Theo de Raadt says Pegasos Bad

[magnetic]

Ok, I was going to stay out of this thread but I got dragged into it. (why was my name mentioned anyway???)

 Firstly, Theo had a Pegasos 2 for a couple of weeks. When bPlan updated the firmware there were problems with the Openbsd build (which was working fine on Peg1 and peg2 OF 1.0)  This would have been a quick matter for Theo to work around but he decided not to do so, making a stance that he wouldnt work on "closed hardware"

 One of the most exciting features of the Pegasos2 is the onboard Gige. This is much better and faster than pci gige. The unfortunate thing is that Genesi has NDAs with Marvell. ATM Genesi/bplan has the only full PPC computer in the world running with that bridge, so combined with the Marvell NDA it could be considered IP.  

 AFAIK they could not release the full documentation of the Northbridge to Theo due to a varity of legal and business decisions. This obviously was unacceptable to the OpenBSD devs, because they want everything totally "open" Its the classic argument of Closed Source vs Open Source.

 Genesi in fact was excited to work with Theo and the OpenBSD guys but could not at that time give out that information. Even though Theo is a well known coding figure, his "antics" are known to alot of people in the industry. If he had exhibited a little more patience he may have had a better relationship with Genesi. As it was they promoted OpenBSD to many industrial clients and at trade shows.

 Anyone who is running a Pegasos2 will tell you its very stable and works well. I have been running a Pegasos 2 G4 with the 2.6.4 PPC Linux kernel and Debian with KDE 3.2.1 and Gnome 2.4.2 and it hasnt crashed once. I've had uptime of over 10 days. Since moving away from Articia the Pegasos is now an industrial class piece of hardware.

 MorphOS makes strides daily and more native PPC programs are coded and/or ported all the time. The community is growing rapidly and the expectations are high. If you look into Theo's record of problems with companies, and yes even his own fellow developers you will see some scope on this. I do not know this individual but will give him the respect that he is an elite coder. However, his politics and attitude are questionable to some.

magnetic

[Floid]

Batman wrote:
@jope
First, Theo is not a troll, he's the boss of the OpenBSD project. Read his website: http://www.theos.com/ :-D

Second, Theo is questioning only about Genesi's "customer care". Basically he's alerting everyone to not make his very experience with BBRV. Well, he has a somewhat flaming way of express himself, but when he talks, he talks with reason. Why don't ask him, via misc-openbsd mailing list, the same questions you asked me? I'm sure he will have some interesting answer.. ;-)

If there was an experience with BBRV... Hoo-boy, especially as all the visible efforts really did look positive on that front.

Basic rules of the game:
-If you want to work with OpenBSD, you can't blow smoke up Theo's ___.
-If you aren't blowing smoke, Theo is generally "the voice of reason," but he's human, and can sometimes get the wrong idea like anyone else.  (Who gave him the idea that Genesi was representing itself as a huge edifice, vs. a muddling startup possibly trying to do right?)
-If you don't have the courtesy to stand up and correct him before he enacts bad policy, he's probably not going to want to work with you anyway.  (The McEwen 'people who piss me off' excuse -- But if you don't have Theo's best interest in mind, you don't have the project's best interest in mind, and it's time to find another project.)
-If it comes down to an unrecoverable disagreement, you can always fork.  However, the 'value' of OpenBSD is in the auditing, so forks usually drop to the attractiveness of NetBSD_without_the_scalability.

If the Barbie were still going and had the same exact problems (including customer support, if applicable), I doubt relations there would've taken a nosedive so quickly, and as far as I know, Theo and friends don't have a problem with expensive 486 boards.  So what happened? ... (For those who can claim authority:  Don't tell me, tell misc@, if there's a chance it's worth salvaging.  But if you can't be 'Open' about it, they'll be happy to hand you plenty of rope.)

Third, OpenBSD's policy is always to support hardware makers that give full and free access to the documentation (Read: no NDA). Look backward in that mailing list for a thread about SPARC documentation for OpenBSD. :-D

Policy can/has been simplified to "Free as in build a suitcase nuke out of it."

---

Being the first 'scene' platform into a major project's CVS was a great bragging right and could remain so... I'd suggest not losing it, or at least, not going down in memory as only wasting the project's time. :-)

[Karlos]

:lol:

What a bizzare thread!

[hirschma]

Jonathan

itix wrote:

I tried but could not find that. Could you provide a link?

First, here's Theo's previous messages, dated 2/20/04 (sorry for the raw language...):

"At this point, I would recommend against anyone buying a piece of hardware from the Pegasos people because their firmware is SO BUSTED that it makes Apple roms look like hot {bleep}.

If you want a powerpc machine, I must relunctantly recommend Apple hardware, since the alternatives are impossible to use.

ps. And their tech support people have proved to be intensely arrogant about the bugs in their software..."

You should be able to find it on the Sigmasoft archives, but they're down right now

I'll also make note that there are many reasons to want to run OpenBSD on non-x86 platforms, and that Pegasos would be a great solution if the project's leader didn't condemn it.

