29-Mar-2024 08:21 GMT.
UNDER CONSTRUCTION
Anonymous, there are 42 items in your selection
[News] ScummVM InterviewANN.lu
Posted on 07-Apr-2002 22:34 GMT by Christian Kemp42 comments
View flat
View list
Andreas Magerl wrote: Today, we release on amigafuture.de a German interview with Rüdiger Hanke (ScummVM developer)
ScummVM Interview : Comment 1 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 08:37 GMT
To summarise: current work is a straight recompile using SDL but this didnt work too well on AmigaOS ( 68k ) for no apparent reason ( bugs in SDL support on AmigaOS most likely ) so he did the work on MorphOS which magically worked. Wants to make it work without SDL ( not sure why bothering would make it hard himself when main maintainers make upstream updates ). Doesnt work too well at the moment, colour corruption et al. He has no plans to port for AmigaOS4 and proceeds to bash AmigaOS 4 whilst lauding MorphOS. Tries to build credible "I was an AmigaOS4 BAF but Ive seen the light" type image in the article. Fell asleep before the end.
Gave me a chance to practice my German tho.
ScummVM Interview : Comment 2 of 42ANN.lu
Posted by Anonymous on 08-Apr-2002 09:39 GMT
In reply to Comment 1 (DaveW):
ok wanted to start the war ....
where is the cross-compiler allowing me to build progs for os4 ?
You cant blame people to try to bash a non existing thing
even if Rüdiger Hanke wanted to do he os4 version of scumvm, he will only be able to create a 68k version or a wup version, 68k and wup will be emulated and not native.
how do you want to compile something for a thing which dont exist (even virtually : there's no cross-compiler)
ScummVM Interview : Comment 3 of 42ANN.lu
Posted by Budda on 08-Apr-2002 09:57 GMT
In reply to Comment 2 (Anonymous):
WarpUP emulated? How can you emulate a PPC binary on a PPC CPU? Stupid.
ScummVM Interview : Comment 4 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 10:31 GMT
In reply to Comment 2 (Anonymous):
> Where is the cross-compiler allowing me to build progs for os4 ?
Ask Hyperion what is required beyond using WarpUP for now and recompiling later.
> You cant blame people to try to bash a non existing thing
Like production ready versions of MorphOS?
> even if Rüdiger Hanke wanted to do he os4 version of scumvm, he will only be able to create a 68k version or a wup version, 68k and wup will be emulated and not native.
68k will be executed in emulation. Wup will be executed natively.
> how do you want to compile something for a thing which dont exist (even virtually : there's no cross-compiler)
If it compile on SDL the work to make it run nicely on AmigaOS4 will be:
* Recompile SDL on AOS4 or for WarpUP depending on what you want to achieve.
* Recompile game
Run.
You see it isnt even a matter of flame wars. You didnt even mildly singe me. The reason my body temperature went up a fraction of a degree is because the air conditioning just shut off.
Any more anti AmigaOS4.0 bleating?
ScummVM Interview : Comment 5 of 42ANN.lu
Posted by David Scheibler on 08-Apr-2002 10:45 GMT
In reply to Comment 1 (DaveW):
I think Ruediger clearly said why he doesn't support AmigaOS at the moement:
Lack of information. He was under NDA but didn't get any more dev infos and
when Hyperion took over OS4 the situation didn't change. Exactly the same
statement which already came from Titan Computer.
On the other side he got the information and support he wanted from the MorphOS
team (he didn't even contact them, Ralph did contact him...).
I hope he'll release his new programs (and also develops his the killerapp he
spoke about) on M-Day.
Every developer has to decide what platform he'll support in the future. Some
chose AmigaOS some MorphOS. Ruediger did chose MorphOS (at the moment), so
where is the problem? If Hyperion wants OS4 to be a success than they have to
attract developers. Otherwise it'll be hard for them. Maybe they will release
the promised OS4 dev kit these days.
