The Mac Observer has posted a reader poll and commentary on the topic of whether Apple should embrace USB 2. Commentary at AppleLinks. In other Apple news, Apple is preparing on launching an online music service. Also, longtime hardware developer Steve Sakoman, a former Palm and Be executive, has rejoined Apple Computer as a vice president. Mr Sakoman was involved in the Mac II development and was the main guy behind the creation of the Newton. At Be, he designed the first generation H0bbit-based BeBox, the one that had 5 AT&T CPUs, back in the early days of Be. He later left and when he came back became Be’s CTO.
Ex-Be/Palm Executive Rejoins Apple; Should Apple Embrace USB 2.0?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
USB 2.0 is truly a sad, sickly technology. It’s inferior in every possible measure to IEEE1394 (after accounting for protocol overhead), let alone Firewire 800.
Furthermore, IEEE1394 is set to become the backbone of digital media.
It’s really sad when a wholely inferior solution being pushed on the public to recoup R&D costs wins simply because it’s backed by the leading commodity vendor.
I suppose you could blame Apple for charging licensing fees for FireWire for as long as they did, but Intel should’ve just eaten their existing R&D costs instead of trying to recoup their losses with yet another disparate standard which covers only part of the problem domain of an existing technology. Or perhaps Intel just came down with a case of NIH…
Unfortunately, that said, it’s here. There’s no reason for Apple not to adopt it at this point, except in protest of what a truly loathsome, inferior standard USB 2.0 is. In this case I wish they’d come to their senses, step off their moral high ground, and let their customers have the usability they would prefer to moral superiority…
You’re mistaking with LOTR are you ?
The machine had 2 hobbits and 3 DSP’s from Ti.
I find amusing the number of ex-Be people going back to Apple, when they where bad mouthing Apple when they where un the Be World.
—
http://islande.hirlimann.net
They are 5 CPUs there, 2 H0bbits and 3 DSPs indeed, *but* they were all from AT&T, not from TI. The DSPs were the AT&T 3210.
We have a H0bbit motherboard here at home and they are all AT&T… 😉
I am looking into the motherboard right now (1993 model)… A really rare item. 😉
but hey, who cares … Eugenia does your MB boots ? (mine does).
>I find amusing the number of ex-Be people going back to Apple, when they where bad mouthing Apple when they where un the Be World.
Sorry, this unfair. They are not that many who went back to Apple. Only a handful of them, and probably not the ones who were “bad mouthing” Apple. Personally, I have _never_ heard a Be engineer/employee bad mouthing Apple.
And if you want to know, a Be-related engineer (very Be-related
went to Apple for a couple of months, and then he couldn’t take it anymore and he now works at MS. Does that tells you anything about Be-Apple relationships? No, it doesn’t and you should not generalize in such a way. It just tells you that people are individuals and they are doing their choices as they see fit. I can tell you for a fact that they are not interested in religious wars, just in having a steady job. Like most people do.
Oh, I think they are. They are just very heavily optimized for specific cases and operations, but they are “CPUs”, IMO.
I’m sorry, they are… of a type.
From the DSP glossary:
” A special-purpose microprocessor designed to handle signal-processing applications very quickly.”
http://www.cnet.com/Resources/Info/Glossary/Terms/dsp.html
From the CPU glossary:
The CPU is your computer’s brain, taking requests from applications and then processing, or executing, actions, a.k.a. operations.
http://www.cnet.com/Resources/Info/Glossary/Terms/cpu.html
In a way, DSPs are also CPUs on some machines, while they are plain specialized pieces of chips on other machines, DEPENDING how they are used. CPUs are handling general requests from apps, and DSPs are handling signal-processing apps. Depending *how* the DSP is used, it can be a CPU, or it can just be a specialized item.
These are very vague terms, pretty much like the flamewars they had back in 1993 in the BBSes: “is win3.1 just an operating environment, or an operating system?”
You should not be flaming each other or vague and unclear definitions, where both can be right and wrong at the same time.
What was the functions of those DSPs on the early BeBox/Hobbit?
The point is not to know if USB 2 is better or not than Firewire (there are some pros and cons). The point is that USB2 is there. And it’s sad that Apple computers can’t take advantage of the bandwidth.
I have an Archos Recorder 20 in USB 2.0. And it’s VERY slow to transfer lots of files (not only music) on USB 1. Since the cost between integration of USB1 and USB2 is very small, I don’t see a reason other than political not to make the move. I’m sure all the USB soundcards, MIDI devices and other peripherals around would benefit from the move (read latency) and at a minimum cost.
