** HOT FLASH ** EFIKA is now the biggest ever Team Aros Bounty! **

[falemagn]

@ takemehomegrandma

Hollywood already runs on AROS.

However, AROS itself already runs on the Efika too, albeit under Linux, so what would AROS running on the bare hw offer more than it running under Linux? It will rather offer less: less drivers, no virtual memory (as in swappable memory), less ability to track bugs down by means of a debugger, etc...

[pixie]

Speed? More 'real' resources? And the fact that if it is not the bes solution you can always have the hosted version on Linux?

[Georg]

I will continue to claim Firefox/Mozilla will not happen

I don't think so. Once one manages to make a working version using X11, one can do it like the RiscOS guys and instead of using real X11 use a X11-to-native-windowing-system wrapper (in RiscOS they have something called ChoX11).

[bhoggett]

@Fabio:

AROS is Open Source, this means that anyone can take its sources and do whatever he wants with it, even port it to the Efika. On the same fashion, anyone is free to offer as much money as he wants in order to incentive the port: there's an offer waiting for demand, a demand which might not be there at all in the first place.

I understand that. However, not all open source projects are bounty driven, and the ones that progress best* are the ones which follow a development roadmap and at least a loose schedule.

The ones that tend to drift wherever the wind (or bounties) take them are the ones which tend to get bogged down and stagnate.

In fact, if you haven't noticed it, no one has taken up on the bounty yet.

It would be damaging to AROS if and only if

1) someone had taken up on the bounty

True.

and

2) by accepting the bounty, that person would subtract valuable time to other kind of AROS development.


Considering that (1) hasn't happened yet and (2) is quite unlikely to be the case, there's simply no damage being done.

Well, we can't say much about (2) until the bounty is taken. Even then it may be impossible to know if the person who accepts it would otherwise have chosen to work on something more advantageous for AROS as a whole.

I remain of the opinion that a port to Efika is primarily meaningless at best and will use up resources best used elsewhere at worst. People who want to run things like Hollywood on Efika would do so on MorphOS - as that OS is considerably ahead of AROS in development and will probably remain so for the foreseeable future. Any suggestion that 68k emulation would be easier to do on PPC is pointless as well because in that case it won't be portable and thus of no use to AROS in general. The only benefit to AROS would be if it helped port more of MorphOS into AROS thus feeding back into the general code base, but as there seems to be very little enthusiasm for convergence from either MOS or AOS4 I don't see much mileage in that either.


* by this I refer to relatively complex team projects. Single developer OSS projects tend to follow a different pattern.

[Cymric]

Piru was not (entirely) referring to the technical problems, but also to the many, many, MANY hours constructing all of the wrappers, interfaces, abstraction layers and more. This is extremely tedious, extremely boring, and extremely not-rewarding work.

[bhoggett]

@takemehomegrandma

Hollywood and the Efika is a *perfect match* IMHO.

Perhaps, but unless I am very much mistaken a MorphOS "port" to Efika will be ready before a native AROS one will, and Hollywood works on MorphOS.

Alternatively, AROS works hosted on Linux so Hollywood on AROS on Linux is already an option - and it benefits from extra driver support compared to what is likely to be available to a native AROS version.

[falemagn]

Speed?

AROS hosted isn't slow at all.

More 'real' resources?

What does that mean?

And the fact that if it is not the bes solution you can always have the hosted version on Linux?

If it's not the best solution, why develop it at all?

[pixie]

AROS hosted isn't slow at all.

I would assume, pedantically I may assume it, that it would be slower then AROS native, not meaning by it that it was slow at all

What does that mean?

That if AROS isn't hosted it bangs for the metal whereas on hosted it doesn't, or atleast it uses linux ones (it was mainly a reference for using swaps memory from linux). On native the resources needed for Linux are gone and only those of AROS are taking into consideration

If it's not the best solution, why develop it at all?

Because the world isn't flat, and if it isn't the best suited tool for one job it might well be for another... take linux, it might not be the best tool for desktop but it is seen as such server wide

[iamaboringperson]

HOT FLASH

I don't like the sound of that!

[Fransexy_]

Cymric wrote:
Piru was not (entirely) referring to the technical problems, but also to the many, many, MANY hours constructing all of the wrappers, interfaces, abstraction layers and more. This is extremely tedious, extremely boring, and extremely not-rewarding work.

$10,373.02  are not rewarding?  :-?  :-?  :-?

[Cymric]

No. Considering the amount of manhours which are sunk into this nigh bottomless pit, $10.000 is a laughable amount of money. I doubt you'd be able to hire a professional programmer for more than 3 or 4 months for that amount. Outsourcing to China or India gets 3, maybe 4 programmers, I think.

Apart from that, the non-rewarding part stems from the fact that it is rather low-level stuff which isn't very visible nor endowed with a high 'wow! cool!'-factor. To the user, it isn't really special when a window with a few gadgets is put on-screen, while in fact this is a very, very tough nut to crack in this project.