ScummVM Interview : Comment 6 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 10:53 GMT
In reply to Comment 5 (David Scheibler):
Maybe my translation wasnt good enough for me to get the right idea but the amount of code modifications required to get something recompiled for AOS4 once development tools become commonly available is understood to be near zero.
ScummVM Interview : Comment 7 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 11:10 GMT
In reply to Comment 4 (DaveW):
What's false in Rüdiger Hanke's sentences ?
ScummVM Interview : Comment 8 of 42ANN.lu
Posted by Jan-Erik Karlsson on 08-Apr-2002 11:13 GMT
In reply to Comment 7 (Nicolas Sallin):
that a straight use of SDL on the AmigaOS /WUP should not work is false
(look at WarpSCUMM, blows his SCUMMVM out of the water in all respects)
SDL has existed for WUP for quite some time (I made a version some time ago that
was for both 68K and PPC and worked flawlessly -none of the display thrashing that the morphos version has-)
ScummVM Interview : Comment 9 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 11:13 GMT
In reply to Comment 7 (Nicolas Sallin):
Who said anything was *false*?
Perhaps we should all write in esperanto.
ScummVM Interview : Comment 10 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 11:27 GMT
In reply to Comment 8 (Jan-Erik Karlsson):
Exactly. Thats why I think this was in part a political piece.
ScummVM Interview : Comment 11 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 11:47 GMT
In reply to Comment 8 (Jan-Erik Karlsson):
Writing a complex & good quality app using warpos and/or powerup isn't
as simple as a real amiga and/or morphos app.
That's the point.
WarpScumm still has many strange bugs, and Chaozer has a hard time
fixing them one by one.
ScummVM Interview : Comment 12 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 11:49 GMT
In reply to Comment 9 (DaveW):
Then if nothing is false, this is not any anti AmigaOS bleating.
ScummVM Interview : Comment 13 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 11:56 GMT
In reply to Comment 12 (Nicolas Sallin):
So what you are saying is that because you are standing on a flat table the Euro is a good currency.
ScummVM Interview : Comment 14 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 11:57 GMT
In reply to Comment 11 (Nicolas Sallin):
It is not hard to write a pup or Wos application.
ScummVM Interview : Comment 15 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 12:05 GMT
In reply to Comment 12 (Nicolas Sallin):
Tell you what post the "facts" from his interview here - YOUR translation not mine so theres no mistake - and also the "reasoning" and the "opinion" and classify them here as such and:
1. Ill demonstrate why logical correctness of factual statements in a dialogue is not sufficient to back up your assertion above.
2. Ill point out where he, in my translation, appears to be taking side swipes at AmigaOS in a very unconvincing way.
Also have you ever ported anything using SDL? Written anything for WarpUP? If so please tell me what was so difficult about it because Id love to know from the point of view about how to pitch it to beginners.
ScummVM Interview : Comment 16 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 12:13 GMT
In reply to Comment 15 (DaveW):
Writing a complex amigaos+wos/pup isn't as simple as
with amigaos/morphos.
Now, continue to understand what you want.
ScummVM Interview : Comment 17 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 12:19 GMT
In reply to Comment 16 (Nicolas Sallin):
WHY dont you find it as simple? What is so difficult????????????????
It is not enough to make an assertion without backing it up. It is the difference between something meaningful and something meaningless.
If you concur that it is subjective then I would agree - I personally found Wup and Pup a joy to use and no harder ( in fact less so ) than writing for traditional AmigaOS.
Im not taking the piss out of you I really am interested in what you find so difficult about it.
ScummVM Interview : Comment 18 of 42ANN.lu
Posted by David Gerber on 08-Apr-2002 13:09 GMT
In reply to Comment 14 (DaveW):
> It is not hard to write a pup or Wos application.
If you want to achieve a good performance, it is. You have to avoid as much cache flushes as possible and minimize OS calls. This require a careful design of your program. If you manage to do it, you get many times the performance over a 68060 program. If you fail, you only get something like twice the speed.