I just had a google search. Texas Instruments many times refers to its DSPs as CPUs, while on other technical PDFs, it says that some of their DSPs are comprised from CPUs, among other things. So in essense, a DSP is a special-purpose CPU (most of the time at least, depends on the DSP).
Agreed with everything you said. It makes sense stategically and marketing-wise for Apple.
I don’t have firewire (nor money) but I was just curious about the amount of external hardware that uses FireWire (800). If there are a ton of USB2 ones, and almost nothing for FireWire I think that I can see the problem quite clear.
My suggestion to Apple would be to start to cooperate with other companies to produce more software/hardware for the Mac, maybe even buying in a little in some select companies (choose the one with the highest quality, otherwise the whole ordeal will tank). Satisfy demand and create a market.
You must realize that USB2.0 is far far far more older in the market than FW 800
Sorry for the silly question, but can DSP chip be treated as a kind of CPU where software is built in hardware (hard-wired or microprogrammed) instead of software stored in memory (eeprom, flash, etc) as it is in ‘pure’ CPU?
Andreas
In any ways there is a memory to store the program, even on your soundcard there is a ROM.
The CPU qualification IMO depends on how close to the others the DSP is (that is, either on the CPU bus, or on an ISA card), and the ease of modification of the program of the DSP (downloadable or fixed ROM).
The C in CPU stands for CENTRAL, is the DSP the central processor? The only thing a DSP and a G4/P4/Hobbit have in common is that they process. Only the G4/P4/Hobbit are central. So, i would say it had 2 CPUs, not 5.
Jesus Christ, does it really make much difference if you count the DSPs as CPUs or not? In any case, it had 5 processors – does not make much of a difference if they were DSPs or CPUs.
As for USB2, Apple will adopt it. They really dont have much choice.
And about ex-Be(/Apple) people (re)joining Apple or joining Microsoft – do we really care that much? It’s their life, they should be able to make the decisions not people on the net.
These had DSPs in them as well, and were never referred to as multiple CPU machines.
Could you upload a few pics of your H0bbit and the CPUs/DSPs?
I couldn’t agree more with Elver Loho’s comments above concerning the ex-Be employees. If Steve Sakoman is reading this, good luck with your move!
As for USB 2.0, the industry will continue developing technologies that will utilize it, and Apple may be forced by consumers to welcome it to their line of systems.
My only taste of USB 2.0 is with the Compaqs we have at work, and it is filled with bugs. Not sure if it’s HW or SW, but someone needs to investigate this at Compaq!
Concerning the DSP/CPU debate, again Elver Loho said it best.
Often DSP are called CPU by reference to the PU part of the name. Too many people think: CPU is doing computing, DSP too, so DSP is a CPU.
Wrong, DSP are not CPU. You will never see a DSP managing a system because there have no IRQ management or I/O capability like CPU. DSP are co-processor not central processor. There are enable to operate by themself. DSP need a CPU in order to work.
May be today you can see DSP on the same die as the CPU. You see a big chip with a DSP and a CPU but only the CPU part is able to manage its environnement not the DSP part. the DSP is the slave of the CPU.
Sakoman is a big loss to Palmsource. Was he not leading product development? This is concerning. Who will lead software dev. now?
Palm and and palmsource seem to be a bit of trouble. Anyone see that article on register.co.uk regarding Sony’s desire to purchase palmsource. Realistically, i’d say apple should purchase palmsource (and sun too) and use it to get into the handheld business.
Ownership by sony, i think, would be the kiss of death to any software firm.
The important thing is that, when Be was making those first BeBoxes, *they* considered it and spoke of it as five processors. And so, that’s how others came to speak of them in that way.
Apple should adopt USB 2, there’s no doubt about that. rajan, there aren’t many Firewire 800 devices ou yet, but tons of digital camcorders and hard drives and disk burning drives of all stripes with 400. Some disk burners have both Firewire and USB ports.
People, settle down. This bebox was built long ago. Think about when the thing was built. The cpu’s in it were not much powerful then some dumb chips on computers today. If I rember correctly steve built the thing by hand. Designed even the video and sound cards for it. The fact he could go to frys (I think it was frys) with JLG and buy some bits and go and built the thing is amazing. I’m with Eugenia on this far as CPU count. I thought it at first had just the 2 cpu’s and then later had the others added as daughter cards. Also thought at some point one of the prototypes had 7 cpus, but i’m not very sure. I think on be’s website they once had the history of the making of the first bebox. It was quit the read. One fact remains. Eugenia has one, and she can look at it and see how it is and is not made. I will go with that over people debating the merits of dsp.
