I seem to recall reading somewhere that springboard is actually the same interface as a current format (someone want to help me out here?)
Springboard is a completely new bus. However, Handspring used the standard 68-pin connector from PCMCIA cards. The pinout is completely different, but the connector is the same one.
Handspring did their best to make it easy to make new Springboard hardware. The connector is standard. The plastic shell is not standard, but you can buy five different shells off-the-shelf. (This is great if you don't want to invest hundreds of thousands of dollars for injection molds and tooling to make your own.)
By the way, Springboard is an open spec: no secrets, no royalties. You can download everything from the Handspring web site Developers page. For example, the Springboard White Paper.
just bear in mind that not *everything* gets backed up in a hotsync
As ashpool7 pointed out, you can force everything to be backed up by setting all the backup bits, and you can use a freeware program to set all the backup bits.
For everyday use, I suggest BackupBuddy. Each and every HotSync will back up everything perfectly.
Traditionally, Intel has owned the high end of the market. (Actually, for a long time they owned the entire market. But even after the clone chips arrived, they always owned the high end.)
For a long time, AMD couldn't match the fastest Intel chips. Each time AMD shipped a new chip, Intel mercilessly lowered the prices on their mid-range and low-end chips. It was hard for AMD to make money; they had to pay for research and development, but they were competing against Intel chips that had already made back their R&D investment. Because AMD was always playing catch-up Intel could always afford to force prices down.
Now, in an ironic reversal, it is AMD that owns the top end of the market. Intel still sells far more chips than AMD does, but AMD now has the upper hand.
In short, AMD is in position to do to Intel what Intel used to do to AMD! And Intel knows it and is running scared.
Gasoline engines in hybrids are more efficient because they run at a constant velocity, and are tuned for that speed.
Some hybrids work like that. Others use the gasoline engine with the electric motor, and do not run the engine at constant velocity.
GNOME a better platform for the future?
on
KDE Strikes Back
·
· Score: 1
This article can be easily summarized: KDE works well, so why would you need anything more? And the GNOME folks are all mean bullies.
On the other hand, here is an article that explains why GNOME is a better platform for the future. It seems fact-filled and flame-free to me.
I would like to see an article from the KDE camp that is as sensible and reasonable.
steveha
Re:KDE/GNOME war hurts developers...
on
KDE Strikes Back
·
· Score: 1
99% of people want to embed a spread sheet into their document, not embed a remote-spread sheet running on a computer 50 miles away.
However, I think it would be a good idea to keep options open for the future. For example, suppose that desktop application servers actually become a big deal in a few years? It may turn out to be a good thing that Gnome used CORBA instead of something with less features.
The key question for me is how much of a run-time penalty you must pay in the here-and-now for CORBA versus something more lightweight. If it isn't too bad, then CORBA is clearly the way to go.
I am very suspicious of quick-and-dirty, roll-your-own hacks for complicated stuff like an object system. Maybe Kparts is brilliant and all you will ever need... but maybe the KDE developers will start banging their heads on it in a few years. (I don't know enough to have an opinion yet.)
I don't think this idea is will happen. What is much more likely is fuel-cell based cars.
Fuel cells react hydrogen with oxygen to create electricity directly. Their waste product is pure water. The main problem is storing the hydrogen.
But the Jet Propulsion Laboratory (JPL) invented a way to use methanol as your hydrogen source. With this technology, methanol fuel-cell vehicles have become possible. They run on methanol and air, and their exhaust is carbon dioxide and water.
Methanol is just an alcohol. You don't need to supercool it. It doesn't crystallize metal. You can pump it from ordinary gas station pumps. The infrastructure changes needed to support methanol are tiny compared to pure hydrogen, or even the liquid nitrogen idea.
I read somewhere that it might even be possible to make a fuel cell that runs on gasoline instead of methanol. I like the idea of an electric car that can stop at ordinary gas stations. No changes needed to infrastructure!
when you've got a few billion in cash, it's not a bad idea to have a few products in your back pocket
True. Remember around 1990, when MS was pushing both Windows and OS/2? Microsoft ported every important Windows app to OS/2. They covered all the bets. When OS/2 fizzled, they dropped the OS/2 products, and ran with the Windows products. If Windows had fizzled instead, the reverse would have happened.
