c-FMR : The questions of this interview were sent just a few minutes before Palm annoucement.
1)What happen to you since the Met@box ?
Well, I haven't been away from Metabox all that long. Last fall, I
started to work with the Metabox Corporation (the US company) and an
outside engineering company, with the intention of developing a US
engineering group. This was going well, but lacked the necessary
funding.
I put some of my own cash into the deal, and in May, officially left
Metabox AG to work for the US company fulltime, as Chief Technology
Officer. However, Metabox AG didn't meet a number of promises they'd
made in regard to the US company, and as well, their flirtation with
bankruptcy directly affected the US company's ability to seek funding.
In June, Metabox USA closed their door, and I was basically left
without a job.
How did you get involved into Merlancia Industries/MISEL?
I have known Ryan C. for a while before this. Ryan was interested in
reselling the Metabox "Phoenix", a set-top-box, or more properly,
"convergence computer" (eg, a full fledged computer, but for plain old
everyday users, centralized around media applications for your
livingroom, rather than business/engineering applications for your
office).
2)For how long have you been working on that project?
Well, the stuff I'm initial doing for Merlancia began while I was
working for Metabox Corp. They couldn't pay me, so I kept all my work
under my own name (the AG was supposed to be paying, but they didn't).
The goal, while at Metabox Corp, was to design a short-term
architecture
improvement to the Phoenix. In Europe, the target is clearly DVB, and
the Phoenix architecture handles that just dandy. But here in the USA,
the goal was advanced broadband entertainment, and I felt that we had a
number of shortcomings in the Phoenix architecture.
Why starting again with a PPC machine?
Several reasons. PowerPC is a fairly simple replacement for the
ColdFire
used in the Phoenix, but with greater performance for the price. Other
Merlancia projects (cooperative developments with bPlan, for example)
use the PowerPC, so there's value in maintaining the family. For a
set-top box, I really can't care all that much about the CPU, since
you'll either have proprietary/bundled applications, or you'll have
machine independent applications, like JavaTV, MHP, perhaps something
like Tao/Intent provide. I don't plan to get locked into any specific
CPU in the "device" space, that's as bad a move as any other hardware
lock we've seen in the past. And PowerPC has it over x86, for embedded
applications, in spades.
3)Isn't that machine a resurrection of the Pios One/TransAm machine
?
Not even remotely. The PIOS One was a mid-to-high end desktop media
workstation. The new thing is a fairly low cost livingroom computer.
Most users won't even think "computer", they'll think of it as a
multi-featured appliance.
4)What are the major differences between the 2 machines ? Isn't the
Tsunami a simplier version of the Pios One ? Is it CHRP ? Why only 2
CPUs ? Isn't that machine a bit limited compared to today hardware ?
I wasn't involved in all the bPlan stuff, the decisions, etc. They're
not really doing something like the PIOS One architecture. You
definitely wouldn't clone a PIOS One today, anyway. Whether there'd be
any value in following that gestalt, I'd really need to think for
awhile,
and perhaps work up the next generation. The bPlan stuff is so close,
there's no need for me to be involved. With Merlancia building a real
engineering department, with me at the top of that, I'll most
definitely
be more involved with future directions.
5)What market(s) do you target ? Is it a geek machine?
Some of the bPlan system applications are definitely for geeks or Amiga
fans who refuse Windows. A huge company couldn't launch such a system,
but Merlancia small enough to do this. I'm also a believer in the idea
that you can be profitable via one huge market, or numerous niches.
We're not building single-use systems.
As for the STB stuff, it's not a geek machine, though it's quite likely
geek/gadget freak types will want one in their TV room.
This is a
multi-purpose machine that does a wide variety of things. Sure, you
could do that with a PC, but it wouldn't do the video stuff as well,
and
the interfaces that work on the desktop aren't what'll work in the
livingrooms of most people. Go play with a TiVo, then call that
"VIC-20".
We're extending the logical evolution of that concept to "Amiga 1200"
or
so, metaphorically speaking.
6)Who would buy a PPC machine today ?
People who want PPC-Linux.
People who want AmigaOS 4.0/MorphOS/etc.
People who don't care what's inside, but what it does.
Other stuff I won't get into now.
7)Will it ship with several OSes? What's the best OS for that
machine?
