@umisef

I guess the next question is, how hard would it be for them to entirely fake all the output?

_________________
Pegasos2 G3 running AOS 4.1 and MorphOS 2.0
Amikit user, tinkering with Icaros VM (AROS)
EFIKA owner
Amiga 1200

Quote:

I guess the next question is, how hard would it be for them to entirely fake all the output?

How hard ?

Quote:

If (cpuID == PASemi) printf("you'll never know he he ;)\n");

_________________
http://www.warpdesign.fr/

Quote:

I guess the next question is, how hard would it be for them to entirely fake all the output?

How hard ?

Quote:

If (cpuID == 0x5000) printf("you'll never know he he ;)\n");

_________________
http://www.warpdesign.fr/

@umisef

Or maybe the coders by mistake put the exe in the wrong place (like C:) so the original one in SYS:Tools (or wherever AOS4 has it) was still there. Now coders will typically run ShowConfig from a shell which would load it from C: So they might not have noticed that the original one is still there. But users would typically start it from WB by double clicking the ShowConfig icon in SYS:Tools which will start SYS:Tools/ShowConfig, ie. the "wrong" one.

Does "showconfig debug" work with os 4.1.2?

@vidarh

Quote:

Fair enough. Though I did ask for something "reasonably realistic" :)

Well, if you want to do something revolutionary, you must be willing to ignore your past. AmigaOS is very much an 80s OS, and *anything* you could derive from it, or turn it into, if still being recognisably AmigaOS will simply have no chance of being revolutionary, the way AmigaOS was back in 1985, in its way.

So, let's say rather than "take forward AmigaOS" or something like that, you have some (a lot of) money, time, clout and dedication, and you want to create something that is to the current crop of "mainstream" stuff as AmigaOS was to the mainstream stuff of the mid 80ies....
Then I'd suggest looking into abandoning the concept of files, file I/O, file systems and anything else that differentiates between "in memory" and "on storage" data. Instead, simply treat hard drives as the last level in the (64 bit) memory hierarchy; Any data, and any programs you have will always be "in memory" (even if only some of it will be "cached" in RAM). In order to keep track of things, and be able to find them, you probably want to provide OS services to deal with sharable in-memory objects (such as describing them, indexing them, access control both for security as well as for concurrency, etc). However, chances are that with everything always in memory, you might be able to reduce the "kernel" equivalent to something that merely handles the very basics (i.e. the memory hierarchy, process scheduling, and inter process communication), and move everything else into non-kernel objects --- which would provide some AmigaOS spirit :)

Now, this is obviously neither a new idea, nor is it an easy one to actually make work, let alone make work well. There's a myriad of details to get wrong, and if it isn't done quite right, it's not going to provide the benefits. It's a huge job to get *anything* like that off the ground, unfortunately, which is almost certainly beyond the means of anyone but some very big corporations.
What's worse, this being 2010, making a revolutionary new OS is not even half the job. Users don't want an OS, they want to interact with their computer (so need a user interface), and want to do stuff (so need applications). And once you do away with one of the major concepts of all current mainstream OSs, tapping into the efforts of open source by "porting" becomes very, very hard.

So no, given the tiny amounts of money which exist in the current Amiga market, it is extremely unlikely that a paradigm-changing development like that would come from this source; Nobody is going to sink tens of millions of dollars (if not hundreds) into it. But *if* someone were to do so, they could, in a few years, have a few years head start on everyone else :) (Or they could find, in a few years, that the whole thing was a bad idea. That's the nature of speculative risk investment :)

@fairlanefastback

Quote:

I guess the next question is, how hard would it be for them to entirely fake all the output?

Fairly trivial, of course.

However, there are a few reasons to strongly doubt that's what happened:

(a) It has been stated that the command line ShowConfig suppressed all PCI output. Why would anyone go to the effort of creating two distinct doctored ShowConfig executables (or two distinct doctorings of the one executable)? Spending precious developer time on hiding the CPU's identity is understandable. Wasting it on frivolously doing it twice, not so much.

(b) Outputting something made-up is easy. Outputting something realistic-looking is much harder --- and the PA6T list shown on that screenshot is quite believable. AFAICT, there is the correct number of all the devices. Why go to such lengths?

(c) Related to (b): If all you want to do is hide your CPU (and be a smug smartalec about it), why would you put in data for a CPU that *could* be in the box? It doesn't give anyone the "ha, I thought I were clever, but you are obviously cleverer" feeling. Instead, it gets people to believe in the faked CPU, and they may either (i) like it, at which point they'll get upset when they find out it was all a fake, or (ii) not like it, at which point they might just decide to give their money to someone else before finding out the truth about your product. Neither of which is good for you. So instead, you would rather put something like "0x1057 0x5806" in there... That way, people will know they were outsmarted.


And, of course, HyperionMP had the chance to say "Ha! Suckers! You all fell for it! Aren't we clever-sneaky :)", and quite pointedly avoided doing so.

@-Sam-

Quote:

This guy has seen every NDA ...ever. In between being profitable and doing his wife.