If Microsoft can send a few dollars to a developer somewhere and cover the Linux bet, it makes sense for them to do it, and there is even historical precedent.
it's painfully obvious that M$ apps have had an interface to the OS advantage over everyone else.
Actually, it isn't obvious to me. When I worked on Microsoft Word, what was obvious to me was that the code base was huge and very, very little of that code base was the OS interface.
The vast bulk of Word is code specific to Word. Think about it. How much code do you need to draw menus?
I never saw a secret back door to the OS. I never saw a secret back door to Excel or the other parts of Office. And, when people wave their hands and shout about secret back doors, I notice they never spell out just exactly what those doors do and what advantage MS gets as a result.
I did much of the initial work in porting Word for Windows 6.0 from 16-bit Windows to NT. I looked all through the code when I did that. Any magic back door to Windows would have jumped up and bit me!
Anyway, the article didn't say "secret back door", just "proprietary interfaces"; what they meant isn't clear. Perhaps they were speaking of COM interfaces. Does WINE help right now if your application depends on COM?
I can understand taking out the suicides, but why the accidents?
Look, this is off-topic for the MP3 thread. It's a topic I care about, but we are abusing/. by discussing it here. I'll answer your questions, but if we want to continue this, we need to take it somewhere else.
In the first place, an unknown number of gun cleaning "accidents" are actually suicides, done as a fake accident so that life insurance will pay off despite a no-pay-for-suicide clause. In the second place, people who quote the "43:1" statistic often say you are 43 times more likely to be shot, not to shoot yourself by mistake, so I think it is valid to pull those out. Third and finally, the number of accidents was so few in the Kellerman study that it doesn't really matter if we leave them in or out.
The 75:1 ratio is a myth too of course. By presenting it this way you make believe that without guns, those 75 people would have died.
This seems fair to me, since the anti-gun folks are claiming that in the 43:1 case that none of the gun suicides would have committed suicide using some other means than a gun, and none of the murder victims would have been murdered by some other means, and so on.
If you read the link I put to Dr. Lott's article, you will see that actual research has shown that resisting violence with a gun is the safest thing you can do. ("Give the attacker what he wants" is, for a woman, 2.5 times more likely to result in harm. Even for a man, it is 1.4 times more likely to result in harm.)
See you on some other forum if you want to keep this discussion going.
I'm listening to Trio's quintessential "Da Da Da" song, which the encoder actually made _larger_ in OGG format. Quality's decent, but you have to wonder what causes some files to become larger.
It's simple: Ogg made the ssong larger because it was protecting the data.
Almost all MP3s are encoded at a fixed bitrate. With a fixed bitrate, when the audio is too complicated to encode properly, it just gets chopped. (Also, when a song is easy to encode, such as a very quiet passage, it gets padded out.)
Ogg is a variable bitrate format. You don't really specify the bitrate to encode, you actually specify a quality level. And for that one song, encoding at that quality level worked out to be a bit larger.
Now, if you were to transcode from MP3 to Ogg, I would be very surprised if the Ogg file got bigger. Even there I suppose it is possible with just the right (wrong) data.
People with guns in their house are 43 times more likely to be shot in their own home than people without a gun in their house. You're better off without the gun.
This turns out not to be the case. The 43:1 ratio is a myth. If you take out the suicides, accidents, and justifiable homicides, and consider actual murders compared to actual killing in self-defense, the ratio looks more like 4:1. But most gun owners avoid shooting people as much as they can, and 98% of the time defend themselves without killing anyone, so the actual ratio looks more like 1:75 (75 people saved by guns for each one lost). Follow the link for support of the 1:75 claim. See also here.
Dr. Lott has shown that using a gun to defend yourself gives you the best chance to escape an attack unharmed.
Why did NeXT fail? Because Jobs ran it. [...] I think he let's his desire for "cool" block his sense of "realistic."
I agree. Jobs has made some decisions that left me shaking my head.
The NeXT should have had a floppy. Back in the 80's, most computer users made heavy use of floppy disks to distribute code or data. I remember thinking, "So much for the software market; who will want to release anything on a MO disk that costs $50?"
The Lisa failed because it cost $10,000. Steve Jobs was there. So, how much was the first NeXT computer? That's right... $10,000. ($6,000 at a student discount, but you can't build a viable business on selling at student discount price.) Later boxes were more reasonable, but what was Jobs thinking with that first box?
