Are there any uses for DAE, besides ripping music? It's seems to me that's pretty much it's sole purpose... used to be, in the days of 8x (and lesser) cd-rom drives, a lot of drives didn't even support DAE and they worked fine for everything but ripping.
Yes there are. On the Mac, there is no audio connection between the CDRom (or DVD Rom) drive and the sound output. The audio data is ripped off the CD, and sent to the computer's sound system in real time. This was done, partially, so that users could listen to their CDs over their USB speakers.
An interesting side effect of this is that if you do it right, you can play two tracks off the CD at the same time, with no skipping. Why you'd want to do this is another matter entirely.
Well, it seems that I had a totally different experience with Myst III. The game itself ran perfectly on my PowerMac G4/400, and I have yet to hear of anyone having problems with the game on a Mac.
What I liked:
- Graphics: As per usual, the game's graphics were simply astounding. The 360 degree rotation technology that was used to produce the game worked flawlessly and the movie clips melded into the scenes perfectly.
- Sound: Again, the audio track of the game was great. The music was not intrusive, and added a lot to the game.
- Gameplay: The gameplay of Myst III was quite good. As oposed to Riven where flipping a switch on one side of the world might change something on the oposite end, all the puzzels and what not were pretty much self contained. Basically, you can immediately see the effects of your actions.
What I didn't like:
- Difficulty: Basically, the only thing I didn't like about the game was that some of the puzzels were too easy.
Overall, in my experience, this game was well worth the money I spent on it. As for the complaint about the packaging, well, quite frankly, Riven didn't come with much else, and I was pleased to see that the four CDs were in a decent jewel case, instead of cheap cardboard. As far as a journal, this game didn't really need one. All I needed was a letter sized peice of paper.
Before you cut any holes, or pull any cable, walk around your house and figure out exactly where you want to put the wall plates. Then go and figure out how you want to run the cable. Drill exploratory holes and/or probe with nails if you need to figure out the structure. When you have a coherent plan, then start laying cable. You will obviously run into unexpected issues, but hopefully they will be minimized.
First some background, my home is totally wired. We have some 30 ethernet drops spread throughout the house, all terminating in a central wiring/server closet. Based on this experience, I have a few peices of advice to offer.
1) Always do it right.
By this, I mean don't do anything half-assed just because it's a home installation. This is a mistake I see a lot of people making. Network cable is rather fragile (if you want to maintain 100BaseT capability). This means that running it over the floor, through heating ducts, and what not is generally not a good idea. Plus, running the cable that way is just ugly, and a potential danger. And besides, drywall is easy to patch.
Also, it's incredibly important to document which cable is which. There is nothing more frustrating then trying to figure out which jack/end in one room matches up with which cable down in another room. The easiest way to do this is to come up with a numbering convention for your home, and then mark these numbers on both ends of the cable with some sort of label (Masking tape/pen works fine).
All the cable in our network terminates is laid to cat-5 specifications, and terminates in a proper wall plate on one end, and a patch pannel (see Ours) on the other. This has saved us hours of trying to track down problems, and allows us to quickly bypass problems should one occur. If I were to do our network again, the only change I would make would be to use a 50 pair horizontal cross connect cable to go from our second floor to the wiring closet.
2) Don't underestimate how much cable you need.
The other mistake I see people making is that they underestimate how much cable they need to run to their bedroom, or any other room. Invariably, this leads to chaining hubs, and all sorts of other problems. It's best to pull lots of cable, and if you don't need it right now, just leave it in the walls for future use.
In our place, we pulled 4 lines to each of the bedrooms, and that's barely enough. As it stands, the three of us that live in our home are all using all four drops in each bedroom.
3) Always pull a string.
Whenever you pull cable through floors/ceilings/walls always pull a string along with the cable, and do your best to keep it from getting tangled or twisted around the cable. This string can then be used, in turn, to pull the next cable. When you're done, leave a string in place, in case you suddenly realise you need a nother cable. Your cheap Polyester twine that can be had from Home Depot will be more then adequate.
4) Don't go totally overboard.
Don't spend too much on stuff that you don't need. Cat-5e isn't really worth it, since gigabit can go over properly laid Cat-5, don't lay fiber unless you intend to actually use it, don't buy plenum cable unless your fire code requires it, ect.. Really, always just keep an eye out for the best price/reliability values out there.