Jonathan

[JKD]

Well, I guess to those of with experience of the Pegasos platform this looks like nothing but pure flamebait...so lets just review:

1. Pegasos 1 had (and still has) some serious problems...there is a lot of information in the public domain about this and it's no secret about the painful Artica S, April I, April II evolution...The Peg I could not be made to work reliably with a G4.

2. 600MHz with a 100/133MHz FSB is slow compared with latest PCs - no arguement with that,  but it does depend on application. People use old Pentium 200 Mhz boxes for firewalls, mail servers etc. running linux and it's certainly way more capable than that.

3. Only a few G4 machines in USA...well pardon me if I don;t fall over in surprise. Again, there is plenty of public domain information stated by the head honchos that they are struggling in the transition from bleeding money into the platform into making it a profitable entity. Sales aren't what anyone expected (same applies to A1 for example)...but I do see progress, getting assosciated with IBM, visibility in Motorola for example. It's also publically stated that production is occurring in small batches (low 100s) for this very reason. A G4 Pegasos II costs $600 to $700..this is no small investment (but see point 5)

4. Open firmware could be bugged for all I know - I'm not qualified to know. I've seen other people discussing issues regarding device tree, CPU enumeration. I assume they will be fixed in time.

5. Cost - it's very competetive for a PPC platform. I don't think he has a point here unless you want to open the old chestnut of comparing Apples with Oranges i.e. PPC vs x86

6. Genesi as a working business entity - again there's plenty of public domain knowledge regarding their financial situation and we all know it's not good - even top management posted this information! (pretty unusual IMO)
As to their business practice _ I really don't have any inside information or knowledge. If Theo does then I might consider myself forewarned, however:

7. The tone of the post is pure flamery...irrespective of how competent the person is (and I surely don't know him from Adam) This just reduces the impact of the statement.

So it looks like some of his points are way of base and on a few there is still work to be done...

Sound fair?

Steve

[JKD]

Sounded like it just didn't work out for his projects and there were communication issues.

He even knew the whol project (Pegasos BSD) was at very early beta stage.

Those are the facts.

Do I need to list the reams of hardware on the PC platform that may or may not be supported on various kernels?
 
If Genesi had the opportunity to get Gig ethernet working and missed it due to not paying attention to their email then it's a shame - developer realtions are key.

If, however, the mails merely bounced then not sure what to make of it. I have had one bounce from genesi.lu and none from pegasos-usa.com in over a year.

[CodeSmith]

The one thing people in this thread don't realize is that Theo (and anyone else outside of the amiga "community") is used to having a certain minimum quality, support and low price on a motherboard.  Theo can walk into any shop in the world and buy a computer that has much better price/performance and accessibility than either the AmigaOne or either model of the Pegasos, and not only x86 - a Mac easily beats even a Pegasos on price.  That's just the way things are out in "the real world".  What happens when he complains about it?  a little red vs blue fight erupts, and he's accused of "impatience" and "not trying hard enough".

Guess what people - in the 21st century, people don't have to "be patient" or "try harder" when they buy a computer.  My Windows XP box is a *lot* better at that sort of thing than my AmigaOne, and I expect it to keep on being that even after I get AmigaOS 4 for it.  Sure, OS4 will be leaner and probably more fun to program (I expect similar things about morphos), but the degree to which users are expected to "grin and bear it" is from back in the computing dark ages.  I have put up with a *lot* of crap, just because it's an amiga; I've read both "reds" and "blues" with similar stories.  If I'd had to put up with  only one tenth of that when I bought a PC, that box would have been taken back to the shop in an instant.  Things have evolved over the past decade, out "in the real world".  Theo's not the one with the problem, the amiga "community" is.

[mikeymike]

Ooh, suspicious single post poster with possible anti-Genesi material.

More than one, seemingly.

[DethKnight]

elite programmer huh

then what about this

Known Problems

    * OpenBSD has a problem finding itself ......

from links within his own webpage

[JKD]

I take exception to your first paragraph. I (personally) absolutely *do* realize this and I'm sure your generaliztion sucks for other too.

With regards to your second paragraph....unfortunately that's the reality of a  non-mainstream platform. If peoples expectations are higher than that nothing either of us can say or do is going to close that gap.

It's a *personal* choice to go against the flow...this applies even one of the biggest 'marginal' markets - Mac OS. Marginalization issues abound even there.

Lets not forget that Windows didn't even become good until Win2k (or NT4 if you were brave enough...) The amount of crap the user had to put up with on version prior to that was comparable IMO.

B-, must try harder! :-D

Steve

[cecilia]

Karlos wrote:
:lol:

What a bizzare thread!

yup!!!
and i'd like to be helpful here, so, if anyone needs some creative cuss-words to fling out at others look here for help:

SwearMaster 3000

isn't technology wonderful?
:roflmao:

[Karlos]

[homer]

It's Amazing. I can't believe some guy I never heard of is slagging of a plaform I know virtually nothing about...