But in the interview with the Microsoft C# guys, one of the guys said that Microsoft will be submitting the C# language definition to a standards body. Why would they even be talking about that if they want as much lock-in as possible?
Do you really want to embed an editable spreadsheet in a document, and deal with the bloat and crashes that will occur? Or is there a Better Way?
Well, I don't want to embed a spreadsheet in a document, but there are lots and lots of people who do. And I think the general concept is the right idea.
Just as it's cool that I can chain grep, sort, and uniq together in a pipeline and make them all work together to solve a problem for me, it will be cool to be able to get AbiWord and Gnumeric to work together to make a report. Business people have to make reports that show numbers and graphs, and then explain what they all mean. Why should they have to statically export the numbers and graphs and manually import them, every time the numbers change? Or would you prefer to turn Gnumeric into a general-purpose word processor?
Don't forget, when you embed numbers from Gnumeric in Abiword, those numbers can be cached. Then, you can work on your Abiword document without needing Gnumeric running at all times. Embedded data doesn't have to mean slowness and crashes.
As for bloat, which would be more bloated: components that can be shared, or duplication of features?
I hadn't heard of labelled break; now you brought it up, I searched the web and found it here. I like it; "break CrashAndBurn" is better than "goto CrashAndBurn".
Looks like labelled break is the Java way to deal with my first case, and try/finally is the Java way to deal with final cleanup. Cool... maybe goto really is dead.
One question: is try/finally an expensive or risky thing to do, or is it commonly used?
Working with such a small team seems just too cool for Microsoft.
I worked at Microsoft, a few years back, in the Applications side. Apps development was done with "feature teams" of 4-8 people or so. There were around 40 developers working on Word, divided up into about 8 feature teams.
I'm certain the Systems side works in a similar way.
1) Apple spent a lot of money marketing their imac and eMachines tried to leech of of it.
Are you implying that it is okay for Apple to "leech" off the Qube trademark because Cobalt is not spending enough on marketing?
2) Cobalt is a virtually unknown company who probably spent less money on marketing their cube since it started selling than Apple spent attending Macworld to annouce their cube.
According to the article on cnet.com, Cobalt paid $4.1 million dollars to Cube Computer to use the Qube name! Apple didn't even misspell "Cube"...
3) Cobalt is making a publicity stunt to leech of the already bigger mindshare of the G4 Cube.
This makes it okay for Apple to dilute Cobalt's trademark?
4) Apple Entreprise (NeXT), made the NeXT Cube in 1989.
And the Qube is a current, shipping product with a current trademark.
What goes around, comes around, indeed. Apple is either going to need to pay up, as Cobalt did, or change the name to "The G4 Right Rectangular Prism" or something.
Apple uses PARC ideas to build a graphical OS. MS lambasts the products, then uses knowledge obtained under NDA to produce their own GUI-based products.
MS... lambasts? NDA? I don't remember it that way at all; can you give more details?
The way I remember it: Apple and MS use PARC ideas. Apple gets a nice head start by designing new hardware. MS agrees to port Word and Excel to the Mac, as long as Apple agrees never to sue over Windows. Nearly all business Mac users buy both Excel and Word, and MS makes piles of money on the Mac. When Windows begins to look less like a joke, Apple sues MS over Windows.
(Cute lawyer tricks department: Apple lawyers take HP New Wave, based on Windows, and change all the desktop settings to look more like a Mac; then they create screenshots showing how much Windows and New Wave resemble the Mac.)
A paperback book would be too big. It would be an ideal size for use but not for portability.
Perhaps I was unclear. I don't want a Palm PDA the size of a paperback book; I want a gadget for reading books, that is the size of a paperback book.
My Handspring Visor IX or whatever can have the paper-like display, too, but it doesn't have to be big or waterproof. I want a dedicated bookreader that is as good as a book in every way, and waterproof besides.
Are Palms going to wind up like Gameboys, where the actual processor power of the thing never goes up that much?
They shouldn't. I suspect that the older Gameboy games are written like Apple ][ and Commodore 64 games were: they depend on the clock speed of the system, and if you speed up the system, the games run too fast and are unplayable.
Possibly the action games on the Palm are written like that, but I doubt it; we already have a substantial speed difference between the basic Palm models and the Handspring Visors.
The Z-80 is of course famous for being in the atari2600s and timex-sinclair boxes.
