The guy who wrote the article clearly has no idea what he's talking about and probably has never seen how this is done in Unix: it uses 'magic'.:)
So that means that compiler front ends on UN*X systems run file or equivalent code on source files to figure out what language they're in, rather that treating.c files as C,.cpp/.cxx/etc. files as C++,.f files as FORTRAN, etc?:-)
In addition, whilst I think the major UN*X+X11 desktops might determine file types by doing file-style processing on the file, the OS X desktop, for better or worse, doesn't.
Presumably by "GSM-style service" you mean that the higher-layer protocols (well, until you get to IP; I have the impression that LTE's voice is done as VoIP) are more like those of GSM/UMTS than like those of cdmaOne/CDMA2000.
over a CDMA style network
I was under the impression that LTE used OFDM, not CDMA.
UMTS (3G) is, I think, more like "a GSM-style service over a CDMA-style network", using GSM-like protocols atop a (W)CDMA radio layer.
the cdmaOne (IS-95) and CDMA2000 (IS-2000) mobile phone protocol stacks, which use Code Division Multiple Access at the lowest layer.
"CDMA", as in "Code Division Multiple Access", doesn't use any particular frequencies. "cdmaOne" and "CDMA2000" have particular frequency ranges (plural) that they use.
GSM is a mobile phone standard that also includes a protocol stack; it uses Time Division Multiple Access at its lowest layer. GSM has particular frequency ranges (again, plural) that it uses. Some frequency ranges are used by both GSM and cdmaOne and/or CDMA2000.
WCDMA, is the broadband (3G) Frequency
WCDMA is a lowest-layer mechanism for mobile phones, using Code Division Multiple Access. It's one of the lowest layers that the Universal Mobile Telecommunications System uses. UMTS is more-or-less the 3G flavor of GSM, just as CDMA2000 is the 3G flavor of cdmaOne.
which is an "add on" to your mobile phone, so you can have things like Internet.
You don't need WCDMA - or any other flavor of 3G, such as CDMA2000 - to have Internet access on your phone. (I have an original iPhone, which has Internet access but doesn't support UMTS, just GSM, including the General Packet Radio Service (GPRS) and Enhanced Data Rates for GSM Evolution (EDGE); GPRS and EDGE are what are used for Internet access on 2G/2.5G GSM phones such as the original iPhone.)
and this can be added on to the phone regardless of whether you have CDMA or GSM.. this is up to what 3G technology (there are others than WCDMA) the carrier decide to use..
A 2G-only phone can't, in general, have 3G support added to it. (Perhaps some such phones have radio hardware that can support it with an upgrade to the firmware; I don't know of any that can, however.)
A particular carrier can upgrade to 3G on its network; GSM carriers will probably go to UMTS, and cdmaOne carriers will probably go to CDMA2000.
Here is what happened.. In the majority of the rest of the world.. GSM is the standard for "phone" frequency.. In the US, Qualcomm came up with the CDMA phone frequencies
Yes, although I'd say "protocol stack" rather than "frequency". Qualcomm were, I think, the first people to use Code Division Multiple Access; their "CDMA", however, involves more than just Code Division Multiple Access as the radio layer - it also involves higher-level protocols, which are different from the higher-level protocols used by GSM.
and the US carriers used it.. although AT&T (Cingular) switched to GSM, and the other GSM carrier "Voicestream" was bought by TMobile....
...so that's probably better stated as "some US carriers used it" - others, such as AT&T (who switched from Digital AMPS - which uses TDMA at the lowest layer, as does GSM - to GSM) and Voicestream", didn't.
CDMA "the phone freq technology" did not take off in the rest of the world.. In the rest of the world GSM is still king..
By and large true, although I think cdmaOne and CDMA
OS X apps have had 32-bit and 64-bit executables in their bundles for quite some time now. At least since 10.4, if not some point in 10.3 (it was definitely soon after the arrival of the first G5),
Nope - the output of "file" on the Mail executable on 10.5.8 is:
No 64-bit code there. Perhaps you're thinking of the libraries, some of which had 32-bit and 64-bit slices in 10.4, and most if not all of which had those slices in 10.5. They would still need their 32-bit PPC slices, even on Snow Leopard, for the benefit of PPC binaries running under Rosetta, although they could lose their 64-bit PPC slice, as Rosetta doesn't support 64-bit PPC binaries.
And, lastly, I know that OS X apps have always utilized Frameworks. But the point is, in Snow Leopard, Apple is utilizing Frameworks more than ever. I mean, how else can Mail.app shrink for 192 MB to 16 MB? It's not just the PPC code being excised.
Where do you get the size figures from? ls -l? size? Activity Monitor? Some other tool? I'm not seeing them.
The only valid way to determine whether, for example, Mail is using more libraries and frameworks is to run otool -L on the binaries and seeing whether it reports the Snow Leopard binary as being linked with more libraries and frameworks. That wouldn't tell you whether a given bit of functionality was moved from Mail to a framework, for use in other applications, but "Mail.app shrunk" won't tell you that, either.