Isn't the Tsunami machine a Amiga machine ?
There will be multiple OSs available. What ships is really a marketing
question, not an engineering one. Best is what's best for you, when
it's
your computer.
As for Amiga, what's "an Amiga machine"? If it's "something based on a
hack of the classic 68K Amiga achitecture", then no, we're not doing
that. If it's "something planned to be able to run new AmigaOS, and
other things", then sure.
8)How can you ship it with BeOS as there's no more BeOS PPC support
from Be Inc. ?
Some of the marketing information on the web site is old. We're working
on that.
9)Where could we find a real photo of the machine ?
c-FMR : Not answered
10)Are the Pilot Systems ready for shipping ?
Prototypes exist today. Pilot production is next, that's being
organized
now. The goal was to have systems at the Koln Amiga show, but I heard
that's cancelled, or perhaps merged with another show in November. So
you'll see them then. Or before then, perhaps, if we have time for the
Web site.
11)What is a realistic shipping date of the machine?
I don't know about that. The stuff I'm directly working on will ship in
2002. The bPlan stuff, certainly sooner than the STB.
12)Will it ship to any country?
Logically, if it ships at all, it's shipping to a country. Will it ship
to _every_ country? I don't know. Merlancia certainly doesn't have
representation in every country. That doesn't mean we won't have
partners to extend beyond.
13)How many third party developpers are working on that machine?
That's an OS-dependent question, really. We're the hardware vendor, not
the OS vendor.
14)Are you in touch with Apple about licencing MacOS?
No. If we did MacOS, it wouldn't be with Apple's help. They don't want
other people running MacOS. They can't necessarily stop it. On the
other
hand, Linux is already bigger than MacOS. There's no guarantee that
MacOS is going to remain interesting enought to consider, though I have
no problems with anyone putting any OS on our hardware.
15)Why should a MacOS user choose a Tsunami instead of a Apple
machine?
Assuming we do have MacOS, it would be, simply, because it's a better
deal than something from Apple. Or because, perhaps, you want an open
system. Apple won't allow other OSs to run on their machines, for the
most part. We will.
16)Isn't the Tsunami a bit expensive?
Some of the systems are high-end, with fancy casework and other things
that drive up the price. Certainly not for everyone.
17)Will the Tsunami machine ship with multiboot software?
Ultimately. Things won't necessarily happen all at once.
Can we expect some kind of VMWare software to run several OSes at the
same time ? Not at first, other than the obvious bits, that classic AmigaOS,
AmigaDE
if we support it, etc. can run in any OS environment.
1)I supposed that you're aware that Be Inc. is looking for one or
more
new investors ?
Palm, Inc. is acquiring them, for about $10 million in stock.
2)Do you believe such as Be people that BeOS is almost a dead OS?
BeOS is technically the best OS of the last 10 years. But without some
kind of support, it's not going to evolve. A thing either evolves, or
it
dies, over time. We'll have to see what Palm thinks about the Be
future.
In your opinion, what was wrong with Be Inc. strategy ?
I could write a book. They had a few major problems. For one, they took
kind of a gamble on the geek/media market, as being large enough to
support them. It could have been, but the time necessary to establish
their SW platform was longer than they could deal with.
The huge mistake was, at the time of the BeIA annoucement, also
officially dropping support for desktop BeOS. They were basically
cancelling one bet, before the horses even got to the last
straightaway,
and put their remaining money on a longshot.
See, BeOS was, I believe,
on the verge of becoming very well supported for high-end DAW (digital
audio workstation) work. I was ready to switch my studio computer to
BeOS fulltime, buy Nuendo at nearly any price, buy new audio/MIDI
hardware if mine didn't get BeOS support. At least half of the
interesting audio carc companies were supporting BeOS, with drivers. I
suspect many of the software companies who weren't yet supporting BeOS
were watching intently.
The "we're dropping emphasis on the desktop", or whatever they said,
doomed all this. The big boys left, practically overnight. And there
was
NO need to do this. Saying "we intend to support applicances" doesn't
have to be predicated on "we're killing the desktop". In fact, they
could have made a good show about how the OS modularity made them two
sides of the same coin.
The problem with BeIA is simple: x86 is a crappy embedded platform.
