An open discussion on where we go from here

As more and more of computing ends up on the internet, the opportunities for an Amiga redux as a viable alternative for those that grew up with it and find it preferable through emotional familiarity is growing. What is required for an Amiga purchasing decision is to want an Amiga in the first place. Arguments that make sense to traditional Amigans are nonsensical, hollow and unconvincing to the rest of the world to the point where those that would be convinced is so insignificant to be indistinguishable from zero.

The key of course is the browser. With tools such as Google Docs growing in popularity and the final arrival of a reasonably good standards compliant browser on the Amiga based platform perhaps the tide in the rock pool that is Amiga could finally turn.

Having poked around the community forums for a few months, the old irreconcilable divisions are ever present. Many of the old bile filled war mongers have moved on, more than enough have stuck it out and spout the same old same old. What has changed is that Amiga Inc is neutered. Not just as a force but now in potentia also. Good. But the handcuffs of that license agreement seems still to be holding Hyperion back in some way, so perhaps the fat lady hasn't really sung yet.

On the more positive side more portability tools and layers have come to Amiga OS4.1 than ever before, Amiga Forever still does good business, and whilst the PPC market looks moribund AROS has come on leaps and bounds.

The hardware position still sucks dreadfully, and whilst the SAM is a valiant effort (we will come to economics in a moment, as it seems that some still need a basic lesson in viability and business costs) it is both not cheap enough to sell the approx 20,000 license copies that Hyperion needs (if we are to consider the rumoured 2,000,000 EUR figure accurate), and nor is it fast enough for some to swallow their concerns about the ?high? price.

On the OS3.x side, the hardware picture is rosy. Conversely because of the death of the classic platform and because if you wanted you could pick up a cheap PC right now, install Linux, e-UAE and yank the ROM and Key off your copy of Amiga Forever. However, it is a brutal reminder in many ways to those of us that have used AmigaOS4.x (and presumably MorphOS) quite how far things have come on.

Returning to AROS, well have you taken a look at Icaros lately? The problem there isn't so much the availability of cheap hardware, or of the OS (download it and try it hosted, OK its not as fast as I'd have expected hosted), but of matching it up with hardware to make a standalone system. It is also terribly rough around the edges. But, here is the thing, it works. It might turn out to be the great hope of Amiga users around the world, it might turn out to be the last man standing.

Temporarily, lets look back at the PPC side. I have a defunct uA1, a defunct A1-XE and now a SAM on the way while I send off the A1-XE for repair (it is bound to be something simple but time is money and patience is a virtue I no longer possess in abundance) and whichever one I want to keep at the end of it will remain my PPC machine. But would I have bought a Mac Mini instead if OS4 had been available for it as well as a Sam? I can safely say no. A work colleague has a Mac Mini PPC and his comments about it have always struck me as significantly underwhelmed. More so with each passing year. When someone starts talking about modding it extensively, you know it has not really achieved purpose. If it had been the ONLY choice as opposed to supporting someone trying to resell new hardware, then I might have considered it. But such decisions are personal. It will be interesting to see how the MorphOS port to that hardware pans out, whether it re-invigorates that operating system. If MorphOS successfully blazes that trail, then others will need to follow suit.

What I was really looking for, and what would have made my decision easier, was an x86 PPC emulator, with U-Boot that would allow me to run AmigaOS4.x as a development environment. This would not actually have to be done in co-operation with Hyperion, although I can't see anything like it coming out in the foreseeable future. A ?hostile? emulator would be ideal. I could see them shifting another few hundred units of Amiga OS4.1 on it, but again I suspect they are more constrained by the license handcuffs on this than they are prepared to declare publicly. Perhaps it has to be sold with hardware in order to sneak in under a loose interpretation of the contractual terms. Piracy is a concern for sure, but when you are this desperate for users, has the balance of risk reward yet shifted for you?

Heck, if I had an emulator for OS4.0, I might even be able to spend half an hour working on Amiga based stuff in the evenings instead of on extra company time. I wonder how many others are in this position? Portability means productivity in all senses of the word ?portability?.