I hope all goes well with Steve at apple, must say this caught me by surprise. Curious about and reasons for leaving Palm Source. Guess return of the Newton rumors can kick into high gear now
. But then again the other Steve could kill it just as fast again. What is with Steves and Apple?
Ahah, one more religious war I didn’t know about!!!
http://www.bebox.nu/images.php“>BeBox , including Hobbit & PPC pictures.
Er, that should be http://www.bebox.nu/images.php
The Saturn had eight CPUs. No wonder the lame PlayStation couldn’t compete!
😉
“Mr Sakoman was involved in the Mac II development and was the main guy behind the creation of the Newton.”
Could this mean Apple is considering a TabletPC or PalmPC design in the near future? A Newton 2.0 perhaps? I’ll have to watch ThinkSecret and SpyMac a little more closely now.
USB 2.0 will have to be adopted. When you have 5% market share you pretty much go with the flow or die.
apple will not target USB neither firewire over the long run IMHO they will use wireless stuff in tandem with usb and firewire adaptor for the device.
I wonder if Steve get good leaving package, if so he must make quite a lot lately. We all know that apple got him so they can relaunch CodyCam website in streaming quicktime
As for DSP, a CPU is a DSP and a DSP is not a CPU. The only reason why this is becoming more a thin difference lately is because they put CPU core in all kind of stuff lately to compensate for crap bandwith distribution inside current computer architecture.
While I’ve heard that Steve Jobs hates PDAs, it’s hard not to suspect they have some kind of handheld in the works. But I suspect it isn’t going to be a conventional PDA. Like most things they do these days, it’ll probably have a few features that are at least evolutionary (if not revolutionary), have a kind of off-the-wall design and/or UI that will instantly polarize critics, and be a minimum of 25% more expensive than devices it’s ostensibly competing with. (As a Slashdot reader observed, “In order for it to be an Apple product, it has to be an Insanely Great idea that is overpriced enough to make you think twice before buying it.”)
Alien Soldier, I agree. And I don’t care if USB, FireWire or wireless wins out, or that each is targeted toward devices with different data rates. I want one way to connect peripherals to my computer, period.
Sakoman worked on the Newton, which had the estimated lifespan of a fruitfly. He worked on the BeBox, which was the computer equivalent to the Columbia. Lots of neat tech which went down in flames. He then joined Palm, which ain’t doing all that stellar revenue wise anymore nowadays. And now he is back to ruin apple?
Apple’s maximum agreed on market share is 3%, where a number of market researchers see them only at 2,2% to 2,6%…
People. Get over it. DSP can be used as CPUs. There are machines where a m68k is used to do processing, but not as the *central* processor. Big deal. The BeBox used hobbits as CPUs and probably *as* DSPs (for sound). All of those Be engineers must have been too stupid to know that they built a machine around non-CPUs. Someone should’ve let them know! OI.
I find it 100% unlikely that Apple would use Newton OS 2.0 on a PDA/Tablet device- after all, the newer models (2100 and 2000) both ran Newton OS 2.1! NOS 2.0 doesn’t even support ethernet cards. Nitpickerey yes, but I couldn’t help it.
At this point in Apple’s game, I think they would be a helluva lot more likely to create something OS X-compatible. Perhaps pared down, but the last thing Apple wants to do now is cause a split in their product line. Every indication poitns at this. That said, I would froth at the mouth at a new machine running Newton- I use a 2100 (running 2.1) all the time as my main PDA.
I recently measured my USB2/FW drive on PC. amazingly USB1.1 holds up quite well, it boasts 12Mb but delivers 8Mb or 1MByte/sec. But USB2 boasting 480MB only delivers about 5x USB1.1. FW boasts less at 400MB but delivers only 20% more speed. No doubt FW is a more capable std but both are highly overstating what to expect at least for HDs.
I don’t don’t know how many EEs there are here, clearly a few people know something, but far more know less.
I have been designing DSPs in one form or another for 20years now so I probably know as much as I ever will.
DSP as a term can refer to a device that is little more than a multiplier and some control logic or it can be a full blown cpu. When a DSP is not a CPU, it is usually a HW implementation of a very specific Digital Signal Processing function such as FIR filter or a hundred other things, signal enters it and leaves it at certain rates, may be slow or fast. When a DSP is a cpu like device, it is usually executing a more elaborate DSP algorithm written in C that might be a JPEG, FFT or whatever, it might also be concurrently executing several smaller DSP tasks. It may even be running non DSP code.