This is more important for things like jacks and connectors then for hubs/switches. With a good quality jack, the wires will not come loose. We've been bitten by this more then a few times, and have had to punch things down more often then we'd like.
Hubs and switches, in general, are pretty reliable no matter who makes them. For home use, there really isn't much point in purchasing the latest and greatest 3Com switch, when something less expensive will do the trick. It's highly unlikely that you will need to use all the bells and whistles an expensive one would provide you with.
5) Think Safety.
When opening up walls to pull cable and/or drilling exploratory holes, always be weary of hitting electrical cabling. 120V is not something that you want to hit with a hand saw. Also, be careful if you're up on a ladder. It's very easy to come un-ballanced while doing a hard pull.
6) Add Toys.:)
The last thing to do, once you have a stable platform to work on is to add toys. 802.11 is fun, and incredibly cool, but still much more expensive then a wired solution (as long as you're handy, and can do a reasonable job yourself).
Anyway, I hope this information helps people out with their plans. If there is enough interest, I can put together a list of materials that we used, to give an idea of what works. (Most of our equipment comes from Lin Haw).
Actually, I have copies of NeXTStep 3.3 for Black Hardware, Intel Hardware, Sparc, and PA-RISC, as well as OpenStep for all these platforms. (My '040 NeXT Cube runs OpenStep 4.2)
Huh? "minimums"? Obviously you don't mean that the rates are guaranteed after the CO... And the min speed would also the max speed between you and the CO, since there isn't any line sharing.
Not quite true. ADSL, and all DSL is based around ATM (Your DSL modem is an Ethernet over ATM bridge). ATM has a couple of nifty features, including the ability to guarantee a minimum amount of bandwidth to a given connection.
This means that an ISP could have 100Mbit of bandwidth from the aggregation point to their uplink and sell 100 people 4Mbit links, with 1 MBit guaranteed.
I still have my entire BBS, still configured, on an old 130 MB HD lying around, complete with voting booth. It should be theoretically possible to slap the drive in a machine, flip it on, and have my BBS reboot.. which I should do, if only for nostalgic purposes.:)
Just thinking of another solution. What if there was a concerted, online effort to track which subnets thes pseudo-legal attackers are using? There are obvious ways around this, such as using free ISPs (or any dialup ISP for that matter), open proxies, anonymisers, and what not.. but then again, I would tend to say that these should be blocked out as a general security risk as it is.
Another thing to consider: Since the community finds this trademark to be distasteful, why doesn't everyone start violating it flagrently? He can't possibly go after everyone, and thus since he hasn't adequately defended his tradmark, will it not be declared invalid?
But... getting the car to a higher ground would be far more challenging... I wonder how they managed to get the VW up onto the bridge towers in the past without getting caught...
It's actually quite simple.. disassemble the shell into small pieces, carry them up the ladders/stairs inside the towers, get on top of the bridge, and re-assemble the vehicle. But again, when time is short, best option is to just toss it over the edge.:)
Essentially, the article says that Raskin doesn't like MacOS X, MS Windows, or any other general purpose operating system for that matter, because he thinks that computers should be pure appliances, relieving the user of having to worry about mundanities like file storage or program launching, rather than infinitely mutable environments.
There is not really any reason why you can't have an esentially transparent operating system, and still maintain a large degree of flexibility. I hate to bring this up again, but in my experience the best example of how this can be done is the Newton. Instead of having a distinct difference between the operating system and the applications, the line is very blurry.
Much of the software available for it simply works as an extension to the operating system, rather thn as a distinct program. A good example is the email program SimpleMail.. rather then having a separate program to read and write the mail, you can compose the message in the the note pad, then send it using an email address out of your (global) address book. Receiving works the oposite way, and the email address and name of the person who sent you the email is added to your address book, if you don't already have an entry.
IMNSHO, This is a very bad idea. No matter how hard Paul tries, the information will leak out. There are simply too many people who would be on that list and don't give a rats ass about any NDA, or will refuse to sign said NDA. Many people live in countries where these NDAs would not even be worth the paper that they're printed on.
The only safe way to deal with this is to diseminate the information far and wide, in the hopes that everyone will update as soon as possible.