If so, then I suspect the Amiga Inc contract that has to be challenged and broken in order for AmigaOS4.x to survive and provide reasonable return for the Hyperion and team investment in it over so many years. It is also this contract that perhaps, when we are down to our last 100 malfunctioning Mac Minis, and our last barely functioning Pegasos 2s, Sams and A1s, needs to be ripped up for all our sanities. After all, as time goes on there are going to be less and less working hardware in circulation, not more. That is the importance of projects like Sam, to replenish the working hardware pool.

Ideally there would be a range of hardware available, PPC, ARM or x86 based is not particularly relevant to this point:

1. Low end cheap hardware. Buy it, plug it in, boot the OS and use it from a distribution of previously selected and tested applications as a fun hobby system.

2. Add on cards, accelerators if you will. Be it for classic hardware or for slots in a PC (if not x86). Just some way of enabling the large volume of idle classic owners to get onto the new platform if they want to part with money and plug in the card. Perhaps in here we can categorise the MacMini as another stop gap solution?

3. High end hardware, for gameheads. But the problem here is, exactly what on the Amiga platform would use it at the moment? Then you have to pose the question do the software writers start to devour those extra cycles when they are available, or do we justify hardware once the existing hardware has been pushed to limits?

4. Emulation of the hardware platform (if not x86).

Why? Well, I think we are beyond the point where anyone can seriously consider the Amiga market to be a viable business proposition. At this moment, and for many a year, it has been on life support and scrabbling in the dirt for survival. The technology isn't superior, there are no real technical unique selling points and what it does have going for it is a nostalgic familiarity for a certain type of user who has stuck it out this far ? or nearly this far. What value there was in the brand once has been successively eroded, depleted and debased. It comes down to a stubborn choice for many, and one that many are forced despite their personal desires out onto other platforms.

In order for a viable business to be made around Amiga hardware and software the price point (or risk taken in advance with projected units) has to break even on costs. This is not just the cost of producing a board from the point of view of renting production line space, buying chips in volume, the cost of components plus the assembly time. A viable business has to cover staffing costs, has to cover taxes, overheads, advertising, marketing and a plethora of costs. If what is being sold does NOT cover those basics then it is being priced at a loss.

The same applies to the software. There is no real difference between ?traditional? and ?new? economic models for business, when it comes down to it you either cover your costs or run out of money. This is before we consider the cost of ongoing development for improvements. Consider this starkly. Think about it. Has the realisation dawned on you yet that none of the companies involved can be doing it for money, or the love of money? The fact that they have survived this long shows not that it is viable, but that they rely on costs being absorbed elsewhere. Don't give up the day jobs, folks. Either that or sell something else too!

The "company" structure might have changed but business is still business.

In the middle of this depressing situation we have a technology divergence going on. Let us consider the three or four main players:

1.Amiga OS4.x
2.AROS
3.Amiga OS3.x
4.MorphOS

If you develop for Amiga OS3.x you have a reasonably good chance of being able to port it to AROS without much modification, and AmigaOS4.x with some modification. Obviously, depending on what you do. It is a laughable thought that you could develop for Amiga OS3.x plus RTG and then rely on the emulation layer of Amiga OS4.x to execute it, presumably the same applies to MorphOS.

The ideal situation would be some convergence going on, perhaps a standards group overseeing not just a set of guidelines for developers of device drivers, portability layers and application software to help them avoid making expensive (in terms of time spent) implementation decisions.

I mention drivers because one of the big challenges is to get enough hardware working effectively with any of the above systems. I mention portability layers to bring more Linux based open source applications onto any of the four. I also mention it from the point of view of porting between them as a stop gap solution. The most important development of the next five years, if any of us are around that long in the scene, would be API convergence.


Ideally you could write for one, and recompile for all others with some idea that it was going to work. For the developer this would be a boon, for the users, well it would allow them real freedom of choice when selecting their Amiga-like distribution. For the OS developers, it would mean some real competition. Having a standards group would not mean an end to innovation. Like anywhere, an idea could be tried out first on MorphOS, giving that the advantage temporarily, and if it was successful the demands of users on the other platforms would help it race through the standards group and get implemented on the other distributions.