Actually, the 2600 used a 6502-related microcontroller, the 6510 I believe. Contemplate, for a moment, a chip even less capable than the 6502! (The 6502 can address 64KB, and I believe the 6510 can only address 4KB.)
Is there some correlation between hand-held devices and a power increase that is more sluggish than Moore's law would predict?
The Palm devices all run on a Motorola 68000-family microcontroller, the Dragonball. When Motorola releases new, faster versions of that chip, you will see new, faster Palm devices.
But keep in mind: one of the best things about the Palm devices is that they don't suck the batteries dry too quickly. Moore's Law stands mute on low-power devices. Expect even low-power devices to get quicker and better, but don't expect them to improve as fast as desktop CPU chips.
Slashdot Mirror
Springboard is a completely new bus. However, Handspring used the standard 68-pin connector from PCMCIA cards. The pinout is completely different, but the connector is the same one.
Handspring did their best to make it easy to make new Springboard hardware. The connector is standard. The plastic shell is not standard, but you can buy five different shells off-the-shelf. (This is great if you don't want to invest hundreds of thousands of dollars for injection molds and tooling to make your own.)
By the way, Springboard is an open spec: no secrets, no royalties. You can download everything from the Handspring web site Developers page. For example, the Springboard White Paper.
steveha
As ashpool7 pointed out, you can force everything to be backed up by setting all the backup bits, and you can use a freeware program to set all the backup bits.
For everyday use, I suggest BackupBuddy. Each and every HotSync will back up everything perfectly.
steveha
For a long time, AMD couldn't match the fastest Intel chips. Each time AMD shipped a new chip, Intel mercilessly lowered the prices on their mid-range and low-end chips. It was hard for AMD to make money; they had to pay for research and development, but they were competing against Intel chips that had already made back their R&D investment. Because AMD was always playing catch-up Intel could always afford to force prices down.
Now, in an ironic reversal, it is AMD that owns the top end of the market. Intel still sells far more chips than AMD does, but AMD now has the upper hand.
In short, AMD is in position to do to Intel what Intel used to do to AMD! And Intel knows it and is running scared.
steveha
Some hybrids work like that. Others use the gasoline engine with the electric motor, and do not run the engine at constant velocity.
On the other hand, here is an article that explains why GNOME is a better platform for the future. It seems fact-filled and flame-free to me.
I would like to see an article from the KDE camp that is as sensible and reasonable.
steveha
However, I think it would be a good idea to keep options open for the future. For example, suppose that desktop application servers actually become a big deal in a few years? It may turn out to be a good thing that Gnome used CORBA instead of something with less features.
The key question for me is how much of a run-time penalty you must pay in the here-and-now for CORBA versus something more lightweight. If it isn't too bad, then CORBA is clearly the way to go.
I am very suspicious of quick-and-dirty, roll-your-own hacks for complicated stuff like an object system. Maybe Kparts is brilliant and all you will ever need... but maybe the KDE developers will start banging their heads on it in a few years. (I don't know enough to have an opinion yet.)
steveha
Fuel cells react hydrogen with oxygen to create electricity directly. Their waste product is pure water. The main problem is storing the hydrogen.
But the Jet Propulsion Laboratory (JPL) invented a way to use methanol as your hydrogen source. With this technology, methanol fuel-cell vehicles have become possible. They run on methanol and air, and their exhaust is carbon dioxide and water.
Methanol is just an alcohol. You don't need to supercool it. It doesn't crystallize metal. You can pump it from ordinary gas station pumps. The infrastructure changes needed to support methanol are tiny compared to pure hydrogen, or even the liquid nitrogen idea.
I read somewhere that it might even be possible to make a fuel cell that runs on gasoline instead of methanol. I like the idea of an electric car that can stop at ordinary gas stations. No changes needed to infrastructure!
Read more here.
steveha
True. Remember around 1990, when MS was pushing both Windows and OS/2? Microsoft ported every important Windows app to OS/2. They covered all the bets. When OS/2 fizzled, they dropped the OS/2 products, and ran with the Windows products. If Windows had fizzled instead, the reverse would have happened.
If Microsoft can send a few dollars to a developer somewhere and cover the Linux bet, it makes sense for them to do it, and there is even historical precedent.
steveha
Actually, it isn't obvious to me. When I worked on Microsoft Word, what was obvious to me was that the code base was huge and very, very little of that code base was the OS interface.
The vast bulk of Word is code specific to Word. Think about it. How much code do you need to draw menus?