PowerUP and WarpUP might be fun, but it's passé (read: people have better things to do than worry about cache flushes and PPC <-> 68k messaging).
ScummVM Interview : Comment 19 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 13:17 GMT
In reply to Comment 18 (David Gerber):
> If you want to achieve a good performance, it is. You have to avoid as much cache flushes as possible and minimize OS calls.
Im sorry but I dont actually find that difficult. If we are talking about good game design ( for example ) then you naturally do that as part of the design/development process anyway.
As I noted on another thread when porting the lizard onto AmigaOS WuP/PuP the constant need to do interact with the OS and the fact that I could not see a good way around it made the port unsatisfying.
For games Pup and Wup are fine. For apps where theres lots of system calls and IO the 68k is a millstone around our necks. Doesnt mean its difficult to write for it just means that come AOS4 this performance problem will be significantly reduced if not a thing of the past.
Does *not* however make it difficult to write for.
Regards
Dave.
ScummVM Interview : Comment 20 of 42ANN.lu
Posted by Alkis Tsapanidis on 08-Apr-2002 15:15 GMT
In reply to Comment 14 (DaveW):
> It is not hard to write a pup or Wos application.
AHAHHAHA! That's new... Go ask Hyperion about that... It seems that you
don't know that unless you spend months getting rid of as many context
switches as you can, your software will be slower than on a 68k and will
load for ages... Trust me, I know this first hand. That's why the developers
got tired of PUP/WUP and want PPC native OSes (OS4/MOS).
Software doesn't magically work nicely on WUP by itself. You have to optimize
THE HELL out of it...
ScummVM Interview : Comment 21 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 15:24 GMT
In reply to Comment 20 (Alkis Tsapanidis):
I know quite well how to get great performance out of Wup and Pup as well as the main issues. I have yet to create or use software that is slower than the 68k version on PPC so I would dispute that too. However I put that down to enthusiastic exaggeration on your part.
ScummVM Interview : Comment 22 of 42ANN.lu
Posted by Alkis Tsapanidis on 08-Apr-2002 15:32 GMT
In reply to Comment 21 (DaveW):
Ok, try to make a photogenics like program to run ENTIRELLY on the PPC.
Run the main loop on the PPC, and call the GUI drawing functions with
the PPC. You'll what context switches are all about:) And cut the bullshit
man... It's very difficult to make COMMERCIAL grade software for WUP/PUP.
You need to have everything running perfectly. Nice load times, (VEEEERY
difficult), and nice speed (a bit easier). I'm not an outsider my friend.
If it was that easy then why would everybody moan that we need a PPC OS,
be it AmigaOS4 or MorphOS? Why do you think that Hyperion&bPlan are so interested in making a PPC AmigaOS(or compatible)? Are they stupid and have
money to waste?
ScummVM Interview : Comment 23 of 42ANN.lu
Posted by Alkis Tsapanidis on 08-Apr-2002 15:34 GMT
In reply to Comment 21 (DaveW):
BTW, what did you actually write? :-)
ScummVM Interview : Comment 24 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 15:54 GMT
In reply to Comment 22 (Alkis Tsapanidis):
What bull? Ive already said that system calls are the bane of a Wup and a Pup developers life and that games software is easier to write on Pup and Wup ( which is after all what the discussion is about ) vs application software. It still is not hard to write for Pup or Wup.
Im finding myself restating the same thing - time to agree to disagree. :-)
TTFN!
ScummVM Interview : Comment 25 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 16:01 GMT
In reply to Comment 22 (Alkis Tsapanidis):
> If it was that easy then why would everybody moan that we need a PPC OS,
be it AmigaOS4 or MorphOS?
Yeah thats right Im sure its all because Wos and Pup usage means you have to think when writing code ( just like you had to think about odd and even cycles back in the A500 days ). That must be it. Definately. If you say so it must be so.