If an operator does not update their system, and gets comprimised due to the security hole, it's the operator's fault. If the system is comprimised, and information regarding the vulnerability was not released, it's the author's fault.
A better example would be AM/FM radios. We've been told that years that radios shouldn't be used on airplanes, because they can cause problems with navigation equipment. I find it hard to believe that a passive receiver can cause more problems than an active 802.11b transmitter.
The problem is that an FM receiver is not a passive receiver. If I remember correctly, part of the FM receiver includes an RF modulator to generate the carrier frequency used for the FM broadcast, thus allowing it to be removed from the inbound signal. If noise from this leaked out of the unit, it could play havoc with inflight systems, since it's down in the 100 Mhz Range. 802.11b, on the other hand, is up in the 2.4 Ghz range, will above what I suspect is being used on most aircraft.
See, those companies are charged the postage for the envelopes when they *buy* the envelopes, not when the envelopes are actually mailed. There is a flat charge the Post Office charges for each envelope, which includes the return postage.
You're forgetting the cost to the company to pay for someone to actually open up the envelope only to discover that it's not what they were expecting. This alone probably costs the company more then buying the envelope/return postage.
Hmm. If I remember correctly, at the time of the Manhattan project the nuclear physics involved in the atomic bomb were already fairly well known. Little boy, being a gun type weapon, was extreemly simple: slam two peices of Uranium together to push it over critical mass and, well, boom.
Fatman, on the other hand, was an implosion type weapon. On the surface, the nuclear physics are again quite simple: compress a plutonium pit to the point of criticality, and watch the fireworks. The trick is the compression bit. This requires absolutely precise use of explosives and control of the gases that the explosion creates, a hugely difficult fluid dynamics/chemistry problem. It's this very problem that keeps a lot of nation-states from developing their own nuclear arsenals. Even if they get their hands on weapons grade plutonium, they need to develop a way to set it off. If the chemical explosion is not absolutely perfect, the bomb will simply scatter the plutonium all over the place and not go critical.
From what I gather, every media device that would conform to this spec would require a unique key. (Otherwise there is no reason why I can see that the data couldn't be transfered to another HD).
If this is the case, why not write a hardware abstraction layer that traps for the command to retrieve said key, and instead return some bogus, bub non-unique key? Obviously this would affect performance a bit, but the impact shouldn't be that big.
Also, how the hell would this affect those of us running RAIDs? or are we not allowed to do that either?
Move to Canada: No DMCA, no UCITA, no software patents, no bullshit.
Yes, but compare specs. You're talking Quad PIII Xeon running at 800 Mhz. Mot did their demo with a single 350 Mhz 7400. I'd like to see you try and stick 4 800 Mhz PIII Xeons in a box no bigger then a hard cover Tom Clancy book, without a fan. (which you could do with a single G4).
The GPL's chief goal is its viral nature, not its copyleft. The BSD license is just as good at keeping code open and it is not viral.
If a company chooses to use my code in one of their products, I expect some form of compensation, preferably in the form of updates and improvements to my code. The GPL forces them to do this, the BSD license simply doesn't protect my rights.
Secondly, if my understanding of the GPL is correct, you can still link GPLd code into a closed source binary if you don't need to make any modifications to the GPLd source, and/or if you release any modifications that you make to the GPLd code. (I think that the only people who'd have a problem with this are nutcase zealots like RMS).
Based on my understanding of the AltiVec unit, I think there are a few kernel level options that AltiVec may be useful for.
The most obvious is optimizations to the firewalling and routing code within the kernel. The AltiVec unit has a "permute" function that allows you to generate a 128 bit vector by picking and choosing words from two other 128 bit vectors. Aparently, Motorola has built a software router around a 7400 capable of doing software routing on multiple T3s. In this situation, the AltiVec unit simply becomes a glorified switching fabric.
The second optimization that I can see would be in kernel level encryption, for doing things like IPSec and/or encrypted filesystems. If I remember correctly, this use was sugested in one of the documents available on Motorola's web site.
Esentially, if you stop thinking of the AltiVec unit as a media processor, and think it more of a parallel processing unit, you can use it for a lot of different tasks.