One huge difference is that DSP cpus process signals at precise rates, running code that takes fixed & well specified execution times. DSP can not run codes that have no known completion times as regular cpus do, ie no spin wheel, no BSOD. They are usually replacements for HW blocks. A general cpu can be used to perform DSP too, it maybe faster, maybe cheaper, but it will still need some support to lock it to the signal rates. Signals don’t wait.
The math that DSPs perform may not even be recognizable to most numerate folks, DSPs often process math that is polynomial in nature that is much harder for regular cpu to process or they can perform math that even a Pentium can do much faster. I would include Error Correction, Compression, Crypto as part of the DSP field, these fields all share alot of common math that is not well implemented on conventional cpus.
DSPs are not neccesarily faster than regular cpus at DSP, Pentiums are now universally in the 1-3GHz region, most DSPs process data at rates anywhere from 100Khz to a few 100MHz. DSPs are used because they perform the job in the right way for the right cost in the right place at the right power levels. Thats why you will generally find a few in most PCs with out usually getting identified.
A 1024FFT may run just as fast on a P4 as it does in special purpose DSP, but it do so 1 op at a time at GHz speed instead of 8ops at a time at more modest speed.
Pentiums/PPCs do have a huge advantage though for floating point, ie FP is about same speed as integer. Almost all DSP is done integer for lowest cost, typically a DSP engine might include 100s of multipliers. It would be painful to do that in FP, but some DSP needs FPU for adaptive algorithms.
Your 56K modem might have one if its not a WinModem, your ADSL & Cable modem definitely has one or more, but then it will likely include a embedded PPC/xyz as well. Most HDs use them now to control the servo systems, they didn’t used to. One can easily argue that a graphics chip is a form of DSP, or a GSP. Audio system has one or more.
DSPs ASFAIK though have never been used to run an OS & desktop GUI, but I don’t think its impossible, just isn’t a useful exercise.
JJ
The difference between a CPU and a DSP is quite simple. A CPU has a control unit (e.g. for jumps) and arithmetic-logic unit (for calculations). A DSP has only a arithmetic-logic unit.
So a DSP can process data but it cannot execute a program.
Bla wrote :
“The difference between a CPU and a DSP is quite simple. A CPU has a control unit (e.g. for jumps) and arithmetic-logic unit (for calculations). A DSP has only a arithmetic-logic unit.
So a DSP can process data but it cannot execute a program.”
Well in that case the so called “DSP” manufacturers (Motorolla, Analogue Devices, Texas Instruments, etc.) are conning us! All their processors have control units and can execute programs… and they sell these as DSPs! The unmitigated cheek of it!!
Or maybe this definition is totally incorrect?
DSPs are quite simply processors which have architectures and instruction sets which are geared towards the processing of time variant signals (e.g. sampled audio or video) – note I use this definition because all binary processors process digital signals (i.e 0s and 1s).
DSPs tend to have a harvard architecture, they almost invariably have a hardware multiplier (both of these also apply to many general purpose processors now, but they were common in DSPs a long time ago), instructions which help in FFT or similar algorithms (bit reversing or reverse carry) are also common.
But none of these things in itself defines the processor as a DSP, in fact there is no clear point at which a processot goes from being “a processor with some extra signal processing functionality” to “a DSP”.
Contrary to other claims in previous posts, DSPs generally have interrupts, and can handle IO (usually they have both memory IO and additional IO for getting the signals to process on and off the chip).
The main defining factor is what it’s intended for, and how it’s used.
That is a no brainer. 2.0 is in no way even close to firewire 400 or 800, but devices like keyboards mice and other devices will start being made in 2.0 when they (wintel manufactures) decide to drop 1.1 . So the move to 2.0 should be made for the eventuallity of devices that will be made for USB 2.0 and not for firewire.
bla
You obviously don’t read!
A DSP and general purpose CPU both have control units and datapath units. A DSP will execute C programs perfectly well if its designed to. Even if it doesn’t execute code, it still has an implied program executed in random logic, ie a controller. DSPs usually executes programs in exact nos of clock cycles (a concept lost on most programmers) in precisely predictable ways. All instructions in a DSP usually but not always take exactly 1 cycle. The branch instruction does exist, whether it branches or not, will not cause much or any change in execution time or eratic pipeline flushes, and generally takes same time either way. It may be possible to write code that can have variable timing, but then it may well have to executes NOPs till it has used up the alloted time.