Of course, there are going to be many out there that will refuse to port using applications as advocacy for a given distribution. Well, so what? Given enough time and developers and most importantly end users for developers to sell (or get kudos from) to competitive applications would appear.

But this is the point, surely now is the time for co-operation? Both AmigaOS4.x and MorphOS face a real threat, not just from the level of bombastic antagonism between a minority of their users but mostly from Anubis. If and when that project is complete, the game will change. In fact, there will only be one game in town. At least, from my point of view, Anubis is a positive threat.

I urge technical co-operation between the development groups. Put personal vendettas aside, and work together on a new AmigaOS API standard that all can share in common. Perhaps this will bypass the license restrictions? If you cannot do this, barring some miracle, all your alternatives (bar AROS and Anubis perchance) face a very slow agonising death. Given the technical brilliance that went into all, I think that would be more than just a waste, it would be criminal. If this is na?ve and out of the realms of possibility, then so be it. I will use AmigaOS4.x and AmigaOS3.x until I get bored with it, and keep trying out AROS (and Anubis when it gets that far) as the potential long term Amiga comfort blanket.

But still, I watch the MacMini port of MorphOS with great interest. Will it prove me wrong and bring in an influx of experimenters whose mind-share we all need, or will it just get called ?a dead OS on dead hardware? on Slashdot and OSNews?

We ignore any possibility of improvement at our collective peril.

@DaveP

Interesting thoughts.

@DaveP

While I salute your idea of a common API for the Amiga-alike platforms, that is going to be restrictively prohibitive. The biggest problem is going to be in the user interface. One uses MUI, one uses a MUI clone and the other, Reaction. To build a common API (wrapper) for those would mean leaving out a lot of functionlity from each GUI system. I have personally implemented a few modern features within Reaction that is not available in MUI, and I know that works the other way around too. I doubt the AROS MUI clone has all the features of MUI 4 either.

This would lead to a common API which misses a lot of the latest features, and would only allow access to any common functions between the systems.

On top of this, creating all these wrappers is a huge amount of work, and even more to debug, who is going to undertake that work, for a less than ideal outcome?

Simon

@DaveP
Interesting thoughts, and many points are valid. I feel however that the divergent os's are divergent for the simple reason that the coders and porters want it that way. As you say the `Amiga' ( I include MorphOs )scene is a hobby for most of us and we choose our respective machines because that is the path we have travelled throughout homecomputing history. I started with the Vic20 for example.
If I wanted a standard I would go out and buy a PC.
I don't want a standard, its as simple as that. I want trailblazing and boundary kicking invention. Os4, Morphos, UAe, Linux, and all other non standard OS's are examples of good independent thinking, whether they will be successfull in the future who knows? and hardware is developing too fast to keep up with anyway. Even in PC land your machine is obsolete the minute it leaves the store.
I am very happy with my hobby, I am sure many others are too, and i am confident that with better access to the internet we will begin to attract more and more hobbyists.

AmigaOS and its associated HW will be fine. I'm not interested in the direction ANY of these OSs (that like to associate themselves to AmigaOS) are heading, NOR do I think incorporating there resources is needed either...

AmigaOS & its HW is going to survive regardless of what people say or think simply because IT IS a Great OS...

I urge AmigaOS Developers to stay quietly working on OS4.x and also stay with the PPC technology and just stay on your targets

It won't be long before ALL the holes are filled in and AmigaOS4.x with PPC HW is going to look very attractive indeed... Those on the fringe better get some nice new winter weather clothing; it's going to get VERY cold out there... heh

@DaveP

I agree with you 100%. We need code compatibility.

We should have ONE AmigaOS API or a converter.

We also should have Linux/MacoS/Windows API mapping software.