Well, it looks like it is time to move to a different country. Most other western nations that I can think of (Most of Europe, Canada, Austrailia, New Zeland, ect..) don't have these stupid patent offices/laws, nor do they have the DMCA, UCITA, of (back when it was relevant) crypto export laws. Perhaps southern california is not the best place to conduct high-tech buisness after all.:)
With my Palm, I snap on a modem, plug in a phone line, and connect PPP to earthlink. Then I can check/send email (Eudora), "surf" the web (EudoraWeb, AvantGo, others), get on IRC, use AIM, use Yahoo! Messenger, and can telnet. All these clients use their respective "standard protocol."
Point taken, however in the experience I've had playing around with a Palm connected to the 'net, I was.. well.. rather underwealmed. Because of the screen size, surfing with the Palm is akin to looking at a mural a postage stamp square at a time.
In reality, I think the differences in my experiences has more to do with screen size, and CPU power then application availability, two things which the palm doesn't have much of. With a 320x480 screen, and a 162 mhz StrongARM, you can simply provide a richer experience to the user.
What it all comes down to, though, is what tool do you find works the best for you.
The other hand-held that hasn't been mentioned here yet is the Newton. Despite the fact that it has been discontinued, I still have yet to see a handheld that comes close to matching the elegance of the Newton.
In my opinion, it is what the other free efforts should be modling themselves after. The UI was written from the ground up to to be pen oriented, rather then being an adaptation of a desktop UI.
Also, it provides for tight integration between the OS and 3rd party applications. For example, the email program I use on my MP 2100 uses the standard contact information for getting its email addresses, and is treated by the system as simply another way to get information in and out of the system.
Thirdly, it has real, honest to God, handwriting recognition. Despite all the jokes that were made about it when the Newton was first released, it was perfected in the MP2x00. I can write at full speed, cursively, and it's accurate about 99% of the time.
Finally, the hand-held should be able to operate independently from a desktop PC. I should not have to sync it on a nightly basis with my host PC for it to be fully useable. The NewtonOS has a full TCP/IP stack, with communications being done over Ethernet, or PPP. The email, browser, ftp, irc, telnet, ect.. clients all use the standard protocols, and you can download new packages directly off of the 'net.
Anyway, if some one could produce a hand-held with these types of capabilities, I'd buy one in an instant. Until then, I'll stick with my MP2100.
FYI: Here in Vancouver, we pay $34.95 for 1.2Mbit/650Kbit ADSL. At current exchange rates, that's $22.46 USD. The kicker is that we're 2 hops off of Telus's Gigapop, with access to 5 major backbones.
One problem: At Apple, the Hardware sales finance the Operating Systems. Apple is primarily a hardware manufacturer, not an OS company. If they started selling OS X for Intel hardware, they'd die a slow and ugly death. Look at what happened to NeXT and Be after they stoped producing the hardware.
Slashdot Mirror
Yes there are. On the Mac, there is no audio connection between the CDRom (or DVD Rom) drive and the sound output. The audio data is ripped off the CD, and sent to the computer's sound system in real time. This was done, partially, so that users could listen to their CDs over their USB speakers.
An interesting side effect of this is that if you do it right, you can play two tracks off the CD at the same time, with no skipping. Why you'd want to do this is another matter entirely.
Well, it seems that I had a totally different experience with Myst III. The game itself ran perfectly on my PowerMac G4/400, and I have yet to hear of anyone having problems with the game on a Mac.
What I liked:
- Graphics: As per usual, the game's graphics were simply astounding. The 360 degree rotation technology that was used to produce the game worked flawlessly and the movie clips melded into the scenes perfectly.
- Sound: Again, the audio track of the game was great. The music was not intrusive, and added a lot to the game.
- Gameplay: The gameplay of Myst III was quite good. As oposed to Riven where flipping a switch on one side of the world might change something on the oposite end, all the puzzels and what not were pretty much self contained. Basically, you can immediately see the effects of your actions.
What I didn't like:
- Difficulty: Basically, the only thing I didn't like about the game was that some of the puzzels were too easy.
Overall, in my experience, this game was well worth the money I spent on it. As for the complaint about the packaging, well, quite frankly, Riven didn't come with much else, and I was pleased to see that the four CDs were in a decent jewel case, instead of cheap cardboard. As far as a journal, this game didn't really need one. All I needed was a letter sized peice of paper.