Have you ever tried Amigafying a program which is breathing Unix at every line of code? I have, and that was a program 100 times smaller than Firefox which didn't use any graphical UI nor publically available low-level libraries such as GDK, GTK+, libpng, and the like---in the end I got so fed up that I simply used the program on my PC instead. (The nasty proneness to crashing of the Amiga made coding a hell, too.) The people toying with the Firefox port are in effect backporting 10 years of software engineering advances on the Unix-platform to a platform which, while sharing some rudimentary characteristics, operates in a radically different way, nor is really supportive of projects of this magnitude and complexity. But hey, it's their spare time, not mine. To each their own.

[falemagn]

pixie wrote:

AROS hosted isn't slow at all.

I would assume, pedantically I may assume it, that it would be slower then AROS native, not meaning by it that it was slow at all

AROS hosted is certainly slower, but on x86 and by a marginal factor. This is essentially due to the fact that AROS native has virtual memory shut off, which saves about 10% or less of speed. As far as I know (but I might be wrong), you can't shut virtual memory off on the PPC.

What does that mean?

That if AROS isn't hosted it bangs for the metal whereas on hosted it doesn't, or atleast it uses linux ones (it was mainly a reference for using swaps memory from linux). On native the resources needed for Linux are gone and only those of AROS are taking into consideration
[/quote]


But in practical terms, why would that be of your concern? You'll see no disadvantages from that, just advantages.

If it's not the best solution, why develop it at all?

Because the world isn't flat, and if it isn't the best suited tool for one job it might well be for another... take linux, it might not be the best tool for desktop but it is seen as such server wide[/quote]
[/quote]

If it's best suited for some jobs, better know that in advance rather than find it out only when the port is done: it's not like time and efforts come for free.

So, if you have in mind some things for which AROS natively running on the Efika is better than AROS running within Linux on the Efika, just say it? :-)

[pixie]

So, if you have in mind some things for which AROS natively running on the Efika is better than AROS running within Linux on the Efika, just say it? :-)

It's good for AROS ego! ;-) I think efficiency is key regarding EFIKA, and AROS being an Am*ga OS represents it, adding a layer goes one step backward, unless you could integrate both user interfaces to run as one, AROS apps would be then seen as the same as linux... and... after 'seeing his work god saw it to be good' ;-)

[FerretSimp]

Reasons for the Efika:

The whole system, a dedicated AROS PPC system , can be built for under £200.

It's Small, so it won't take up your entire office, like a PC based Amiga.

It will be the first "official" AROS computer. That's actually quite a big step when you think about it. A machine that you can go out and buy, specially to run AROS on.

Obsolete? Yes. But so's your 7.14mhz A500, and no one seems to complain about those around here. It gets you a proper Open-Amiga for a low price, in a small form factor that you can stack on top of your PC.

The point is, it's a beginning. Yeah sure, we'd ALL love to see AROS on a 64 bit Three-core Multiple gigahert IBM PowerPC system, but it's just not going to happen, is it!? And even if it did, almost no-one here could afford them anyway.

So stop griping at the project paid for considerably by the company who makes them, ok? The fact that any company is taking an interest in building AROSMIGA's is great.

I welcome it, at least.

[koaftder]

FerretSimp wrote:
Reasons for the Efika:

The whole system, a dedicated AROS PPC system , can be built for under £200.

It's Small, so it won't take up your entire office, like a PC based Amiga.

It will be the first "official" AROS computer. That's actually quite a big step when you think about it. A machine that you can go out and buy, specially to run AROS on.

Obsolete? Yes. But so's your 7.14mhz A500, and no one seems to complain about those around here. It gets you a proper Open-Amiga for a low price, in a small form factor that you can stack on top of your PC.

The point is, it's a beginning. Yeah sure, we'd ALL love to see AROS on a 64 bit Three-core Multiple gigahert IBM PowerPC system, but it's just not going to happen, is it!? And even if it did, almost no-one here could afford them anyway.

So stop griping at the project paid for considerably by the company who makes them, ok? The fact that any company is taking an interest in building AROSMIGA's is great.

I welcome it, at least.

800MHz PC w/ 256MB ram, 20Gig HD from dumpster: free

You can goto a thrift store and pick up a PC with twice the power of an efka for $20 and run aros on it. To me, aros on efika is worthless.

I do hope a port is made avaiable though, for the only reason that it will help force the codebase to be multi arch.

[FerretSimp]

koaftder wrote:

800MHz PC w/ 256MB ram, 20Gig HD from dumpster: free

You can goto a thrift store and pick up a PC with twice the power of an efka for $20 and run aros on it. To me, aros on efika is worthless.

Wow! The wonder of the free country. Over here, they'll charge you $200 for a P2.

And you can't directly compare the processor megahertz, the 400mhz PPC chips tended to be at least as powerful as the 700-800mhz chips they were competing with.

Maybe the PC would have more ram, but it's still a huge, butt ugly PC. PC. Not Amiga, not AROS computer. I hate PC hardware. It's shoddy, cheap and unreliable. My PC's tendto average 5-6 years TOPS before they burn out, while some manage 3 years. I have UNIX workstations and Amigas and goodnes knows what else, running still when they were madein the early nineties.

In fact I think my oldest functional PC is around 1997 after continual repair with NOS parts, whereas I have a BBC that still work perfectly.

If you want to stick with a PC, stick with a PC. GO install windows. I'm happy with getting an Amiga.