That whole debate about a certain "neutral" feature not being amiga or for amiga is missing the point - even I as a non-tech person can see that. You can have an extension of your hw/os that doesn't immediately means it's a so called feature, just as you can have a port of a game that uses the os interface where in other platforms it needs its own interface. You make that connection in the same style that Israel, United States, GB, Egypt, etc have connections to the UN. If you look at the UN does it looks like IL or the US or any country from any part of the world? one could argue that it doesn't and miss the basic model of the UN and the way you can shape it and be shaped by it.

The question is, again, divided to 2:

1. Does the target has anything that can be manipulated a certain way, like for example an AmigaDE player/interface or the way ScummVM can take a core engine and do all sorts of stuff to it?

2. Do you get back a way to make your stuff portable so developers can have one hand in Amigaland and another elsewere so they wouldn't have to think about all sorts of implications for being based in Amigaland if anything happens? Perhaps not only considering negative developments but also positive ones in the open sourced world.

Those questions touch the very core of Amiga. If there is a basic identifable model beyond the way the chips were intertwined with the os. Can this model or module be invested or inserted into a flexible socket to provide other things than originally intended. The original model, I think, was never supposed to be about portability. Our should. Like stuff coming back to amiga dedicated software and hardware that can show a different way like incorporating (...) games/apps into the gui and the one screen and its windows (in contrary to the current tendency to always use a new screen).

Last edited by Srtest on 21-Oct-2017 at 10:10 PM.
Last edited by Srtest on 21-Oct-2017 at 10:00 PM.

@WolfpackN64

What you say may be correct. Being in the same ballpark of the 1070, but a light year behind the 1080TI qualifies as “slightly better than a 1 generation old nvidia”

To put the data in context, these are recent (September) benchmarks. Agreed, there is no old gen nvidia, but you and everyone else can look at the gap and draw your conclusions.

https://www.phoronix.com/scan.php?page=article&item=pre-okt17-radeonsi&num=5
https://www.phoronix.com/scan.php?page=article&item=vega-pre415dc-comp&num=1

As for me, I go Radeon for openess, not speed.

@Tellurium

Quote:

As for me, I go Radeon for openess, not speed.

Yes, and comparing open to proprietary drivers is a bit disingenuous.

_________________
Amiga! "Our appeal has become more selective"

@Srtest

Quote:

Those questions touch the very core of Amiga. If there is a basic identifable model beyond the way the chips were intertwined with the os. Can this model or module be invested or inserted into a flexible socket to provide other things than originally intended. The original model, I think, was never supposed to be about portability. Our should. Like stuff coming back to amiga dedicated software and hardware that can show a different way like incorporating (...) games/apps into the gui and the one screen and its windows (in contrary to the current tendency to always use a new screen).

I think you should have look at RetroMode, what identifies the original Amiga500,

* Lowres/Hires/Interlaced NTSC/PAL signal quality.
* Being able to change colors of palette at any point on the screen.
* Being able to cycle colors
* Being able to flash colors.
* Being able produce effects looks transparent, bits in colors defines different layers, like dual playfields, or things like that.

Well all of this can be done in Emulator like EUAE, but at high CPU cost.
but what if we don't need to emulate hardware, what if we port programming languages that use this features, what if there was a library that gives you what you need to make that happen.

and that is what retroMode.library is trying to do, allow programs to take advantage of yesterday's graphics.

You can find retro mode on github, please note that its under development, things are changing from day to day, what worked like x works like y the next day.

https://github.com/khval/RetroMode-AmigaOS-GFXLIB

Last edited by NutsAboutAmiga on 22-Oct-2017 at 12:26 AM.
Last edited by NutsAboutAmiga on 21-Oct-2017 at 11:55 PM.
Last edited by NutsAboutAmiga on 21-Oct-2017 at 11:53 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@Beans

Quote:

Yes, and comparing open to proprietary drivers is a bit disingenuous.

Which open to proprietary? AOS4.1 to Linux? Linux to Windows?