Lets just pray that they don't get Shatner to sing the theme song. :P I've already been scarred for life by his singing. :D
One other thing I forgot to put in:
Plan out your Network.
Before you cut any holes, or pull any cable, walk around your house and figure out exactly where you want to put the wall plates. Then go and figure out how you want to run the cable. Drill exploratory holes and/or probe with nails if you need to figure out the structure. When you have a coherent plan, then start laying cable. You will obviously run into unexpected issues, but hopefully they will be minimized.
First some background, my home is totally wired. We have some 30 ethernet drops spread throughout the house, all terminating in a central wiring/server closet. Based on this experience, I have a few peices of advice to offer.
:)
1) Always do it right.
By this, I mean don't do anything half-assed just because it's a home installation. This is a mistake I see a lot of people making. Network cable is rather fragile (if you want to maintain 100BaseT capability). This means that running it over the floor, through heating ducts, and what not is generally not a good idea. Plus, running the cable that way is just ugly, and a potential danger. And besides, drywall is easy to patch.
Also, it's incredibly important to document which cable is which. There is nothing more frustrating then trying to figure out which jack/end in one room matches up with which cable down in another room. The easiest way to do this is to come up with a numbering convention for your home, and then mark these numbers on both ends of the cable with some sort of label (Masking tape/pen works fine).
All the cable in our network terminates is laid to cat-5 specifications, and terminates in a proper wall plate on one end, and a patch pannel (see Ours) on the other. This has saved us hours of trying to track down problems, and allows us to quickly bypass problems should one occur. If I were to do our network again, the only change I would make would be to use a 50 pair horizontal cross connect cable to go from our second floor to the wiring closet.
2) Don't underestimate how much cable you need.
The other mistake I see people making is that they underestimate how much cable they need to run to their bedroom, or any other room. Invariably, this leads to chaining hubs, and all sorts of other problems. It's best to pull lots of cable, and if you don't need it right now, just leave it in the walls for future use.
In our place, we pulled 4 lines to each of the bedrooms, and that's barely enough. As it stands, the three of us that live in our home are all using all four drops in each bedroom.
3) Always pull a string.
Whenever you pull cable through floors/ceilings/walls always pull a string along with the cable, and do your best to keep it from getting tangled or twisted around the cable. This string can then be used, in turn, to pull the next cable. When you're done, leave a string in place, in case you suddenly realise you need a nother cable. Your cheap Polyester twine that can be had from Home Depot will be more then adequate.
4) Don't go totally overboard.
Don't spend too much on stuff that you don't need. Cat-5e isn't really worth it, since gigabit can go over properly laid Cat-5, don't lay fiber unless you intend to actually use it, don't buy plenum cable unless your fire code requires it, ect.. Really, always just keep an eye out for the best price/reliability values out there.
This is more important for things like jacks and connectors then for hubs/switches. With a good quality jack, the wires will not come loose. We've been bitten by this more then a few times, and have had to punch things down more often then we'd like.
Hubs and switches, in general, are pretty reliable no matter who makes them. For home use, there really isn't much point in purchasing the latest and greatest 3Com switch, when something less expensive will do the trick. It's highly unlikely that you will need to use all the bells and whistles an expensive one would provide you with.
5) Think Safety.
When opening up walls to pull cable and/or drilling exploratory holes, always be weary of hitting electrical cabling. 120V is not something that you want to hit with a hand saw. Also, be careful if you're up on a ladder. It's very easy to come un-ballanced while doing a hard pull.
6) Add Toys.
The last thing to do, once you have a stable platform to work on is to add toys. 802.11 is fun, and incredibly cool, but still much more expensive then a wired solution (as long as you're handy, and can do a reasonable job yourself).
Anyway, I hope this information helps people out with their plans. If there is enough interest, I can put together a list of materials that we used, to give an idea of what works. (Most of our equipment comes from Lin Haw).
Actually, I have copies of NeXTStep 3.3 for Black Hardware, Intel Hardware, Sparc, and PA-RISC, as well as OpenStep for all these platforms. (My '040 NeXT Cube runs OpenStep 4.2)
Not quite true. ADSL, and all DSL is based around ATM (Your DSL modem is an Ethernet over ATM bridge). ATM has a couple of nifty features, including the ability to guarantee a minimum amount of bandwidth to a given connection.