General CPUs are far more complex because they handle multiple levels of memory and HDs that are upto millions of times slower than the core, and a C program executed on modern x86 executes in x cycles has timing depending on such tiny things as word alignnment, as well as being in Cache at right time. Word misallignment can change execution times by 2x or more. Cache or page misses can be far worse.
DSPs generally execute smaller programs that are held in local ROM or SRAM, (maybe DRAM) and I don’t usually use HDs. Page swapping, cache design, speculation, branch prediction play little part in DSP cpu design.
DSP cpus are more akin to older 68K like cpus where the design goal was to get everything to one cycle by simplifying the architecture, then get the speed by adding instructions that do DSP kernal functions in 1 cycle, such as a FFT rad2 butterfly, or a FIR/IIR filter tap section.
I could go on, you could read a book!
Hal wrote:
Sakoman worked on the Newton, which had the estimated lifespan of a fruitfly… [other slams snipped]
If you had ever met Sakoman, you wouldn’t say those things…he worked on the Newton project very early, and when it went in a different direction than he thought best, he left…there are other retorts to your oddly personal attack, but let’s just say I’ve never met anyone who’s worked with him say anything bad about him.
The memories of being able to go to a store and buy parts off the shelf and just build ie wirewrap a BeBox are long gone unfortunately.
But all is not lost
Would be CPU builders have the newer option of designing almost anything they want and programming into an FPGA. Its now perfectly possible to build a BeBox with some off the shelf Intellectual Property or bake your own blocks. It isn’t easy, the IP blocks are either expensive or free, and not all blocks are available. Even the SW tools are free for smaller projects.
Today Sakoman Jnr could obtain various soft cpus, usb1, FW, ethernet, SDRAM, video controllers and build a basic Apple 1 style PC. It would take alot more work to create an interesting PC like a BeBox but all the things it did could easily fit in a modern FPGA. The speed would not get much more than 120MHz and the FPGAs can get expensive. One nice thing is there are no rules, you can load up many cpus copies if they are small. One free design out there can potentially be copied almost a 1000 times in biggest FPGA, but noone has found a use for that many so far.
Gee you can even get built in quad 300MHz PPCs in the top of the line Xilinx, or a humble Arm in the Altera line, or do your own cpu architecture if you can live nearer 100MHz.
opencores.org would be a place to start for some of the free blocks and even some cpus.
Then again, a 3GHz P4 can simulate any PC you could design, just think of Basilisk, Bochs,VPC,VMWare, but they simulate known cpus.
I just hate it when a simulation is faster than any HW I can build.
It was nice to look at them. Interesting IO board connectors!
The irony about hobbits is that they were made for embeded primitave palm tops and PDA’s by todays standards, the faliure of hobbits then should have been a warning sign for BeIA
Heh! How does that reasoning flow? By that, you could say that because the Apple Newton had an ARM processor, that certain doom is spellt for PocketPC PDAs. MS and PPC manufacturers: take that as a warning sign- switch to a MIPS processor for the PPC!#%!
My script of “Rechnerarchitektur” (german for computer architekture) says a CPU consits of an arithmetic unit and a control unit. If a DSP has both it is a CPU.
From my understanding a dsp takes some digital values, does some operations on them and creates some analog signals (or the other way around) like the early graphics or sound cards.
All thes terms were created long ago where intergration was far away from what it is today.
If Motorola (or someone else) takes a cpu, io-controller and some dsps and put them on one chip, they can call it what they want, it’s still a cpu, io-controller with some dsps.
Classification was never about what you actually do with things but what things actually can do.
>I could go on, you could read a book!
True but the sad part is, i can go on, too, but apparently you can not read a book.
President
Vice President
Senior Vice President
Do we also have a Junior Vice President, Senior President and Junior President?
How about an Executive Vice President? Or a Senior Executive Vice President? Or a Senior Executive Junior Vice President of The Senior Board of Junior Directors?
Corporate structure baffles the hell out of me. Is there any wonder these companies take forever to get things done…
Most people seem to be ignoring the “central” part of Central Processing Unit. While DSPs are certainly processors (being… Digital Signal Processors and all) their purpose, at least in general purpose computers like the one Eugenia is describing, is to offload certain specific signal processing tasks off of the Central Processing Unit.
I would say a DSP is not a CPU because there’s nothing “central” to its purpose or function.
But this is all arguing semantics, which is a boring and pointless exercise…
“Most people seem to be ignoring the “central” part of Central Processing Unit. While DSPs are certainly processors (being… Digital Signal Processors and all) their purpose, at least in general purpose computers like the one Eugenia is describing, is to offload certain specific signal processing tasks off of the Central Processing Unit.