A user on the Natami forum posted where Java is heading:

Quote:

SWIXML describes the layout of GUI gadgets in an XML file, so porting GUI gadgets from one platform to another is simple by reading the XML describing the GUI and mapping it to the target platform GUI.

lets just hope things get done.

I must just add that programs created with any sort of common API will not take full advantage of any of the platforms due to feature sets differing.

Simon

@Rigo

If we had a good supply of designers and implementers, we could design a new API towards which all competing platforms would converge over time. The aim would be to design and converge towards a new hardware-independent platform. If the GUI interface were called Wigwam, then Reaction would get a compatibility wrapper to implement Wigwam calls, so would MUI, etc. Parts of the new platform would be coded, tested on all hardware and added to the individual platforms as new API calls.

Slowly developers would begin using the native Wigwam calls instead of Reaction or MUI, eventually dropping the need for new software to use the wrappers at all. The wrappers would only be used by emulators, to run legacy software. The same goes for kernel, library and DOS calls.

To make it work, we would need a steering committee that had the support and input of all platforms towards a common outcome. The committee would need the courage, vision and independence to be able to design and specify the unimplementable, just as NTSC colour TV was beyond the capabilities of the available hardware when it was designed. They would also generate software for conformance testing.

Do we have the people and the will to implement such an approach? It's the sort of project that Microsoft would spend tens of millions on, but in our corner, there are too many Indians and not enough Chiefs. I can't see it happening in my lifetime.

@Rigo
Quote:

Without to deny obvious missing features from MUI4 to MUI3.x and Zune, you can always obtain more modern UI using MUI3.8 APIs instead of use Reaction for OS4 today and you will obtain an application which goes on each Amiga flavour. One between all features which actually Reaction misses is Drag&Drop between UI objects like gadgets. You can't implement a Toolbar configuration like on Firefox with Reaction, while you can with MUI3.8, and of course it will work on Zune:



And this is olny one of many features which MUI3.8 API gives to a dev and which are missing on Reaction OS4 today...

Quote:

If you want use a particular feature of a specific AmigaOS flavour, for example AmiDock support, resource tracking etc of OS4 you can always use #ifdef __amigaos4__ #edif....

@DaveP

Quote:

There are some standards de facto you use if you want to be compatible with all Amiga flavours, they are MUI, AHI, CGX. It's so difficult to understand?

Quote:

You can't see any kind of cooperation from each side, each OS dev group is interested to its OS only. You can see some cooperation from third part devs, and these cooperations have just pruduced things like SDI headers and common framework files to create libraries and mui classes portable (you find it into MCC classes opensources).
The only thing you can't have in common are drivers, you can't write a portable driver. But for normal development you can develop whatever you want following Amiga standards de facto API... the only reason some software were not developed on the Amiga was lack of a fast hardware, because with MUI+AHI+CGX you can develop much more of you have seen during these years. If you want use some OS specific features you can use conditional code, it's simpler than what you believe.
There are many software actually follow Amiga API standards de facto, they are Yam, SimpleMail, WookieChat, SabreMSN, NoWinED, NewsCoaster, AutoDocReader, etc...
So the only difficulties I see with Portable Amiga Programming is into mind of some developer. If you have in mind to write only OS4 software of course you can choose to use API like Reaction, Picasso96, etc... but it's only an your choose, not a tecnical one.

If MUI3.8 API would be so obsolete then Zune which is very young and with some missing features of MUI3.8 wouldn't give any chance to build for AROS an OWB version with dynamic adding/removing of tab pages.
On MUI4 there is a builtin class to make this, on AROS OWB developer has implemented a custom class which is opensource and soon or later will arrive on MUI3.x... and so all MUI flavours will have a common opensource class for these kind of things, like it's happened with TextEditor, BetterString, NList, TheBar... which are currently default installed on OS4 and so a dev has not worry an user has to install these classes.

What you miss yet to write from scrratch an Amiga Portable application? A guide? Here:

http://code.google.com/p/guidetoamigacompatibleprogramming/

@ShInKurO

My point is that none of these things you mention are actual standards (de facto you say, well, not so sure), A "Reaction fan" could doubtless write in saying something that Reaction can do well that MUI finds awkward. And so on.