This means that an ISP could have 100Mbit of bandwidth from the aggregation point to their uplink and sell 100 people 4Mbit links, with 1 MBit guaranteed.
I still have my entire BBS, still configured, on an old 130 MB HD lying around, complete with voting booth. It should be theoretically possible to slap the drive in a machine, flip it on, and have my BBS reboot.. which I should do, if only for nostalgic purposes. :)
Just thinking of another solution. What if there was a concerted, online effort to track which subnets thes pseudo-legal attackers are using? There are obvious ways around this, such as using free ISPs (or any dialup ISP for that matter), open proxies, anonymisers, and what not.. but then again, I would tend to say that these should be blocked out as a general security risk as it is.
Another thing to consider: Since the community finds this trademark to be distasteful, why doesn't everyone start violating it flagrently? He can't possibly go after everyone, and thus since he hasn't adequately defended his tradmark, will it not be declared invalid?
It's actually quite simple.. disassemble the shell into small pieces, carry them up the ladders/stairs inside the towers, get on top of the bridge, and re-assemble the vehicle. But again, when time is short, best option is to just toss it over the edge.
There is not really any reason why you can't have an esentially transparent operating system, and still maintain a large degree of flexibility. I hate to bring this up again, but in my experience the best example of how this can be done is the Newton. Instead of having a distinct difference between the operating system and the applications, the line is very blurry.
Much of the software available for it simply works as an extension to the operating system, rather thn as a distinct program. A good example is the email program SimpleMail.. rather then having a separate program to read and write the mail, you can compose the message in the the note pad, then send it using an email address out of your (global) address book. Receiving works the oposite way, and the email address and name of the person who sent you the email is added to your address book, if you don't already have an entry.
IMNSHO, This is a very bad idea. No matter how hard Paul tries, the information will leak out. There are simply too many people who would be on that list and don't give a rats ass about any NDA, or will refuse to sign said NDA. Many people live in countries where these NDAs would not even be worth the paper that they're printed on.
The only safe way to deal with this is to diseminate the information far and wide, in the hopes that everyone will update as soon as possible.
If an operator does not update their system, and gets comprimised due to the security hole, it's the operator's fault. If the system is comprimised, and information regarding the vulnerability was not released, it's the author's fault.
The problem is that an FM receiver is not a passive receiver. If I remember correctly, part of the FM receiver includes an RF modulator to generate the carrier frequency used for the FM broadcast, thus allowing it to be removed from the inbound signal. If noise from this leaked out of the unit, it could play havoc with inflight systems, since it's down in the 100 Mhz Range. 802.11b, on the other hand, is up in the 2.4 Ghz range, will above what I suspect is being used on most aircraft.
You're forgetting the cost to the company to pay for someone to actually open up the envelope only to discover that it's not what they were expecting. This alone probably costs the company more then buying the envelope/return postage.
Hmm. If I remember correctly, at the time of the Manhattan project the nuclear physics involved in the atomic bomb were already fairly well known. Little boy, being a gun type weapon, was extreemly simple: slam two peices of Uranium together to push it over critical mass and, well, boom.
Fatman, on the other hand, was an implosion type weapon. On the surface, the nuclear physics are again quite simple: compress a plutonium pit to the point of criticality, and watch the fireworks. The trick is the compression bit. This requires absolutely precise use of explosives and control of the gases that the explosion creates, a hugely difficult fluid dynamics/chemistry problem. It's this very problem that keeps a lot of nation-states from developing their own nuclear arsenals. Even if they get their hands on weapons grade plutonium, they need to develop a way to set it off. If the chemical explosion is not absolutely perfect, the bomb will simply scatter the plutonium all over the place and not go critical.
From what I gather, every media device that would conform to this spec would require a unique key. (Otherwise there is no reason why I can see that the data couldn't be transfered to another HD).
If this is the case, why not write a hardware abstraction layer that traps for the command to retrieve said key, and instead return some bogus, bub non-unique key? Obviously this would affect performance a bit, but the impact shouldn't be that big.