Wake me up in the next 5 years if the open linux amd gfx driver has the windows qt amd features. When you can present in the open source driver a way for the enduser to do all he wants easily like the proprietary driver can do then you can say it achieved something. I don't care about vdpau and the other one and all its backends on top of layers that the non-geek has to go through. I like the choice model but if you can't simply make your own launcher fro gfx/video/internet then you adhere to a bubble.

@NutsAboutAmiga

I actually didn't think about retro for once...

I went to the original design like you go to your mom and dad... or like you approach history from your own culture. Just like in culture, today, when you view the origins differentlly then let us say, 20 years ago that basic concept of original design and idea change as well because the world changes.

Back then the original amiga was given form through the need to take whatever brainiac Jay had in his mind and in all of those minds that could, perhaps produce a machine that control nuclear reactors... to make it an amiga for the people (like someone here mentioned recently). That what the market did. I guess that single element of at the same time integrating and specification can be viewed and adapted to our time. Today you can view it as the intangible that resides between your system and the outside world. An octopus of sorts that lets you do whatever you do in amigaland but on the open source habitat in the way you make it on amiga. Then, that same animal can also make something that can be generic and not personal be something special that has that amiga mark that is hard to simply say what it is. I mean, what platform did Java originated from? You can say it was inflluened by the IBM alliance and you can say it's the example of carying itself from one place to another and being its own platform. If AmigaDE is like that can it still behave like an origianl amiga? Can the opposite occur?

I believe that in terms of emulation it needs to move away from being a kind of a dirty word to do stuff that mimics the past and what was great back then, while also realising that thinking software-only will lead it to a place it can't advance from and make you feel closer to that past. I would suggest that those pushing developments in amiga always try to think about both past and present. I downloaded RetroMode and tried to get my head around it. I like some of the output planned for appearing in-windows? Am I putting this right? streaming classic "forms" inside modern implementations?

Last edited by Srtest on 22-Oct-2017 at 02:34 AM.
Last edited by Srtest on 22-Oct-2017 at 02:21 AM.
Last edited by Srtest on 22-Oct-2017 at 02:20 AM.

@Srtest

AMIGADE ??? the momentom has gone.. the trian has left the stasion that boat has sailed.

@Srtest

I'd probably have to read that a few dozen times to decipher it.
That said I just PREFER AMD gpus to Nvidia, and with the performance excesses available I don't really need to worry what is the fastest.

Besides, Linux always lags behind Windows in its ability to use gpus to their fullest.

As far as emulation goes, wish I could say I didn't have much use for it, but running 68K software on any NG system IS emulation.

And I can sort of see the point made about hardware features inherent in the Amiga chipset, but instead of running legacy software that needs them, I can just run a modern equivalent. Problem solved.

Hard to believe we're still talking about ideas like AmigaDE.
Personally, if we want to dig up old ideas that might have worked, I'd prefer to discuss QNX.

_________________
Amiga! "Our appeal has become more selective"

How could Java have been so loved and despised at the same time? At a certain time in the past it was considered a breakthrough, even mimicked by evilcorp with their m$ virtual machine. Then after that it suddenly become hated. Now it's some sort of a standard.

How can Amiga be alive and in color yet still preserve its zombie state?

What's old can be new and vice-versa, espcecially since it didn't come to pass.

Look at Nuts Retro library, isn't it kinda like bringing an amiga player to current amigas?

The question is - what is a Java-like thing as it relates to the cpu and the gpu? It can have gfx implementations or none at all. It can be cpu-based or simply be about the web. A player of sorts that tries to lack any needed roots might limit the reliance on the cpu or cause eventually to need it because it isn't optimized for anything we use. If you move away from that axis you can see the different question - If the cpu is actually a soc and the player embodies chipset-like elements or adequacies between them then it already has something in common. I guess this is why a developer will target the language. Can a language after the age of assembly be Amiga? Were there any thoughts about an amiga-made language?

@Srtest

Quote:

How could Java have been so loved and despised at the same time? At a certain time in the past it was considered a breakthrough, even mimicked by evilcorp with their m$ virtual machine. Then after that it suddenly become hated. Now it's some sort of a standard.