I appreciate your clarifying things, but it is obscuring my main point--apps in Snow Leopard, and the OS itself, are VERY lean compared to any previous version of OS X, and there is a noticeable speed boost.
That wasn't your main point, it was the main point in the posting to which you were responding; the points in your article were claims as to the reason why that was the case:
"many (all?) of the Apple-supplied apps have been slimmed down from Universal Binaries to Intel-only executables" - as I noted, that can't be it, as they were slimmed down and then fattened up;
"there has been considerable "tightening up" of the Apple-supplied apps in that they use Frameworks (what's known as dynamic libraries or shared libraries on other OS's) as much as possible, instead of having nearly all code stuffed in their.app bundle" - speculative at best.
Yes, because many (all?) of the Apple-supplied apps have been slimmed down from Universal Binaries to Intel-only executables.
And then fattened up again to 32-bit Intel+64-bit Intel executables.
Also, there has been considerable "tightening up" of the Apple-supplied apps in that they use Frameworks (what's known as dynamic libraries or shared libraries on other OS's)
Well, not exactly. There are conventional dynamic shared libraries and there are frameworks. Conventional dynamic shared libraries are pretty much the same as they are on other UN*Xes; frameworks *include* one (or more) such shared libraries, but they also include other items, such as header files, nibs, etc.
as much as possible, instead of having nearly all code stuffed in their.app bundle.
That's not new in SnowLeopard - they've always been linked with shared libraries and frameworks.
Isn't NT Microsoft's CP/M-ish attempt to have a proprietary XENIX-like OS
No. NT was originally Microsoft's project to implement the OS/2 API atop a portable kernel that also supported implementing other APIs atop it as well; they eventually implemented the Win32 API atop it and treated that as the primary API. It also originally supported a character-mode OS/2 API and a character-mode POSIX API atop it as well.
We've never been at a point where we could run a full OS in a mobile profile.
What's your definition of "a full OS"? Presumably you're not counting iPhone OS, even though it's based on much of the same core OS code that Mac OS X is based on (heck, jailbreak it and you can even get a terminal emulator with a shell prompt). (I don't know whether the rumored Nokia N900 would be a phone or not, or whether it'll be Maemo-based.)
The one company to really take a unixish kernel and succeed with it? Apple. Many of your arguments could have been made about OS X and the BSD kernel it is based on. I suspect this will be similarly non recognizable to the other OS's using its kernel, and probably have a similar port ability. Taking all the obvious unix-like parts out of it really is required to get your grandmother to use it on her netbook. Think about explaining/usr/lib/etc to grandma. It requires a complete rewrite.
You are aware that Mac OS X has/usr and/etc and, although it doesn't have/lib, it does have/usr/lib, right? And, yes, they are used by parts of the system. (Heck, just about every process on the system uses/usr/lib/libSystem.dylib.)
OS X didn't "[take] all the obvious unix-like parts out of [Darwin]", it just didn't require the average user to know about them. (Is Ubuntu "relegated to a niche" because it hasn't done as good a job as OS X does at that? Or is there more to it than that?)
Open Solaris is *much* further away than OS X is -- it even disposes of the normal UNIX permissioning system, along with many other things.
To which "normal UNIX permissioning system" are you referring? Solaris might have ACLs in addition to permission bits, but 1) that's "in addition", not "instead of" and 2) OS X also has ACLs in addition to permission bits, so that's presumably not it.
Simple. Apple, as a content distributor probably have to guarantee that they can control where their itunes content is played.
Which certainly explains why I can burn a CD with music bought from iTunes (yes, even the old protected tracks), and then play the CD on any CD player - or put it into a computer and grab tracks from it.
"Two of the most famous products of Berkeley are LSD and Unix.
I don't think that is a coincidence."
Neither of them were, of course, invented at Berkeley; one might, at best, argue that Berkeley perfected both of them.:-)
Re:Worth thinking about
on
Unix Turns 40
·
· Score: 3, Interesting
I think that The Open Group would disagree with your overly broad definition of Unix
Then he needs to put a * in the right place; they can still whine about it, but they can't pretend the notion to which "Un*x" or "Unix-like" refers doesn't exist.
Jeeze what the hell man, we elected a socialist so you guys would stop hatin' on us.
Presumably you're Icelandic, as Iceland's the most recent country I know of to have elected somebody from a political party somebody could conceive of labeling as "socialist".; you're certainly not from the US, unless you're completely insane, as nobody sane from the US would assert that the US had recently elected a socialist (the last time I know of that a socialist was elected in the US was back in 2006).