Also, how the hell would this affect those of us running RAIDs? or are we not allowed to do that either?
Move to Canada: No DMCA, no UCITA, no software patents, no bullshit.
Yes, but compare specs. You're talking Quad PIII Xeon running at 800 Mhz. Mot did their demo with a single 350 Mhz 7400. I'd like to see you try and stick 4 800 Mhz PIII Xeons in a box no bigger then a hard cover Tom Clancy book, without a fan. (which you could do with a single G4).
If a company chooses to use my code in one of their products, I expect some form of compensation, preferably in the form of updates and improvements to my code. The GPL forces them to do this, the BSD license simply doesn't protect my rights.
Secondly, if my understanding of the GPL is correct, you can still link GPLd code into a closed source binary if you don't need to make any modifications to the GPLd source, and/or if you release any modifications that you make to the GPLd code. (I think that the only people who'd have a problem with this are nutcase zealots like RMS).
Based on my understanding of the AltiVec unit, I think there are a few kernel level options that AltiVec may be useful for.
The most obvious is optimizations to the firewalling and routing code within the kernel. The AltiVec unit has a "permute" function that allows you to generate a 128 bit vector by picking and choosing words from two other 128 bit vectors. Aparently, Motorola has built a software router around a 7400 capable of doing software routing on multiple T3s. In this situation, the AltiVec unit simply becomes a glorified switching fabric.
The second optimization that I can see would be in kernel level encryption, for doing things like IPSec and/or encrypted filesystems. If I remember correctly, this use was sugested in one of the documents available on Motorola's web site.
Esentially, if you stop thinking of the AltiVec unit as a media processor, and think it more of a parallel processing unit, you can use it for a lot of different tasks.
Well, it looks like it is time to move to a different country. Most other western nations that I can think of (Most of Europe, Canada, Austrailia, New Zeland, ect..) don't have these stupid patent offices/laws, nor do they have the DMCA, UCITA, of (back when it was relevant) crypto export laws. Perhaps southern california is not the best place to conduct high-tech buisness after all. :)
Point taken, however in the experience I've had playing around with a Palm connected to the 'net, I was.. well.. rather underwealmed. Because of the screen size, surfing with the Palm is akin to looking at a mural a postage stamp square at a time.
In reality, I think the differences in my experiences has more to do with screen size, and CPU power then application availability, two things which the palm doesn't have much of. With a 320x480 screen, and a 162 mhz StrongARM, you can simply provide a richer experience to the user.
What it all comes down to, though, is what tool do you find works the best for you.
The other hand-held that hasn't been mentioned here yet is the Newton. Despite the fact that it has been discontinued, I still have yet to see a handheld that comes close to matching the elegance of the Newton.
In my opinion, it is what the other free efforts should be modling themselves after. The UI was written from the ground up to to be pen oriented, rather then being an adaptation of a desktop UI.
Also, it provides for tight integration between the OS and 3rd party applications. For example, the email program I use on my MP 2100 uses the standard contact information for getting its email addresses, and is treated by the system as simply another way to get information in and out of the system.
Thirdly, it has real, honest to God, handwriting recognition. Despite all the jokes that were made about it when the Newton was first released, it was perfected in the MP2x00. I can write at full speed, cursively, and it's accurate about 99% of the time.
Finally, the hand-held should be able to operate independently from a desktop PC. I should not have to sync it on a nightly basis with my host PC for it to be fully useable. The NewtonOS has a full TCP/IP stack, with communications being done over Ethernet, or PPP. The email, browser, ftp, irc, telnet, ect.. clients all use the standard protocols, and you can download new packages directly off of the 'net.
Anyway, if some one could produce a hand-held with these types of capabilities, I'd buy one in an instant. Until then, I'll stick with my MP2100.
My $0.02
FYI: Here in Vancouver, we pay $34.95 for 1.2Mbit/650Kbit ADSL. At current exchange rates, that's $22.46 USD. The kicker is that we're 2 hops off of Telus's Gigapop, with access to 5 major backbones.
One problem: At Apple, the Hardware sales finance the Operating Systems. Apple is primarily a hardware manufacturer, not an OS company. If they started selling OS X for Intel hardware, they'd die a slow and ugly death. Look at what happened to NeXT and Be after they stoped producing the hardware.