That, my friend, was absolutely hilarious!!!!!



Rude!

_________________
2017 Camaro 2SS

@umisef

You could still make Amiga OS 4.x truly revolutionary by porting it to Quantum computers.

_________________
Sent from my Quantum Computer.

@umisef

Quote:

Rob has stated that he has a photo which shows the 12th PCI entry as "working" rather than "defective"

Here it is.

http://i49.tinypic.com/2niqsk5.jpg

@Georg

Some filesystems have recycled dir (i.e. SFS and PFS) and if old executable was just replaced it could be possible someone just found old binary from there... it is also possible someone just plugged in his usb memstick and ran showconfig from there. Or found an install cd somewhere...

Many many years ago when there was an Amiga event in Finland I saw how one guy wrote small app on a demo machine. If you leave machines unattended anything can happen..

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@DiscreetFX

Quote:

You could still make Amiga OS 4.x truly revolutionary by porting it to Quantum computers.

Again, I'll give you the benefit of the doubt, and assume that you were making a joke...

Ive been working on th idea for years. .theoretically, my proof of concept quantum logic gate can be built using a PC with sound card, speakers, but preferably more than one mic.. and a Large cardboard box.

Unfortunately, for the wavelengths and frequencies involved, I cant afford a box that big.. at least the triple layer thing used to carry commercial photocopiers. 8(

Why do I say my quantum CPU idea for Amiga works so easilym, and others dont?.. Everyone else is working on quantum particle forms.. Im using Waveforms. ..you know.. the ones you can DVBT send 8000 simultaneous carriers through the logic?

Theres lots of fun things out there.. but teh researchers dont talk to each other. .theyre doing paair analysis to understand music, speech, text. .which in unified symbol coding. .just means theyre repeating what Collosus does. .theyre breaking the code of reality. So.. work on teh code directly.. And you find out how to understand an awful lot more of reality than you realised.

_________________
The older and more respected a scientist is, the longer it takes to prove him wrong.

@all


http://www.linuxfordevices.com/c/a/News/Freescale-QorIQ-P3014-P5020-P5010/?kc=rss

@umisef

Quote:

umisef wrote: @xeron Quote: The ShowConfig tool was completely replaced with one that just printed a message. I am not convinced.

I'm not guessing, i'm telling you that I know for a fact that ShowConfig, at least the binary in its usual location, was replaced with one that printed a one-line message and exited.

If someone found a copy of the original binary somewhere else, i do not know.

_________________
Playstation Network ID: xeron6

@amitv

VERY interesting find!

It could very well be the freescale e5500 and Varisys is one of thier partners!

Looks pretty sweet if it is! Makes a bit more sense to me because it JUST got announced!



But its still speculation until the official word.

@amitv

Here's a better link.

Clearly PPC is not dead yet, that thing is rated faster per clock than a G5.


However there is something very interesting about the specs of the p5020, It looks remarkably like an enhanced PA-Semi chip.

There's clearly changes, slightly different ports some bits faster as well, less Serdes but they're twice the speed. The CPU cores appear to have changed a bit as well, they've got their own small L2s and the pipeline length has been cut in 2. Looks like an effort to boost control plane (i.e. integer) performance.

However, P5020 isn't due until 2011, but given Freescale's record that means 2012. Last time Freescale did a 64bit PPC (The PPC620) it was 4 *years* late and then only shipped in very limited quantities.


My guess is X1000 will ship with a Freescale branded CPU that is basically a rebadged PA-Semi chip. Then, when it's ready there'll be an upgrade to the p5020 (assuming A-eon can actually sell enough to keep going that long).

_________________
Whyzzat?

@koft

Quote:

You wrote: It's not standard for the industry to do this.... NDA saying you can't tell potential customers what part you are using in your product, nah. Nobody does that. That's absurd man...

So are you saying it's normal practice or 'business as usual' for hardware designers (specifically CPU or SoC developers) not to require their partners (regardless of size) to hide details about the future use of their (SoC producer) new design in upcoming product releases?

Do I understand correctly that you're saying it's routine to have NDAs which don't include clauses which limit the information which can be given out during the development phase of any product using newly released SoC/CPUs?

Are you also saying that IF any NDA actually should happen to (senselessly, according to you) include such a limitation, it would be written in said NDA as a specific, set time and date, after which the clause no longer ceases to be valid, and can be treated as non-existant, while the rest of the NDA would still be held to be current and true and indeed still as restrictive as it was before, with the exception of being able to reveal that small bit of information, which the targeted market is so keen to hear?

Well... I guess PA Semi and A-Eon must be a right bunch of ####s then! What do you say? Who could imagine ANYONE ever using a clause such as "won't reveal details until first production run is complete", so if dates slip, the NDA will still protect. Who could imagine a company including such clauses in their NDA because they wouldn't want a bunch of 'startups' announcing upcoming products, but never actually relasing them, and the SoC developer becomes associated with 'vapourware' announcements. Why would they care about such nonsense? Surely there couldn't possibly be people in the community that would actually keep a list of vapourware and continually bring it up, smearing the names of anyone involved or attached to that list? Couldn't be.. could it?