Java never lived up to its promise; java was never ported to ATARI, AmigaOS, and other older systems.
Java security was hyped as result it was used in the wrong placed, for banking.
Java was just a vertical machine, that emulated none existing computers, as result the GUI was poorly integrated with rest of the system, it had It own style and look.
because nothing was native, graphic function, byte code, GUI or even the disk access it horrible slow.
and despite it security clams, it was horribly, it had lots security holes.

Now days web pages has moved to more secure 2 level security, where need mobile phone and PC, to login to net bank.

Quote:

How can Amiga be alive and in color yet still preserve its zombie state?

Because is noting revolutionary about Amiga graphics,
How a look at this MSDOS demo from Future Crew.

https://www.youtube.com/watch?v=rFv7mHTf0nA

The rest of the world don't get what we were so special about Amiga graphic. VGA killed AGA in so many ways.

As shown there was time, where PC tried to mimic Amiga effects, but it moved over to 3D quickly, there was period where 3D was not impressive, it had to be animated 3D objects like fishes and dolphins.

Quote:

Can a language after the age of assembly be Amiga?

Assembler is just a language as well, but there different Assembler languages for different CPU's. The machine code is the real deal, HEX / binary numbers stored in memory.

You might in theory create Motorola 680x0 assembler language for PowerPC, instead of generating 68000 machine code you might get PowerPC machine code instead.

But it will be limited by fact the PowerPC is more advanced CPU, more CPU registers more instructions and so on, so in that sense it bad idea.

There is trend now to make things that look old school retro, indie developers. And this where think retro mode has place, because retro is a thing.

Telling people to adapt C code, there is something holding developers back. it might be that are used to coding for AMOS, or BlitzBasic, and think, if can bring this languages into NG, we can bring more developers over.

Last edited by NutsAboutAmiga on 22-Oct-2017 at 09:58 AM.
Last edited by NutsAboutAmiga on 22-Oct-2017 at 09:51 AM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@Beans

Quote:

but running 68K software on any NG system IS emulation.

It's recompilation, JIT (just in time), (680x0 to PowerPC machine code) sure memory layout chip-sets are emulated. but it can not compete with native machine code.

In every conversion there is a loss, no matter what you do, you see it image compression, or you see as factor of speed, like you see in JIT compiler, fact that you have to convert some ting from x to y, always result in extra energy loss.

I was going to say Entropy, to look smart, but I have no clue about thermodynamics , so maybe should keep Entropy out of it.

Last edited by NutsAboutAmiga on 22-Oct-2017 at 10:20 AM.
Last edited by NutsAboutAmiga on 22-Oct-2017 at 10:19 AM.
Last edited by NutsAboutAmiga on 22-Oct-2017 at 10:16 AM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@Beans

The Linux GPU problem has been getting better with leaps and bounds. AMDGPU on Mesa is about as fast as it's Windows driver now.

@NutsAboutAmiga

Quote:

I was going to say Entropy, to look smart, but I have no clue about thermodynamics , so maybe should keep Entropy out of it.

No, leave it, I like the word, although I'm not that familiar with thermodynamics either.
But I like Phillip K. Dick's use of the term.

As to emulation/recompilation/JIT, to me its all pretty much the same thing.
Although I will agree that the latter concepts lead to lower performance losses.

And Java? I'm glad to see it waning. The concept (and AmigaDE) was interesting, but with the number of common ISAs reduced to just a few, I'm not sure transportable end code is that important.

I am curious why you guys didn't bite thew hook I threw out about QNX (although that IS off topic).

I just reviewed that again last night and its still pretty slick.

_________________
Amiga! "Our appeal has become more selective"

@WolfpackN64

Quote:

AMDGPU on Mesa is about as fast as it's Windows driver now.

Not quite. And then there's Vulkan and DirectX, which offer performance advantages over OpenGL.

But, I will agree with you on "The Linux GPU problem has been getting better with leaps and bounds".

Performance is certainly more than adequate for virtually any use.

_________________
Amiga! "Our appeal has become more selective"

@TRIPOS

Quote:

Not at all. I wasn't talking about new languages, new toolchains, virtual machines or whatever.

I was focusing on your statement about CPU agnostic bytecode.