We need a PPC OS because:
1. The 68k series has been discontinued for a LONG time ( ignore coldfire ).
2. The OS still running on 68k is a joke. Its slow, outdated and we need to move on.
Yeah of *course* its all Wup and Pups fault because they are so *hard* to write software for. I have *ALREADY* said that application and system software is slowed by the 68k side. I get the impression that you are not really arguing with me you are revisiting an old argument that you have had before with someone else - perhaps on comp.sys.amiga.misc?
> Why do you think that Hyperion&bPlan are so interested in making a PPC AmigaOS(or compatible)? Are they stupid and have money to waste?
Thats right thats exactly what I think. What a very silly thing to write.
Dave.
ScummVM Interview : Comment 26 of 42ANN.lu
Posted by Alkis Tsapanidis on 08-Apr-2002 16:03 GMT
In reply to Comment 24 (DaveW):
Noticed that the mouse is jumpy on all WarpOS versions of Scumm? Ask yourself
why:)
ScummVM Interview : Comment 27 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 16:06 GMT
In reply to Comment 26 (Alkis Tsapanidis):
Notice that the mouse is all jumpy using a rtg workbench on a decent gfx card when you are using Aweb and it keeps interrupting your mouse drags?
Ever ask yourself why even tho wb and aweb are 68k?
Stop telling me things I already know please :-)
Dave.
ScummVM Interview : Comment 28 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 16:06 GMT
In reply to Comment 26 (Alkis Tsapanidis):
Because Dave has not written them.
ScummVM Interview : Comment 29 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 16:12 GMT
In reply to Comment 28 (Nicolas Sallin):
Wow sniping now. Look be sensible about this please:
Here are my views based on my own experience, if you disagree then by all means carry on but at least back off the provocative statements.
1. Games software on Wup and Pup is not hard to write and to get decent levels of performance out of it.
2. Application software is a b*stard because of system calls through 68k.
3. If and when we get a fully PPC based OS any performance issues you or others might be having will be reduced significantly.
Im leaving it at that, if you want to make followup remarks, flame, abuse or otherwise be obnoxious its your affair.
Dave.
ScummVM Interview : Comment 30 of 42ANN.lu
Posted by Nicolas Sallin on 08-Apr-2002 16:19 GMT
In reply to Comment 29 (DaveW):
We agree on these 3 points. :)
You think writing a complex amigaos+pup/wos application is as simple
as writing a normal amigaos/morphos one.
But you think "application software is a b*stard because of system
calls through 68k".
All that collide somewhere :-)
ScummVM Interview : Comment 31 of 42ANN.lu
Posted by DaveW on 08-Apr-2002 16:29 GMT
It is *not* hard to write for Pup/Wup but the 68k legacy code slows everything down - thats a given. It is more difficult to end a Pup/Wup application that uses system calls or 68k library calls that doesnt underperform compared with what the PPC is capable of but there really is nothing you can do about this until you put a fully PPC OS on there. The APIs for Pup&Wup isnt difficult, the semantics of the issues are not difficult to understand.
Moving on because I can see we are just disagreeing on the interpretation of "difficult".
Still doesnt answer what the problem he had with getting ready to support AOS4 was. What does he fear about it? Oh no he cant go and recompile or execute for AmigaOS 4 until its either in beta or development phase. Schlock horror. Unless my translation was incorrect.
Certainly no reason to throw his toys out of the pram and say he will not support AmigaOS 4.0.
My suspicion is that there is an underlying fear of AmigaOS4 becoming production ready soon and lack of committed support for MorphOS versions of applications - fear of being marginalised - so some of those that have got on the MorphOS bandwagon are saying as loudly as possible that they wont support AmigaOS4 and doing a little bit of political mud slinging on the side about Hyperion when they get the chance.
Understandable but very very sad and will just cause a backlash.
ScummVM Interview : Comment 32 of 42ANN.lu
Posted by Alkis Tsapanidis on 08-Apr-2002 16:43 GMT
In reply to Comment 31 (DaveW):
Ehm, any software for MorphOS should be easily recompiled for OS4.
