Discussing strategy of parties involved in the amiga-market (Hardware/OS)

[Crumb]

Last generation of G5 include PCIe (not that it matters much) and some models 4 cpu cores.

As long as "new hardware" doesn't provide the same performance as nearly 10 year old hardware I don't see the point in "new hardware". If I was concerned about that I would buy a brand new Efika fron Directron but I prefer more performance.

Surely, as indicated, MorphOS support for all Macs have gradually brought interest and grownth, but not beyond some 1500-2000 licences. So its a small bunch of people after all.

That would still be 10 times more than the number of x1000 sold and zero investment in hardware development costs. In case you haven't noticed yet "Amiga" word could increase the interest so you could expect more sales.

[amiman99]

I don't know why nobody made AmigaOS for Xbox 360, it is PowerPC afterall! And very common hardware.

I think PS3 would be better! It's got the OtherOS, well, at least mine still has.
If someone makes a port, I'll test it.

[spirantho]

Last generation of G5 include PCIe (not that it matters much) and some models 4 cpu cores.

True - but that's where it ends. Once you've got there there's nowhere to go.

As long as "new hardware" doesn't provide the same performance as nearly 10 year old hardware I don't see the point in "new hardware". If I was concerned about that I would buy a brand new Efika fron Directron but I prefer more performance.

Effectively we're playing catch-up in real performance, yes... but you've got to catch up before you can overtake something. At the current rate of progress, Cyrus will be effectively up to or surpassing the G5 (which I believe isn't very fast per MHz, and rather inefficient), which means that whatever comes next should be better than the opt-of-the-range PowerPC Mac.
As a small company, A-Eon have a large disadvantage for up-to-G5 levels (as Apple had a million times the R&D budget of A-Eon), but after that Apple stopped so it's all forward.

That would still be 10 times more than the number of x1000 sold and zero investment in hardware development costs. In case you haven't noticed yet "Amiga" word could increase the interest so you could expect more sales.

The x1000 isn't the only Amiga being sold, though - nobody expects it to outsell the low-end machines.

[Crumb]

True - but that's where it ends. Once you've got there there's nowhere to go.

Funny to hear that taking into account that PA*Semi cpus are dead end.

The point of releasing AmigaOS for G5 would be gaining time to release it for x86-64.

Effectively we're playing catch-up in real performance, yes... but you've got to catch up before you can overtake something. At the current rate of progress, Cyrus will be effectively up to or surpassing the G5 (which I believe isn't very fast per MHz, and rather inefficient), which means that whatever comes next should be better than the opt-of-the-range PowerPC Mac.

I heard that before with x1000 PA Semi cpu and all in all it's slower than a G5 at the same frequency.

As a small company, A-Eon have a large disadvantage for up-to-G5 levels (as Apple had a million times the R&D budget of A-Eon), but after that Apple stopped so it's all forward.

In case you didn't notice we are talking about machines released 8-10 years ago, not even about competing with nowadays hardware

If you want to look forward to the future the final solution is switching cpu architecture instead of relying on low end embedded cpus. Freescale has always been poor acomplishing their roadmaps.

The x1000 isn't the only Amiga being sold, though - nobody expects it to outsell the low-end machines.

Care to explain how selling OS4 for powerbook g4 would decrease the number of x1000 sold?

Sams have performance similar (or worse) than antique Pegasos boards sold almost 10 years ago in our little market (and higher price, something pretty sad).

[spirantho]

Funny to hear that taking into account that PA*Semi cpus are dead end.

Yes, hence Cyrus using Freescale. But without the X1000 development of things like multi-processor support would be years behind what it is today.

The point of releasing AmigaOS for G5 would be gaining time to release it for x86-64.

Only if AmigaOS is going to get ported to x64 - and the jury's still out on that one.

I heard that before with x1000 PA Semi cpu and all in all it's slower than a G5 at the same frequency.

Maybe, I'm not sure to be honest. But I do know that the PA6T is more or less equivalent power-wise but does it at a fraction of the power consumption. I believe also the memory bus is faster on the PA6T but I could be wrong there.

In case you didn't notice we are talking about machines released 8-10 years ago, not even about competing with nowadays hardware

As I said - the Amiga is playing catch up to the Mac. It's not surprising when you consider the size of Apple compared to the size of A-Eon. Apple could go to Freescale and ask for 100,000 G5s. A-Eon can't. It's a whole different ball-game when you're small company.

If you want to look forward to the future the final solution is switching cpu architecture instead of relying on low end embedded cpus. Freescale has always been poor acomplishing their roadmaps.

I see where you're coming from, but switching architecture just isn't practical. For one thing, I'm not certain they legally can switch architecture, and even if they did they'd need to do a heck of a lot of work (and who's going to pay for that?) AND all the OS4 software would need to be re-compiled.

Care to explain how selling OS4 for powerbook g4 would decrease the number of x1000 sold?