Quote:

I was talking about what we have today, like a Trance JIT compiler, but instead of compiling native binaries out of 68k binaries it would compile it out of "Morph" binaries.

Would there still be direct Trance 68K to native? Or would there be Trance 68K to "Morph" translation?

Sounds like another use of the A/Box with the Q/Box being portable to any CPU.

@NutsAboutAmiga

Quote:

I was going to say Entropy, to look smart, but I have no clue about thermodynamics , so maybe should keep Entropy out of it.

Making a correlation between entropy and thermodynamics suggests you do have a clue about it.

1. Bits and bytes are neither created or destroyed.
2. Lossy compression will always run data down.

Ha!

@Srtest

Quote:

How could Java have been so loved and despised at the same time? At a certain time in the past it was considered a breakthrough, even mimicked by evilcorp with their m$ virtual machine. Then after that it suddenly become hated. Now it's some sort of a standard.

preface: Do you use apps on an Android phone? If so, then in all likelihood you are running Java programs all the time. Java has been the main development language on Android. Google seems to be moving towards Kotlin now, or at least they've opened up to it and are promoting it on Android Studio. But even so Java remains the default, virtually all Android programs are written in Java, the support libraries Google supplies are written in Java, and the tutorials are in Java. When people say they hate Java, keep in mind that they are putting themselves in opposition to places like Google, who use it all the time in place of languages they've developed in-house.

loved: Java was the "first"*** solution an important problem: write once, run anywhere. For most applications, that really worked. It had a well-thought-out library to back it up, was object-oriented, and took care of garbage collection. Unlike C++, a hacked-together language whose overriding priority was speed and compatibility with C (and suffers correspondingly -- only now, 35 years later, are they talking seriously about adding a module system rather than persisting with include files), Java is a well-designed language that is relatively simple to learn.

hated: Some reasons have been pointed out (the Swing framework for GUIs) but never underestimate the Microsoft machine. They broke their contract with Sun by implementing a compiler that compiled a variant with Java, refused to fix it, and after Sun revoked their license they set up C# as an alternative. After that it became a marketing war.

I'm not really sure why else Java is hated, except perhaps that a lot of lesser coders want every language to be like the language they first learned. It's a bit like 68k v. x86 architecture: I've met loads and loads of people with solid reasons why 68k was a much more elegant architecture that is easier to work with, in part because it was a clean break with Motorola's 8-bit processors. Whereas Intel made a conscious decision to keep what was bad about the 8086 and 8088 designs when they went to 80286, 80386, etc. Yet people who were raised on Intel processors and didn't want to learn anything new stuck with Intel in their design decisions.

Similarly, Java made a clean break with previous languages, and it showed, especially at the time. Whereas C++ made a conscious decision to keep what was bad about C, which made necessary monstrosities like copy constructors, and also meant that you can find C++ programmers who whine with a straight face that they have to work with code that uses references instead of pointers, because all those crashes caused by Kernighan and Ritchie's decision back in 1969 to pass arguments by pointer instead of by reference were such a great idea. (A rather strange decision, given that safer methods were already known.)

***Not really the first, but by the mid-90s Pascal's P-code wasn't backed by a major corporation, and no one in his right mind uses LISP (outside of academia, apparently).

_________________
I've decided to follow an awful lot of people I respect and leave AmigaWorld. If for some reason you want to talk to me, it shouldn't take much effort to find me.

@Beans

Quote:

I am curious why you guys didn't bite thew hook I threw out about QNX (although that IS off topic).

I'll bite. Are you talking about that Gateway-related project 20 years ago, where they were going to collaborate with QNX on the next Amiga? I thought that was a great idea, though I could find almost no solid information on it in the press at the time. I remember something about QNX devs being happy because they'd always liked Amiga.

Alas, Gateway's upper management got wind of it and killed it. As far as I can tell, that was the last nail in the coffin, the one that led us to the situation we're in today.

_________________
I've decided to follow an awful lot of people I respect and leave AmigaWorld. If for some reason you want to talk to me, it shouldn't take much effort to find me.

It's funny how things change because the perspectives change and like you can't seperate a cpu from a gpu (while the "big picture" model becomes more and more fragmented) and now you talk about the language as something that's supposed to be an enabler of hardware and cpu from one side while being agnostic and universal from another side.