ScummVM Interview : Comment 33 of 42ANN.lu
Posted by David Scheibler on 08-Apr-2002 17:33 GMT
In reply to Comment 31 (DaveW):
>Still doesnt answer what the problem he had with getting ready to support AOS4
>was
Again: He did not get any developer support from Amiga Inc.(where he was under NDA BTW) or Hyperion.
ScummVM Interview : Comment 34 of 42ANN.lu
Posted by DaveW on 09-Apr-2002 04:56 GMT
In reply to Comment 32 (Alkis Tsapanidis):
Exactly my point so what developer support did he need? Not clear from the interview.
Not clear at all.
ScummVM Interview : Comment 35 of 42ANN.lu
Posted by Alkis Tsapanidis on 09-Apr-2002 12:10 GMT
In reply to Comment 34 (DaveW):
Dev docs, what to do and what not. Programming guidelines.
ScummVM Interview : Comment 36 of 42ANN.lu
Posted by Anonymous on 09-Apr-2002 12:43 GMT
In reply to Comment 34 (DaveW):
>Exactly my point so what developer support did he need
Well, AmigaOS4 is a heavily expanded OS, isn't it? At least the specs say
Exec-SG, VMem, new library system etc. - if you are under NDA and you get
zero information, what would you do?
The same thing with Titan, they said they would consider OS4 port if they
get the neccessary infos, but there seem to be none atm. So it's all a wait
and see.
ScummVM Interview : Comment 37 of 42ANN.lu
Posted by DaveW on 09-Apr-2002 16:14 GMT
In reply to Comment 36 (Anonymous):
But if its API recompile compatible which is what we are all lead to believe then the internal semantic improvements are not immediately relevant.
Any minor issues once beta was finally resolved would, in theory, be based on misunderstandings and quite quick to sort out.
Still do not see enough reason to throw toys out of the pram and start getting political. Like I said Im willing to be corrected if I have mistranslated.
ScummVM Interview : Comment 38 of 42ANN.lu
Posted by Alkis Tsapanidis on 09-Apr-2002 18:32 GMT
In reply to Comment 37 (DaveW):
I agree.
Politics suck! :)
The scene would be MUCH better if we didn't have political clashes... :(
ScummVM Interview : Comment 39 of 42ANN.lu
Posted by Nicolas Sallin on 09-Apr-2002 19:48 GMT
In reply to Comment 37 (DaveW):
I fear it's not as simple.
If you want to write an application using dynamic binding like any
Quake2 port for exemple, then you have three ways:
1. you do your own API from scratch, wasting some weeks on it, and not
as good as something using the OS functionnality
2. you use the documented API (f.g both powerup & morphos has that
since many years) and write your dll code in 30mn
3. you wait some months/years until someone publish the needed API
(AmigaOS4 case) and your project is frozen in the meantime
And this was just one exemple...
ScummVM Interview : Comment 40 of 42ANN.lu
Posted by DaveW on 10-Apr-2002 03:43 GMT
In reply to Comment 39 (Nicolas Sallin):
OK lets ignore Wup and Pup for a bit given that seems to have raised hackles in a way I didnt expect.
If he was to take the AmigaOS 68k version of SDL and recompile on top of that then once the API compatible PPC OS where all these nice runtime libraries are PPC native the job should be a straight recompile... so wheres the difficulty?
Like I said, toys, prams and politics - too much of it in the scene at the mo.
ScummVM Interview : Comment 41 of 42ANN.lu
Posted by Nicolas Sallin on 10-Apr-2002 09:30 GMT
In reply to Comment 40 (DaveW):
If you want to limit yourself to SDL use, then it's your probleme ;-)
I know it was just an exemple you gave, but everything isn't as
simple, really...
ScummVM Interview : Comment 42 of 42ANN.lu
Posted by DaveW on 10-Apr-2002 10:50 GMT
In reply to Comment 41 (Nicolas Sallin):
It was just an example and how he claims to have implemented Scumm.
Anonymous, there are 42 items in your selection
Back to Top