Easy. There wouldn't be an X1000. If you want proof of that just look at the amount of new hardware that's been made for MorphOS since the Pegasos II.

Sams have performance similar (or worse) than antique Pegasos boards sold almost 10 years ago in our little market (and higher price, something pretty sad).

You can get a brand new Sam440 flex motherboard for 667MHz for 270 Euros, that's cheaper than the Peg II was isn't it? I believe launch price for the Peg-II was 299 Euros for the G3 (which is slower at 600MHz, but not far off the same) or 499 Euros for the G4.

You have to remember, though, that the availability of G3 and G4 CPUs back then was way higher than the availability of PPC SOC CPUs these days, so it's understandable there will be a difference.

[vox]

Funny to hear that taking into account that PA*Semi cpus are dead end.

Yes they are. We exhausted almost only stock of them for X1000 at heavy price.
But its interesting CPU arhitecture, somewhat kind of laptop version of G5 with some enhancements.
Sadly it never lived beyond 1.8Ghz and two cores, because Apple was not interested. If it lived
its full frequency range and up to 16 cores (this one was lowest model) it might have better name.

But yet X1000 is and will be only desktop machine to ever use it.

The point of releasing AmigaOS for G5 would be gaining time to release it for x86-64.

There are no current G5 hardware avail at all (as well as G4) so Freescale is natural choice. Plans you speak of are MorphOS plans, not AmigaOS 4 plans - which is OS 4.2 and X2000/4000 as X1000 replacement board.

I heard that before with x1000 PA Semi cpu and all in all it's slower than a G5 at the same frequency.

Yes,that is true, CPU performs about G4 level performance, just at higher clock + Altivec but with faster memory, disk speed, PCI-E and second core. Once all its abilities are used it will surely outperform G4, in some aspects like memory and disk speed as well as GFX cards you can put to it, it might be better then G5 Mac. We`ll post some Linux test once I get the machine, and everyone will be free to compare it to Linux results on G5.

In case you didn't notice we are talking about machines released 8-10 years ago, not even about competing with nowadays hardware

Thats normal as we are stuck in Amiga world.

If you want to look forward to the future the final solution is  switching cpu architecture instead of relying on low end embedded cpus.  Freescale has always been poor acomplishing their roadmaps.

Ah, now we spit on beloved Motorola continuation. Well ,it wasn`t the PPC leader, but kind of like them more then AMCC CPU`s Acube has chosen.

Care to explain how selling OS4 for powerbook g4 would decrease the number of x1000 sold?

While it would make AmigaOne laptop (as it makes MorphOS laptop) there would be no room for A-EON systems.

[QUOTESams have performance similar (or worse) than antique Pegasos boards  sold almost 10 years ago in our little market (and higher price,  something pretty sad).][/QUOTE]

G4 Mac, AmigaOne or Mac outperfor SAM 460, but since it has faster RAM, disk speed and PCI-E GFX it might at least offer comparable overall performance.

We have very few choices in little market, but having some choice is better to none. I well remember dark ages when there was AmigaOS 4.0 out with no hardware to get to run it.

[agami]

Here's my analysis:

Every new business playing in an existing market, or creating a new market, and the Distribution of Innovation follow the same curve. They all invariably face what is referred to as The Chasm; A point at which the ideology and the business practices that got it off the ground need to change in order to gain a firm foothold for continued growth. This is also sometimes referred to as The Tipping Point.



The Hi-Toro guys got past The Chasm by having Commodore acquire them.

For Escom it made sense to buy into the Amiga market even though the Amiga business was flawed. Through it's wholly owned subsidiary Amiga Technologies they had their own plans to change this. Alas Escom had its own troubles and eventually went belly-up.

Escom/Amiga Technologies were continuing an existing business with an existing line of products so there was no Chasm to face.

Gateway also no doubt considered the deal they got was a good one. A diminished but still substantial Amiga market mostly carried by the 3rd parties and the loyal community, and a business that had some nice IP that would fit into their Multimedia strategy. Piggy-backing on top of the Amiga market time meant that they would not need to face the Chasm, or at least when they did it would be easy to traverse as they are not starting from the ground. In the late '90s you could still sell the Amiga as a turn-key business.

This brings me to today and the issues I have with A-Eon's strategy or, dare I say, the apparent lack there of.

Unfortunately things are not so simple for A-Eon. Most of the 3rd parties are gone, most of the development community is gone and therefore most of the users are gone. The market has also fragmented itself with the rise of MorphOS and AROS.

All the signals they have been sending out since their inception is that of a business that believes it is building atop an existing market and therefore can avoid The Chasm. In reality, A-Eon are starting from the ground and are building up a new business and will inevitably face The Chasm. What concerns me is that there is nothing in their displayed behaviour, both direct and indirect, that leads me to believe that they will be able to traverse it. A couple of thousand Amiga loyalists with large sums of discretionary income may get them up to The Chasm, but it's going to take a lot more than that, even if it just modest growth.