The Amiga's use of assembler although implemented as something specific to amiga was pretty universal in what it attempted to do. I remember from high school how people that learned it hated it with a passion. Now you have something that resembeles the desire to talk with the hardware directly via Vulkan (or Nova) like the old amiga created that relationship between the chipset and the os/software.

Then there's the perspective of Java of being entirely unrelated so to speak to that argument and simply trying to be universal without the need to be connected to an architecture in any way, shape or form. In theory that's nice, in practice that's adhering to the bigs who can put a lot of resources to make it a non-issue in terms of performance.

Now, this is where I'm confused. In 1985 amiga showed itself as the host and then something like msdos as the hosted system. Now you have a language like Java that tries to be a host for a computer system? that seems kinda illogical. It might work because you either push a lot of horse power through a bottleneck or optimise it to oblivion.

Funny thing about Pascal as it was the only language I ever learned and caused me to leave amiga back in the day when I was in high school as I was really good until it was made clear to me that my deficiencies in math and physics will hinder my progress so I better choose something else.

Regarding the topic, I wonder what is the relationship between a language that characterizes the web like Java and a cpu of any sort. Is the load a general load of the web that gets carried by the cpu or does it leaves pathways for you to offload certain tasks to other parts of your system like the gpu and memory. Maybe the question is - why isn't evertyhing rendered if they even left out a 2d unit on the gpu? I also question the ability to inflluence a development like Java if you come from a small habitat like amiga so maybe the thought about presenting an alternative becomes a basic need. m$ in that regard behaved like they are small and the underdog not the dominating force with resources.

@Nuts

I said in living color vs zombie as a metaphor for it not being a contradiction. People from the outside might say it's illogical and a bubble yet for those who wants all sorts of different things from their systems they use it and have a passion about it that make it alive. It might be a small group in rational market-based terms however every one brings something to the table.

You wrote: Quote:

the GUI was poorly integrated with rest of the system

However, when people simply needed something like that without it having to do with any system in particular, it didn't matter if it sucked according to perfomance and stuff - they just ran with it and it was logical like laying down infrastructure.

Last edited by Srtest on 22-Oct-2017 at 05:57 PM.
Last edited by Srtest on 22-Oct-2017 at 05:52 PM.

@Srtest

Quote:

Then there's the perspective of Java of being entirely unrelated so to speak to that argument and simply trying to be universal without the need to be connected to an architecture in any way, shape or form.

As I wrote elsewhere, Java wasn't the first. Some Pascal compilers relied on a bytecode called P-code so that you could compile in one place and run anywhere you had a P-code interpreter. LISP did something similar. I'll note that today we now have llvm as an alternative; Apple seems to be using in their devices (as I understand it, clang compiles to llvm bytecode).

The thing is, adding two int's is a pretty universal and important task in most computer programs. Loading data from memory, and writing it back out, is even more fundamental. Every CPU does this; they give it different names, and some CPUs will have particular features that others don't, but most of the tasks are the same. So a virtual machine need merely address this universal background; translation is then easy. The VM can also at times optimize bytecode depending on various circumstances.

Where Java excelled was in a universal, consistent, well-designed interface to network and graphics programming. There was nothing like it at the time. Did you suffer a penalty for it? Yes, but typically the penalty was worth it.

Quote:

I wonder what is the relationship between a language that characterizes the web like Java and a cpu of any sort. Is the load a general load of the web that gets carried by the cpu or does it leaves pathways for you to offload certain tasks to other parts of your system like the gpu and memory. Maybe the question is - why isn't evertyhing rendered if they even left out a 2d unit on the gpu?

I'm not sure what you're saying, so I'll try to answer a question I think you're asking.

The way Java works is that you have to have a bytecode interpreter on the client machine. The host delivers the bytecode (a ".class" file) to the client, whose CPU then runs the virtual machine, which loads, interprets, and runs the code. The host machine doesn't have to understand Java at all; it merely delivers data.