I never saw a secret back door to the OS. I never saw a secret back door to Excel or the other parts of Office. And, when people wave their hands and shout about secret back doors, I notice they never spell out just exactly what those doors do and what advantage MS gets as a result.
I did much of the initial work in porting Word for Windows 6.0 from 16-bit Windows to NT. I looked all through the code when I did that. Any magic back door to Windows would have jumped up and bit me!
Anyway, the article didn't say "secret back door", just "proprietary interfaces"; what they meant isn't clear. Perhaps they were speaking of COM interfaces. Does WINE help right now if your application depends on COM?
steveha
Look, this is off-topic for the MP3 thread. It's a topic I care about, but we are abusing /. by discussing it here. I'll answer your questions, but if we want to continue this, we need to take it somewhere else.
In the first place, an unknown number of gun cleaning "accidents" are actually suicides, done as a fake accident so that life insurance will pay off despite a no-pay-for-suicide clause. In the second place, people who quote the "43:1" statistic often say you are 43 times more likely to be shot, not to shoot yourself by mistake, so I think it is valid to pull those out. Third and finally, the number of accidents was so few in the Kellerman study that it doesn't really matter if we leave them in or out.
This seems fair to me, since the anti-gun folks are claiming that in the 43:1 case that none of the gun suicides would have committed suicide using some other means than a gun, and none of the murder victims would have been murdered by some other means, and so on.
If you read the link I put to Dr. Lott's article, you will see that actual research has shown that resisting violence with a gun is the safest thing you can do. ("Give the attacker what he wants" is, for a woman, 2.5 times more likely to result in harm. Even for a man, it is 1.4 times more likely to result in harm.)
See you on some other forum if you want to keep this discussion going.
steveha
It's simple: Ogg made the ssong larger because it was protecting the data.
Almost all MP3s are encoded at a fixed bitrate. With a fixed bitrate, when the audio is too complicated to encode properly, it just gets chopped. (Also, when a song is easy to encode, such as a very quiet passage, it gets padded out.)
Ogg is a variable bitrate format. You don't really specify the bitrate to encode, you actually specify a quality level. And for that one song, encoding at that quality level worked out to be a bit larger.
Now, if you were to transcode from MP3 to Ogg, I would be very surprised if the Ogg file got bigger. Even there I suppose it is possible with just the right (wrong) data.
steveha
This turns out not to be the case. The 43:1 ratio is a myth. If you take out the suicides, accidents, and justifiable homicides, and consider actual murders compared to actual killing in self-defense, the ratio looks more like 4:1. But most gun owners avoid shooting people as much as they can, and 98% of the time defend themselves without killing anyone, so the actual ratio looks more like 1:75 (75 people saved by guns for each one lost). Follow the link for support of the 1:75 claim. See also here.
Dr. Lott has shown that using a gun to defend yourself gives you the best chance to escape an attack unharmed.
steveha
I agree. Jobs has made some decisions that left me shaking my head.
The NeXT should have had a floppy. Back in the 80's, most computer users made heavy use of floppy disks to distribute code or data. I remember thinking, "So much for the software market; who will want to release anything on a MO disk that costs $50?"
The Lisa failed because it cost $10,000. Steve Jobs was there. So, how much was the first NeXT computer? That's right... $10,000. ($6,000 at a student discount, but you can't build a viable business on selling at student discount price.) Later boxes were more reasonable, but what was Jobs thinking with that first box?
steveha
But in the interview with the Microsoft C# guys, one of the guys said that Microsoft will be submitting the C# language definition to a standards body. Why would they even be talking about that if they want as much lock-in as possible?
Well, I don't want to embed a spreadsheet in a document, but there are lots and lots of people who do. And I think the general concept is the right idea.
Just as it's cool that I can chain grep, sort, and uniq together in a pipeline and make them all work together to solve a problem for me, it will be cool to be able to get AbiWord and Gnumeric to work together to make a report. Business people have to make reports that show numbers and graphs, and then explain what they all mean. Why should they have to statically export the numbers and graphs and manually import them, every time the numbers change? Or would you prefer to turn Gnumeric into a general-purpose word processor?
Don't forget, when you embed numbers from Gnumeric in Abiword, those numbers can be cached. Then, you can work on your Abiword document without needing Gnumeric running at all times. Embedded data doesn't have to mean slowness and crashes.