Sure,
you can fool a fe PClone companies into using underpowered
National/Cyrix chips, or expensive laptop Celerons/Durons, but it's not
what you need to build an appliance. Look at Compaq's various ventures.
They want $800 for a thing that just plays CDs and rips MP3 to an
internal hard disc. The "Phoenix" I worked on at Metabox AG can do
this.
And browse the web. And play games. And play full motion MPEG-2 videos.
And deliver digital TV, and record shows for you when you're not
around.
And all for a bit less money. And as well, there's a certain bit of source code support you will
demand in an embedded market, since your hardware may well be changing
during development. I want the stuff that hooks the OS into my system
to
be under my control, not Be's. If I run QNX, I get a couple of
documented modules, with source code, that hook the OS to my hardware.
If I run some of the other embedded OSs, I get full source code. I
don't
know the BeIA deal, but Be totally screwed the pooch on this, for the
desktop. I even busted their chips, at serveral DevCons, about their
not
having a HAL (Hardware Abstraction Layer). Even Windows NT/2K/XP has a
HAL. I think they had too many high-level software guys, no one
thinking
about the low-level stuff. You could almost get away with that on the
desktop. You can't, for embedded stuff.
Thankfully, some of this is just policy. Under Palm, they could very
well extend the reach of BeOS/BeIA. I would use onumy set-top-box at
Merlancia, without reservation, if they fixed the CPU/source support
problems.
3)Could BeIA save Be Inc and BeOS?
Nope. And I told you why, above. BeIA would have been a nice adjunct to
normal BeOS. It's not a workable replacement.
Is there a real market for Internet Appliances?
Not the way most companies have built them. If you had to pay $200 for
a
web terminal, and $150 for a DTV receiver, and $200 for a DVD player,
and $800 for an MP3-CD jukebox, and $150 for a home gateway, etc. and
so
forth, you'd be happy to spend $600 or whatever for a device that does
it all, and more. There's a kind of basic, can't be any cheaper price
for hardware, based solely on volume, casework, power supply, drives,
and basically, all kinds of stuff in common with any device, regardless
of whether it's a PC or a web tablet. Then you add in the price of the
motherboard and the software, and you have a product. Thing is, adding
a
second function to that product is likely to be software and little to
no hardware. Next function, same thing. I think it of it as a non-PClone personal computer optimized for the
livingroom, for home entertainment. Users don't realize there's a whole
computer in there, any more than they realize their DVD or CD player
is,
in essence, a computer, just one with fixed functionality.
4)Could Merlancia Industries purchase BeOS?
Nope. Too much money, and, well, Palm has it already.
5)Do you think that you could get a BeOS licence from Be Inc and
continue
BeOS PPC developpement on your own ?
I will be interested in the kinds of things Palm plans to do with BeOS.
I'm not short an OS, I have too many, but then again, there's a bit of
work I'm willing to do to get "ideal" rather than "ok", for any
application.
6)Do you believe that Be Inc is of any interest to any big business
company? Who could purchase Be Inc ? Sony, AOL, Apple, Samsung... Microsoft ?
Palm, Inc.
7)What else would you like to tell our readers about BeOS and Be Inc
?
I pretty much have. Overall, BeOS was to the mid-90s, tech-wise, what
AmigaOS was to the mid-80s. As much as people still see AmigaOS doing
things better, or more cleverly, in AmigaOS, compared to where Windows
has gone, BeOS is demonstrating where Windows probably never will be
able to go. It has the Amiga feel of "simple, clever, and functional",
rather than "complex, convoluted, bloated, and weak" you too often find
in Windows.
As well, BeOS is extremely easy to code for. Easier than Amiga,
dramatically easier than the convoluted web of APIs, subsystems,
library
frameworks, language interfaces, thunks, and other garbage you find in
Windows.
That's why Metabox approached Be, in 1988, to use BeOS in our
second-generation set top box, the Metabox 500, which was, oddly
enough,
x86 based (it wasn't ideal, we knew it, but it did make development
very
fast, adapting a PC design rather than starting from scratch). Back
then,
Be didn't have clue about applicance computing. So we wound up using
OS/2. They learned, but still managed to make that learning become the
wrong answer.
--
Dave Haynie d.haynie@merlancia.com
http://www.merlancia.com
Chief Technology Officer, Merlancia Industries