As a language, Java doesn't "characterize the web;" you didn't have to do any web programming with it at all, and it was still a pretty good language. Rather, Java characterizes typical useful tasks of general-purpose programming, in an object-oriented paradigm, and included APIs that were well-designed to work with the network. Much as Pascal was the right language at the right time, but in its basic form wasn't actually that useful, Java was the right thing for its time.

When communicating on the network you don't necessarily know your interlocutor's architecture. With Java you don't have to learn a different API for 10 different systems in order to make something work (BSD, Linux, AIX, Windows, MacOS, OSX, SunOS, QNX, etc.); you write for one target, and that target took care of the 10 different systems.

The trick is (a) to design the language and the APIs in a sufficiently abstract way that they'll work anywhere in principle, and (b) to implement the virtual machines to so efficiently that the performance hit from translation is, on average, less than the performance hit from an average coder's result. In addition, some virtual machines are look for certain patterns in bytecode and adapting the corresponding native code for them.

_________________
I've decided to follow an awful lot of people I respect and leave AmigaWorld. If for some reason you want to talk to me, it shouldn't take much effort to find me.

@simplex

Quote:

The thing is, adding two int's is a pretty universal and important task in most computer programs. Loading data from memory, and writing it back out, is even more fundamental.

What you are writing sounds to me at its basic form as computer logic. What happens when you need to compare to elements that are based on a different kind of logic or at least doesn't correspond to each other? I can think about a SoC because it entails the basic model of a system, which is processing power, gfx, I/O, net, etc. It's the basic model so the SoC mirrors it. There is a logical continuation. Now, if you take the entire net as an example then its logic completely departs from that of the computer model. You have cloud-based modeling and things that live in the virtual world. Is Java like the web in itself or is like the web if it can characterize it first? This is like a universal language that creates its own layer and I don't understand how the basic computing model like that of amiga (old and new) is manifested. I understand the enviroment argument I just don't get the merit of that as amiga is also a kind of corresponding signals of a language.

Quote:

So a virtual machine need merely address this universal background; translation is then easy. The VM can also at times optimize bytecode depending on various circumstances.

The way I think about it, I come to a place where I see two different axis or perhaps even a second dimension by comparing vertical to horizontal. If my computing model, let us say, with something like a Xena chip that wishes to send commands to net transportations can't process efficiantly that way of communication then I can see a lot of resources going to waste and a chip being declared insufficiant. If you want you use a library in Nuts's style with his Retro library then you at least think about the classic model you wish to simulate in software and then host the system. If AmigaDE was made then perhaps it would be able to be something that can benefit us more than a something like java that simply has its own basic elements which are easy to move from platform to platform but maybe hard to shape to benefit you specifically. The net doesn't need to benefit you as it is a benefit in itself. I would think a programming language should have merit for different uses and implications in a way that is not generic.

Quote:

Where Java excelled was in a universal, consistent, well-designed interface to network and graphics programming. There was nothing like it at the time. Did you suffer a penalty for it? Yes, but typically the penalty was worth it.

A good enough theory (from my field) is something which has its own core (based perhaps in a way you view people and the world) yet you can meld it to better suit your needs, not the other way around. You can observe the fragmented world of devices how usage can vary. I have a feeling a lot of smart devices don't use Java if all they need is talk to each other and either be about enabling a few features or be based on a streamlined version of linux that both can communicate directly and fit a wide area.

Quote:

I'm not sure what you're saying, so I'll try to answer a question I think you're asking. it merely delivers data.

I think I understand the merit of hosting something on a cpu and reducing the level of interaction to "data". When you wish to present different charactaristics of that data or maybe if you will, datatypes, then you can also think of offloading different parts of the data to whatever suits its processing, be it a library like Nuts's Retro or a video chip.

Quote:

sufficiently abstract way

I think I will stop here as I'm way out of my element and between us it feels a little bit like hijacking. Thanks a lot for broadening my horizon. This sounds like what Linux gives you with its abstraction of api and layers and backends and then you also have Mozilla and their contribution to that. Maybe i'm wrong I don't know if we do it similiarly here. Maybe we do and we are those that know how to simply talk to data.

Last edited by Srtest on 23-Oct-2017 at 12:21 AM.