I would say a DSP is not a CPU because there’s nothing “central” to its purpose or function. “
Sorry about that, hit return by accident!!
was I was going to say was:
bascule, you’re actually quite right in that a CPU, by definition should be Central, but that’s because CPU is a role, not a specification as such.
An Intel x86 is an MPU (Micro Processing Unit), which may or may not be the CPU (Central Processing Unit) in a system – it so happens that it usually is. A motorolla DSP56k is also an MPU, it also has features which make it fairly clearly a DSP, and in certain circumstances it may be used as the CPU in a system – for example some audio processing equipment uses a single 56k to do DSP and general processing work.
“From my understanding a dsp takes some digital values, does some operations on them and creates some analog signals (or the other way around) like the early graphics or sound cards.”
bla, your understanding is contradictory to accepted industry definitions of at least 20 years (The Texas Instruments 32 series “single chip DSP” was announced in 1982). I recommend you get a more up to date book 🙂
“All thes terms were created long ago where intergration was far away from what it is today. ”
Terminology is not set in stone, it changes over time, sometimes to encompass the changes that occur in the world, sometimes for no apparant reason. In old English “quick” meant alive, not fast, to a chemist all vegetables are “organic” – but now people think of it as meaning grown without artificial fertilizers or pesticides, and there is a whole industry which accepts the TMS320, dsp56k and SHARC as DSPs.
Apple’s stance is that USB was never designed with those speeds in mind. True, it’s been advanced forward… but it’s counter productive IMHO for Apple to add them. True, you can get a USB 2.0 component for a modest fee less than a Firewire… but it usually amounts to $15-30 per device if you look around as Firewire’s support has gotten bigger and therefore has gotten more affordable.
Apple could make all machines come with USB 2.0 as standard, and do away with the 1.1… but if Apple doesn’t change their mice and modem to USB 2.0, and there’s little point in doing so… then any 2.0 devices you connect on a chain with a 1.0 device will… ::gasp:: not perform at 2.0 speeds. You can get around this with hubs, but at that point… isn’t it cheaper to have just spent the extra $15-30 and get a Firewire device?
Then there’s the theory that Apple could add both USB 1.1 and 2.0 ports to a machine, along with Firewire 400 and 800… but then you get to a myriad of ports that Apple and many others are trying to get away from. It’d also add to the costs of machine purchases to have both, and even the faster USB 2.0 controllers themselves cost marginally more. Apple includes USB 1.1 because it does what it was intended to do, and something Firewire wasn’t intended for. It’s works as it should. If you need a faster performing port… Firewire is definitely worth the $15-30 per device over USB 2.0. If you want USB 2.0, run OS X and buy an add-on card (PCI or Cardbus). There are options, and they do work, not to mention it’s a specialized USB 2.0 port you can dedicate to those oddball necessary functions that Firewire 400 and USB 1.1 can’t handle already.
DSP’s… I’ve come to realize there’s a lot of terms that once were defining, but now things have transcended to where the term is a fallacy. There’s nothing more generic in terms that I can think of than the DSP, or hell… even the CPU is transitioning as many include SIMD(s) onboard the chipset, which makes them a CPU but a SIMD is a form of a DSP both the old fogey definitions. To me, it’s all “processors”, as they “PROCESS” some form of data. Who cares if they’re central, outer, inner, on top of, or outside? LoL Ya’ll have way too much time that you could devote to other things than bicker about semantics to boost your egos and wield your schtick.
Mr. Sakoman… welcome back.
I figure you’re more intended to speed up and advance OS X for the desktop than any PDA/Tablet PC crap like is the widespread rumor. Of course, a color PDA-based iPod that can play audio, movies, and the like is probably in the future… but who knows when that’ll happen. My wager… OS X will begin to develop more of the efficiency you were known for with Be, while having an application’s base worthy of keeping the shebang alive. PDA’s are overrated at the moment (and underpowered; we’ve not made much progress since the Newton came out)… but desktops aren’t.
“…which makes them a CPU but a SIMD is a form of a DSP both the old fogey definitions.”
Should be read as “…by the old fogey definitions”
Must of been the tear in my beer from thinking of the possibility of the melding of the best elements of Be with the best elements of OS X.
Pardone moi…
::hiccup::
“…mice and modem to USB 2.0.”
::hiccup:: I meant keyboard… darn Saki, why does it have to taste so good? ::hiccup::