4 http://zzd10h.amiga-ng.org/AmiDark/BoingBall.lha

it crashed here with a guru.
when launched from shell and RAM, a window opens and then a crash (ISI) that can not get over.

@OlafS25

Quote:

At it current state it is propably alpha software and should only be used for testing and not for development. But I think the donators donated mostly for charity and because it is a "amiga based" software, not because they expected a ready or usable engine.

They donated to open-source what he has already written. The bounty was pretty clear that this is what was going to be delivered.

Of course there's a lot to be done afterward before we have a finished and fully usable game engine.

Hans

_________________
http://hdrlab.org.nz/ - Amiga OS 4 projects, programming articles and more. Home of the RadeonHD driver for Amiga OS 4.x project.
https://keasigmadelta.com/ - More of my work.

@lionstorm

Ooops...
Thank you for the test.

@AmiDark
?

_________________
http://apps.amistore.net/zTools

@lionstorm :
Which Amiga platform SAM ? X1000 ?
Which video card ?
Can you send me more informations concerning the crash ?

@ZZD10h : That latest *stable* release was the 0.8.
I have made some changes and improvements ... So, it's maybe some of these changes that cause problems. I'll compile a special version of the library with checking inside to see which part of the library crash (if it's during the init process).

Regards,
AmiDARK

Last edited by AmiDARK on 14-Jan-2015 at 08:02 PM.

@AmiDARK

Quote:

AmiDARK wrote: @Jupp3 OpenGL 2 (that is our Amiga standard) does not allow something better for sprites. I may have "never added them" but I decided to add them only for compatibility with DBPro/DarkGDK. More of this, I've looked everywhere on the net to find how we can do perfect pixel system under OpenGL and even all the informations given on the net. It's impossible to reach on Amiga OpenGL (or I didn't found it yet).

If you can draw the background once, you (usually) can draw it all again, which should be faster (and more "clean"), unless you draw extremely complex graphics.

The best solution might be to render to FBO the background, and use that to redraw background before redrawing sprites. Of course that needs to be supported. Otherwise... Well... Anything that avoids glReadPixels() etc - it's good only for random screenshots.

@Jupp3
I know all these, but due to the "way it work" under DarkBASIC Professional, I currently cannot use something else than "glReadPixels".

In fact, imagine a coder that use the AmiDARK Engine.
1. He display many "DEPasteImage" tiles that are pasted on screen to create the game background.
2. On this background, he draw several sprites (imagine ship, monsters, projectiles, etc.)...

I must be able to :
1. Save the tiles pasted
2. Add the sprites over the display.
3. Display everything on screen.
4. Restore the tiles previously saved.

As on each frame user can add what he thing, it can become really complex to handle everything on a list. So the principle to use glReadPixels to save the tiles on screen is the easier way to do this. As I know that in our current OpenGL 2 it is sooooo Slow .. I decided to put sprites "only for compatibility" and recommend to NOT use sprites if possible. And, there is an extra function, not available in DarkBASIC Professional, to disable/enable the internal "backdrop saving" routine.

EDIT : Source code updated : I removed all unused files
+ I opened a "bug report" forum in the repository.


Regards,
AmiDARK

Last edited by AmiDARK on 14-Jan-2015 at 09:00 PM.

@AmiDARK

"That latest *stable* release was the 0.8."

Euh...
It seems that only the 0.9 is available on Sourceforge, right ?

_________________
http://apps.amistore.net/zTools

I donated to this bounty and am happy that it has been fulfilled.

I understood that the project was unfinished and was released "as is".

I have zero aptitude for programming and therefore never expected to be able to use this software, I donated because I like to support future Amiga developments where I can.

@Zzd10h :
You're right pointing that ... in fact, 0.0r1 was the last releas under os4depot and was stable with the default samples (I did forgot that).

I must check :
1. In my dev diary what I changed from this release and that may cause the crash...
2. Your makefile for compiling the AmiDARK Engine under AmigaOS4 (can you send me your makefile per e-mail, you know it ?, or copy it here)

Last edited by AmiDARK on 14-Jan-2015 at 09:39 PM.

@Caratacas

5 / 0

@AmiDark

Finally, I checked and as it miss the drawer AmiDARKEngine:SDK/ in your package, it is the 0.9 version that I had still january 2014 that I used.
My libAmiDark.a is from 10 january 2014 (367052 bytes)

I didn't modified the Makefile, I just renamed it to Makefile.

Concerning, Lionstorm crash, I retrieved a post on AmigaNG and it worked in early 2014 (when I used a R9250) because I had to add a 640x480 screenmode.

Maybe is it the problem on XEG4 of LionStorm ?


Just for fun, remember one of my comments in january 2014
"-Chemin Codebench en dur dans les .cbp de certains projets (raster par exemple)"


http://www.amiga-ng.org/article.php?sid=893&mode=&order=0

_________________
http://apps.amistore.net/zTools

@zzd10h
Ok.

Thank you for your *research* :)
Concerning the Path in *hard* in CodeBench,
I'll have to create a setup for the AmiDARK Engine that can create an assign for the engine demonstration and then I can use the assign in the CodeBench project.
But I have never created "install" under Amiga ... So I don't know how to create one ... I may check ... But with only an Amiga1200 available .. It can be "outdated" for what I need. not ?