But there are not just the differences of Reaction and MUI now are there? There is the OS4 library system and how that has changed the programming interface. The entire mechanism, in fact.

We could indeed go down the road that has been mentioned and choose the lowest common denominator to get as much working cross platform as possible but that isn't really getting anyone anywhere more advanced fast now is it?

The idea of a standards oversight (and I agree there are no chiefs, many Indians and plenty of mischief to go around) would be to establish a series of revisions that you could write against now and next.

So, perhaps Amiga API 3 would be the Amiga OS API prior to 4.0, RTG master, AHI et al given standardized names and given where needed a canonical wrapper. Perhaps even a common set of build tools.

Amiga API 4 might be AmigaOS 4.x way of doing things plus Reaction or MUI (or some intermediate canonical API).

Amiga API 5 could be something more advanced yet again.

The idea being that yes, the OS developers and the "third party addon" developers write something to implement the common API. The bake off of technology is done as part of the standards process and is done not on ego, not on who is using what, but on a multi-faceted criteria which includes what allows us to achieve more in future.

A good example might be multi-user support. How does that effect the API? How does that get implemented?

Instead of writing something and using it because "we decided to do it that way", writing something to enable a broader architectural vision might be more sensible. If it can't be achieved when the market and groups are this small, it can't ever be done.

Plus, it doesn't require millions to do it, it requires cool heads, diplomacy and a bit of organisational savvy. Perhaps that, rather than money, is what is truly lacking.

(Note: I'm not engaging in advocacy, I'm identifying items by name to illustrate commonality and deviation)

@DaveP

Hello my friend.

I have been an amiga user since my childhood. There are some problems regarding OS4.x that imho prohibit many users from actually using it. First, as you have stated it's the hardware. Expensive and low perfomance. But let's suppose for instance that SAM was a very - very cheap machine. Let's say that you could get a compete machine (case, psu, hd, etc.) for 300 euros, including vat, shipping cost and a copy of OS4.x.

So now we have solved the problem of hardware, in a way. What about software? A middle user wants some certain things to work straight out of the box. Will his internet experience be at least equal like on every other system? What about a complete office suite? Games?Will apps like AmiCygnix be fast and responsive?

So, imho the main problem is the severe lack of well know apps which will cost at the presentation of OS4.x to a new user in adittion to the lack of high perfomance machine.

Will you say to him that: Hey, there is a fine spreadsheet of 1997 to use? Imagine the look in his eyes. Or: Wait, OWB is moving fast you will have a capable browser of doing this and that in 20xx. Also when he will use apps like Amicygnix or Mame or some SDL app that is either badly written or needs more cpu horsepower to be executed he will see a system crawling and not the fast experencie you have promised him.


I am not saying that the current programmers of OS4.x have done a bad job here. I am telling that we need more of them or a miracle from Hyperion so the platform moves from the hobbish/amiga user target group to a wide variety of users. Only when some major companies see AOS4.x as a system where they can develop and use certain apps for their benefit, we will benefit too. If not, then AOS will conitnue to evovle but in a very slow way like we have now.

@DaveP

You've expressed a great many individual discussion points here. Whereas we can already see in this thread a honing in on some specifics regarding UI, I think the important part is that we're having this general discussion in the first place. The willingness for this conversation to expand and draw more individual's thoughts into the mix, is somewhat "key" to me, rather than any specifics.

The hardest thing to do, and often the solution to many problems is changing how people think. Since UI has been touched on, I'll use that for an example and use ShinkurO's quote:

Quote:

The first inclination of anyone who's worked with something so long is to defend it. Hence, that "defense" is used as the principal argument for never changing.

To get passed such thinking, one has to examine the work itself, be more critical in analysis, and ask themselves what "parts" of that work would fit into a better model.

At that point, you are thinking co-operatively, which is the only way any of what you mentioned in your post will succeed. Co-operation feeds on itself in a positive way, much like defense does..in the other direction.