Oh and Trevor mentions that early in 2008, they got a CPU partner. When was PA Semi sold to Apple? Would the possible sale of all IP to Apple POSSIBLY have cause PA Semi to be extra careful with the wording and restrictions used in their NDAs at the time? Was there a period of negotiation, do you think - in your professional opinion, during which PA Semi had to continue to grow their business with new partnerships, and yet try to move forward with these negotiations, privately, all the time ensuring that their existing commitments would be met, whatever the outcome of such negotiations? Or does everyone run a business in exactly the same way because everyone has exactly the same set of circumstances in which they are running said business? All of this is assuming of course that PA Semi is the mysterious partner. Without knowing the partner, you can't assume to now the NDA clauses.

Give us a break. Conspiracy theories are fun... but only for a few minutes, or until you actually use your brain and think things through.

tiffers

Last edited by tiffers on 23-Jun-2010 at 05:29 AM.

@umisef

Why not 0x8086 ids - (0x8086 0x0960 comes to mind - It'd drive people nuts wondering how they'd source an old microcontroller and why they'd want to use it in new product, and how they were running a full OS on it.)? Or, why not some random fake ids? That'd be cool, and it'd keep everyone guessing!

Are you proposing this was *allowed* to stay in

@tiffers

Quote:

tiffers wrote: @koft Quote: You wrote: It's not standard for the industry to do this.... NDA saying you can't tell potential customers what part you are using in your product, nah. Nobody does that. That's absurd man... So are you saying it's normal practice or 'business as usual' for hardware designers (specifically CPU or SoC developers) not to require their partners (regardless of size) to hide details about the future use of their (SoC producer) new design in upcoming product releases? Do I understand correctly that you're saying it's routine to have NDAs which don't include clauses which limit the information which can be given out during the development phase of any product using newly released SoC/CPUs? Are you also saying that IF any NDA actually should happen to (senselessly, according to you) include such a limitation, it would be written in said NDA as a specific, set time and date, after which the clause no longer ceases to be valid, and can be treated as non-existant, while the rest of the NDA would still be held to be current and true and indeed still as restrictive as it was before, with the exception of being able to reveal that small bit of information, which the targeted market is so keen to hear? Well... I guess PA Semi and A-Eon must be a right bunch of ####s then! What do you say? Who could imagine ANYONE ever using a clause such as "won't reveal details until first production run is complete", so if dates slip, the NDA will still protect. Who could imagine a company including such clauses in their NDA because they wouldn't want a bunch of 'startups' announcing upcoming products, but never actually relasing them, and the SoC developer becomes associated with 'vapourware' announcements. Why would they care about such nonsense? Surely there couldn't possibly be people in the community that would actually keep a list of vapourware and continually bring it up, smearing the names of anyone involved or attached to that list? Couldn't be.. could it? Oh and Trevor mentions that early in 2008, they got a CPU partner. When was PA Semi sold to Apple? Would the possible sale of all IP to Apple POSSIBLY have cause PA Semi to be extra careful with the wording and restrictions used in their NDAs at the time? Was there a period of negotiation, do you think - in your professional opinion, during which PA Semi had to continue to grow their business with new partnerships, and yet try to move forward with these negotiations, privately, all the time ensuring that their existing commitments would be met, whatever the outcome of such negotiations? Or does everyone run a business in exactly the same way because everyone has exactly the same set of circumstances in which they are running said business? All of this is assuming of course that PA Semi is the mysterious partner. Without knowing the partner, you can't assume to now the NDA clauses. Give us a break. Conspiracy theories are fun... but only for a few minutes, or until you actually use your brain and think things through. tiffers

Wow, that's a wall of text. TL;DR comes to mind... You're over reacting to the points I have made. Optimally, A-Eon would be shipping product right now. They originally claimed that we'd have this before summer, guess what, it's summer. A-Eon claims to have to be silent about the CPU because of an NDA agreement, but it's already past the date of their most optimal release schedule. The only way to fit their claim to something logical is if the NDA restricted them from talking about the "secret part" until *after* they were shipping units. That's anal to some absurd degree and it makes no sense whatsoever.

NDA restrictions for price lists and access to technical literature is understandable, though annoying. I've even seen NDAs regarding parts that were *about* to be publicly released with samples being delivered before hand to engineers. Never once have I seen an NDA which restricted talking about the part until the final end product was scheduled to hit customers doors. That's absurd. A-Eon isn't talking about the CPU because they choose not to, and it's part of their bizarre marketing strategy.

In my opinion it's a shame they've chosen to go this route in the PR department. We've seen endless waves of holy wars because of it. I'm glad though, I've gotten sick of all the BS. I picked up an emac earlier today and installed MorphOS and was thoroughly impressed to be honest. Half the power of an A1X1K for a grand total of 256 dollars. And man is that sweet that the cost of the machine + MOS license happened to be a power of 2.

_________________