[/homer]

[takemehomegrandma]

(Someone actually registered a new account here on amiga.org just to
bring us this "news item". I find this somewhat interesting. And
touching. An "exodus individual" that has come back home perhaps? ;-))

OpenBSD on Pegasos makes sense in lots of ways. The main goal
(AFAIK) with OpenBSD is security, and OpenBSD would certainly find
itself useful on this PPC platform.

I have a hard time understanding his issues here. As an OS programmer
he should not be unfamiliar with the fact that different firmware has
different characteristics, and certainly he must have seen changes
made in firmware before? What's the fuss about?

As for availability, sure, it's a new platform and the production has
unfortunately not been able to meet the high demand. This has been
annoying to many "Pegasos eager" people (as we have seen here on these
community forums as well), but at least *he* (they?) got their boards,
and the production volume is increasing! So what's the fuss about?

As for the unreachable documentation for the Gigabit Ethernet, which
is one of the main features of the Pegasos board, this could certainly
be frustrating, but hey, it's only one feature after all. Frustrating,
 yes, but it should not be a reason to get so upset on a personal
level to such a degree to write a public letter in this tone! And why
would a compiled binary file (for that feature only) be so bad?
"More closed than PC", yeah right. Why so impatient?

As for the stability - here he is totally up in the blue IMO. I have
two Pegasos II motherboards, one G3 and one G4, I have been running
the G4 CPU at 100% *non-stop* for three weeks, and not a single
problem has arised. I have heard no reports from other users that
contradict this (I have not heard any reports from Pegasos OpenBSD
users though). The current marvell based Pegasos motherboards runs
much more stable than some x86 motherboards I have owned. The Pegasos
is *rock stable* since it moved away from the Articia. If his OS got
unstable with the firmware update, why didn't he simply correct the
problems instead of writing such a post? BTW, for reference it would
be quite interesting to send him some certain other Articia based
motherboard and let him test some stability ... ;-)

As for Genesi's finances, corporate structure, and the price of the
motherboard, heck, what does that has to do with anything? He is an OS
programmer! BTW, my latest PC (motherboard+CPU) was more expensive
than my Pegasos (but a lot cheaper than some other PPC solution).
Sure, that PC is more powerful, but it's also hot and noisy as hell
(besides being more expensive), and it's not PPC. It would IMHO be
better if he simply coded his OS instead of taking on to a pricerunner
role. Let the customers decide what they want. Ultimate horsepower
and heat is not allways needed (nor wanted) in every single hardware
application. And it's not like this current price is written in stone
anyway, this price is AFAIK set according to current production
volumes.

OpenBSD could be great for many purposes, and that OS in combination
with the Pegasos hardware makes sense for many applications. I hope
that Genesi will care to try to find a way of resolving this, despite
this blatant public letter. However, an eventual discontinued official
support from some OpenBSD developer will definitely not be a show
stopper in any way for the Pegasos platform IMO ...

[Acill]

Well since the Peg is made to run MorphOS as its native system I dont see the big deal. If they want it on the Peg they need to program to its specs. I for one just ordered a new Peg II G4 and cant wait for it to come! Just thinking of running all my Amiga apps on a 1GHZ CPU is awsome! I dont plan to put linux or BSD on it at all. Its simply a replacement for my old A3000 and nothing more. If I can use itas a full time computer as I did on my Amiga its good for me!

[JKD]

Likewise here (waiting for a Peg II G4) and also sold my most beloved Amiga to get this opportunity...a machine I probably sunk over $5k into over the years.... j'ai regret rien..or something along those lines.

As far as other operating systems go.. I never tried BSD but I certainly will get round to installing and using linux this time (never did on my Peg I). I will be showing it off and demoing it to anyone who's interested (and a few who probably aren't :-D)

Linux and other OS'es present a more viable entry point to the Pegasos than MorphOS from a familiarity perspective of course.

Back on topoc = I thought BSD was integral to the ShopIP Guardian product...so this guy could be important, at least to that effort?

Steve

[hirschma]

takemehomegrandma wrote:
(Someone actually registered a new account here on amiga.org just to
bring us this "news item". I find this somewhat interesting. And
touching. An "exodus individual" that has come back home perhaps? ;-))

I guess I could be called an exodus individual - I ran a business on Amigas back in the '80s/'90s. Put myself though school doing graphics for local car dealership commercials.

I'd love to have a new Amiga platform - hell, I'm toying with the idea of recreating my old Amiga desktop - but the Pegasos was particularly interesting to me since I'm a big OpenBSD supporter - and Genesi identifies it as one of their major OS's.

So, when someone that I greatly respect, and who is widely respected, for his OS work, I figured it was newsworthy. Yes, Theo De Raadt may be consider an a-hole by a lot of folks, but he's smart, he's dealt with all the weirdness of the PC world, he's accomplished a lot - and when he makes such charges, I think that we all should at least take note. He's earned that much, IMO.

I wouldn't read much more into it. I'm not affiliated with any of the pretenders to the Amiga throne. I just like cool computing.

Jonathan