#6

@DaveP
That's a nice "pipe dream" that's unlikely to lead anywhere. For one thing, I don't think OS4 performs the same on the existing variety of hardware. Wait until your SAM arrives and you may see what I mean. You will probably find yourself scratching your head and saying "gee, it didn't do that on my A1" or visa-versa. We had to have 2 different Quickfixes for SAM & A1 systems. Imagine the complications if the target hardware diverges even further. Your ideas make for interresting speculation but are unlikely to become a reality.

@Cool_amigaN

Quote:

I'm not sure where you are going with this, but Hyperion is not a huge company with a bottomless pit of money burning a hole in their pockets.

As with other platforms, applications are generally developed by third parties, and it is their domain to make the decisions required to bring those products to market. Hyperion are currently involved in a full time endeavour with the Operating System itself.

The latest quickfixes show that work is still ongoing, and indeed more work is needed judging by the mixed reaction to those quickfixes.

Some of the OS4 developers are also working on applications, but to have Hyperion tie themselves up with application development would inevitable stall the OS for that duration.

Simon

@xenic

I sense your post is more about your frustration with your OS4.x computing experience thus far on 2 distinctly different pieces of hardware.

Before you connect that with the idea that Dave is being too ambitious with his thinking, please remember.

The target for OS4.x is not for desktop computing.
Don't forget the constantly presented thought that OS4.x has to get outside of the community with product to gain a safer financial grounding.

Quote:

In addition, the connection to "desktop" has been expressed as more of an emotional tie, than a practical one. It's also been expressed as a "gift" to the community.
Not to mention that those thoughts were expressed years ago, when desktop itself had more relevance than today.

If you think back a bit, you'll recall the AHT stb and the PDA (mystery device).
If either had come to light, then support would not have had the same issues as you just expressed for desktop.
Income from such devices could have funded more of what I think you hoped to see...a stronger desktop to satisfy both community desire and future development.
Either way, though, it is secondary to Hyperion's stated goals. Perhaps we hate to admit such things because Amiga Inc. ALSO talked about having to go "outside" the community to succeed.

#6

@Rigo

Quote:

Yes, of course I understand and accept all the above you mentioned.

What I would really wanted to tell is that if/when Hyperion makes a contract or business cooperation with a large company on let's say.. the mobile indrustry then OS4.x will become widely know. Other companies will look at it (4.x) from a different point of view then. Imagine OS4.x on Nokia for example. Or OS4.x at PS3, XBOX, Wii etc. I am not talking for porting for the sake of porting. I don't even have these systems to run OS4.x. But think the benefit for Hyperion. They will place themselves in the map of serious OS developers more crearly. The more Hyperion earns money, sales etc then OS4.x will expand more and more, faster and faster. IMHO -always- what Hyperion (and OS4.x) really needs is a cooperation with a company that has deep pockets.

Until now, unfortunately I do not see OS4.x running on anything else than a few old A1, PegII and the new SAM but here we are talking for a underpowered and expensive machine that is targeted for a specific industry. We, amigans, don't care so much about cost/perfomance ratio. But the middle user outside of our community will. But OS4.x on a powerfull machine would show its true power. This machine hasn't has to be desktop unit. Make OS4.x run on some product (hardware) alreay know and used widely. Then you will have more potential customers for the desktop version of your OS.

What do you think?

@Cool_amigaN

Quote:

I think you are trying to cleverly rehash the same old argument, of which there is no room here for.

OS4 runs on what it runs on. That situation may change in the future, but when, and if, it does, it will be up to the company in charge, which I sincerely hope to be Hyperion.
No amount of discussing the benefits of porting to a different architecture will influence that decision, regardless of who has to make it.

Simon

@Rigo

That is precisely why we need a detached group (steering committee, if you like) that can ignore the distractions like hardware availability, GUI appearance and third party software. Design what we want to see in ten years' time and then let the developers work towards it.

@tonyw

Exactly ! Im gald a few of us think this way.

Also remember, we need to include/arrange structure for GUI elements from other platforms too for making porting a no brainer.