[Boot_WB]

@Boot_WB

Thank you. That's exactly the sort of thing we need. Healthy discussion - not bitter sniping and carping.

\o/ It is possible!
Choosing a suitable place helps a lot (ie not within a 'Good news' hardware announcement thread - I've made that mistake myself in the past, and it's understandable that criticism/discussion isn't kindly received at that point. It's the forum equivalent of the morose guy at someone's 40th Birthday party saying "It's all donwnhill from here..."

I completely understand what you mean about the price of the X1000, but I think you misunderstand the reason for its existence.

Yes, it exists to help AmigaOS develop by feeding it cash via sales, but that's not the real issue. The real issue is simply one of market forces.

There is a market for the X1000. This has been proven by its success, and the repeated production runs. That's what it's sold.
It also has the benefit that because it's a very powerful (for an Amiga) machine, productivity of users developing on it will go up slightly. Not a lot, but when you're recompiling MAME for the 10th time, believe me it helps!

And in that sense it's a complete success. I do take my hat off to Trevor for perservering through what sounds like an agonising set of hardware revisions with costs spiralling (hence the increased sales price and delayed release vs initial plans).

It's unfortunate that there's not more money from sales of the X1000 going into the OS4 development, but that wasn't really the purpose.

If its purpose was simply 'to exist' as the top AmigaOne model, then again it has succeeded.  The fact that this was the vehicle of the MAP, and Hyperion themselves chose the multicore hardware, suggests that Hyperions expected to develop the OS alongside the hardware to utilise this.
 

As concerns multi-processing, there are the foundations of multi-processor code in 4.1.6, but no - not really multi-processing yet. I think there may be beta testing going on, but that's really a 4.2 thing I think.

I'd love to see 4.2 come out with a good multiprocessor implementation (Hell, even offloading HD video decoding to a second CPU would be cool), OpenGL, Gallium, etc - but from Hyperion's track record on announcements that don't pan out, or are just never mentioned again, or turn out to be cripplingly slow, or years later than suggested (this is consistent behaviour over a decade) - I just don't trust them to deliver even as a bystander, and I'm most certainly not going to hand over money on the 'gentlemanly understanding' that seems to somehow exist between Hyperion and their customers atm.

If AmigaOS 4.2 does get released by the end of 2014, I'll quote you on this.

Please do. :-) I love fruit cake, and it's be a fun video to make.
EDIT: Just to be claer about timescales, you'll have to allow me a few days lead-time to prepare - the dried fruit needs soaking for at least 48 hours in advance of cake-making. :-)

And again - thanks for the post.  It's good to have healthy discussions, and criticism.

Thanks for the constructive engagement - it's nice to be able to discuss things without being labelled a troll, and interesting to gain insight into the 'happy customer's interpretation.

For me (not sure about others) I think I stopped seeing Hyperion/AmigaOS as a desirable option for my own use once I could no longer bring myself to swallow any doubts over ambguous statemtents and 'have faith' in their chosen direction. If they did ever release a product that is desirable in price point and features I'd consider buying it. But that would have to be features included in the current release, not 'promised' for a future one.

[Boot_WB]

Here's my analysis:

Every new business playing in an existing market, or creating a new market, and the Distribution of Innovation follow the same curve. They all invariably face what is referred to as The Chasm;

Isn't it amazing how if you squint, and make a line of best-fit, you can retrofit most any business theory onto past data (and especially future trend projections!).

Such theories can be a useful (if very crude) tool to promote discussion in boardrooms full of out-of-touch executives. The rest of the world outside the administrative reality-distortion field usually recognises these inane attempts to define the blatantly obvious as a 'theory' for what they are.

"Companies face challenges and have to adapt to engage with a changing market, changing opportunities and changing threats as time goes on."

No graph needed for that though, so wouldn't go down as well during a Powerpoint presentation. If you start proletysing about Sales Funnels I'm getting my coat. :-)

[spirantho]

@Boot_WB

I think it sounds more like you're mistrusting of Hyperion than that you have a real problem with OS 4.

I can understand that, to be honest. From an outsider's point of view, there have been problems, to be sure.
As a developer who spends a fair amount of time reading Autodocs, though, I can see some of the changes that have been made that nobody else sees. I get to see some of the reasoning behind their choices. I get to see the progress that has been made towards things like multi-processing.

As a small business owner, I understand the problems that they face in the real world. What I see them do mirrors my own experiences.

I guess, therefore, I have a somewhat more charitable attitude towards Hyperion.

What I'm really hoping for now is that 4.2 comes out with working multi-processor support, and proper 3D support. It's quite do-able, and it's the only way to prove people wrong.

Until then, though, any amount of talk is worth nothing. But don't let that get in the way of healthy, mature discussion