As for bloat, which would be more bloated: components that can be shared, or duplication of features?
steveha
Looks like labelled break is the Java way to deal with my first case, and try/finally is the Java way to deal with final cleanup. Cool... maybe goto really is dead.
One question: is try/finally an expensive or risky thing to do, or is it commonly used?
I know you're trying to be funny. Try harder.
How pathetic is goto? Depends on what you use it for. We all agree it's pathetic for normal control flow.
...
When writing C code, there are a couple of cases where I actually use goto.
Breaking out of a multiply-nested loop:
for (x = 0; x < XMAX; ++x)
for (y = 0; y < YMAX; ++y)
if (DoSomething(x, y) != 0)
goto CrashAndBurn;
General cleanup:
thing = AllocateThing();
if (DoSomething1(thing) != 0)
goto CleanupReturn;
if (DoSomething2(thing) != 0)
goto CleanupReturn;
return 0;
CleanupReturn:
FreeThing(thing);
return 1;
I am most familiar with pure C. Do you use exceptions or something like that to handle these cases cleanly in Java?
I worked at Microsoft, a few years back, in the Applications side. Apps development was done with "feature teams" of 4-8 people or so. There were around 40 developers working on Word, divided up into about 8 feature teams.
I'm certain the Systems side works in a similar way.
steveha
Are you implying that it is okay for Apple to "leech" off the Qube trademark because Cobalt is not spending enough on marketing?
2) Cobalt is a virtually unknown company who probably spent less money on marketing their cube since it started selling than Apple spent attending Macworld to annouce their cube.
According to the article on cnet.com, Cobalt paid $4.1 million dollars to Cube Computer to use the Qube name! Apple didn't even misspell "Cube"...
3) Cobalt is making a publicity stunt to leech of the already bigger mindshare of the G4 Cube.
This makes it okay for Apple to dilute Cobalt's trademark?
4) Apple Entreprise (NeXT), made the NeXT Cube in 1989.
And the Qube is a current, shipping product with a current trademark.
What goes around, comes around, indeed. Apple is either going to need to pay up, as Cobalt did, or change the name to "The G4 Right Rectangular Prism" or something.
steveha
I didn't want to see that picture. How did this get a score of 2?
MS... lambasts? NDA? I don't remember it that way at all; can you give more details?
The way I remember it: Apple and MS use PARC ideas. Apple gets a nice head start by designing new hardware. MS agrees to port Word and Excel to the Mac, as long as Apple agrees never to sue over Windows. Nearly all business Mac users buy both Excel and Word, and MS makes piles of money on the Mac. When Windows begins to look less like a joke, Apple sues MS over Windows.
(Cute lawyer tricks department: Apple lawyers take HP New Wave, based on Windows, and change all the desktop settings to look more like a Mac; then they create screenshots showing how much Windows and New Wave resemble the Mac.)
The 6507 is in fact a 6502 that can only address 4KB of memory. The 6510 is a 6502 with an extra 8-bit I/O register.
My source: http://www.atarihq.com/danb/6502page.htm
Yeah... 6502, 6503; whatever it takes...
steveha
Perhaps I was unclear. I don't want a Palm PDA the size of a paperback book; I want a gadget for reading books, that is the size of a paperback book.
My Handspring Visor IX or whatever can have the paper-like display, too, but it doesn't have to be big or waterproof. I want a dedicated bookreader that is as good as a book in every way, and waterproof besides.
steveha
They shouldn't. I suspect that the older Gameboy games are written like Apple ][ and Commodore 64 games were: they depend on the clock speed of the system, and if you speed up the system, the games run too fast and are unplayable.
Possibly the action games on the Palm are written like that, but I doubt it; we already have a substantial speed difference between the basic Palm models and the Handspring Visors.
Actually, the 2600 used a 6502-related microcontroller, the 6510 I believe. Contemplate, for a moment, a chip even less capable than the 6502! (The 6502 can address 64KB, and I believe the 6510 can only address 4KB.)
The Palm devices all run on a Motorola 68000-family microcontroller, the Dragonball. When Motorola releases new, faster versions of that chip, you will see new, faster Palm devices.
But keep in mind: one of the best things about the Palm devices is that they don't suck the batteries dry too quickly. Moore's Law stands mute on low-power devices. Expect even low-power devices to get quicker and better, but don't expect them to improve as fast as desktop CPU chips.
steveha