Canada will be added to the Priority Watch List in 2009. The United States appreciates the high level of cooperation between our two governments in many important bilateral and multilateral IPR initiatives. The United States also welcomed the Government of Canada's reaffirmation earlier this year of its 2007 and 2008 commitments to improve IPR protection and enforcement. However, the Government of Canada has not delivered on these commitments by promptly and effectively implementing key copyright reforms. The United States continues to have serious concerns with Canada's failure to accede to and implement the WIPO Internet Treaties, which Canada signed in 1997. We urge Canada to enact legislation in the near term to strengthen its copyright laws and implement these treaties. The United States also continues to urge Canada to improve its IPR enforcement system to enable authorities to take effective action against the trade in counterfeit and pirated products within Canada, as well as curb the volume of infringing products transshipped and transiting through Canada. Canada's weak border measures continue to be a serious concern for IP owners. The United States hopes that Canada will implement legislative changes to provide a stronger border enforcement system by giving its customs officers the authority to seize products suspected of being pirated or counterfeit without the need for a court order. The provision of additional resources and training to customs officers and domestic law enforcement personnel would enhance IPR enforcement. The United States will continue to follow Canada's progress toward providing an adequate and effective IPR protection and enforcement regime, including near term accession to and implementation of the WIPO Internet Treaties and improved border enforcement.
It's not clear that it's "[claiming] that Canadian copyright and intellectual property laws are as bad as those found in China and Russia" - or Algeria or Argentina or Chile or India or Israel or Pakistan or Thailand or Venezuela, to give the other countries who appear after China and Russia in the list of Priority Watch countries in the report.
(I'm not saying that the report is justified in thumping Canada - or any of the other Priority Watch or Watch countries; I'm just suggesting that "Canadian copyright and intellectual property laws are as bad as those found in China and Russia" might be an overstatement of what the report is saying.)
Google was contemplating compiling JavaScript to pure native code in a story I read here on/. a while back, but how well they would maintain this for both x86 AND ARM remains another story,
Fat binaries are much bigger than binaries targeted for a single platform. (Hence the name). On the PC this may be ok, but on a cell phone, I seriously doubt you'll want to be transferring fat binaries to it.
...which is why the handheld device could tell the machine on the other end of the connection over which the binary is being transferred what instruction set architecture it wants, and the machine on the other end of the connection could thin the binary down and hand the device the thinned-down binary.
The new SDK will allow developers to control accessories attached to the dock adapter. I'm really hopeful someone will make a card reader
Or, at least, an IBM parallel channel adapter, so you can hook up one of these card readers. Unfortunately, the iPhone SDK's terms of service would probably disallow a port of Hercules, so you'd need to jailbreak in order to get that classic mainframe experience on the iPhone....
Slashdot Mirror
The guy who wrote the article clearly has no idea what he's talking about and probably has never seen how this is done in Unix: it uses 'magic'. :)
So that means that compiler front ends on UN*X systems run file or equivalent code on source files to figure out what language they're in, rather that treating .c files as C, .cpp/.cxx/etc. files as C++, .f files as FORTRAN, etc? :-)
In addition, whilst I think the major UN*X+X11 desktops might determine file types by doing file-style processing on the file, the OS X desktop, for better or worse, doesn't.
I guess if you still have one of the mice that came with Macs years ago you might still have to hold down control and click
Or one of those Macs that has a trackpad.
All of the 4G technologies
Which technologies are those? LTE? WiMAX? Others?
are basically a GSM style service
Presumably by "GSM-style service" you mean that the higher-layer protocols (well, until you get to IP; I have the impression that LTE's voice is done as VoIP) are more like those of GSM/UMTS than like those of cdmaOne/CDMA2000.
over a CDMA style network
I was under the impression that LTE used OFDM, not CDMA.
UMTS (3G) is, I think, more like "a GSM-style service over a CDMA-style network", using GSM-like protocols atop a (W)CDMA radio layer.
Just a simplified FYI.. CDMA, and GSM.. These are the phone frequencies..
No, not really. "CDMA" is a term that can either refer to
"CDMA", as in "Code Division Multiple Access", doesn't use any particular frequencies. "cdmaOne" and "CDMA2000" have particular frequency ranges (plural) that they use.
GSM is a mobile phone standard that also includes a protocol stack; it uses Time Division Multiple Access at its lowest layer. GSM has particular frequency ranges (again, plural) that it uses. Some frequency ranges are used by both GSM and cdmaOne and/or CDMA2000.
WCDMA, is the broadband (3G) Frequency
WCDMA is a lowest-layer mechanism for mobile phones, using Code Division Multiple Access. It's one of the lowest layers that the Universal Mobile Telecommunications System uses. UMTS is more-or-less the 3G flavor of GSM, just as CDMA2000 is the 3G flavor of cdmaOne.
which is an "add on" to your mobile phone, so you can have things like Internet.
You don't need WCDMA - or any other flavor of 3G, such as CDMA2000 - to have Internet access on your phone. (I have an original iPhone, which has Internet access but doesn't support UMTS, just GSM, including the General Packet Radio Service (GPRS) and Enhanced Data Rates for GSM Evolution (EDGE); GPRS and EDGE are what are used for Internet access on 2G/2.5G GSM phones such as the original iPhone.)
and this can be added on to the phone regardless of whether you have CDMA or GSM.. this is up to what 3G technology (there are others than WCDMA) the carrier decide to use..
A 2G-only phone can't, in general, have 3G support added to it. (Perhaps some such phones have radio hardware that can support it with an upgrade to the firmware; I don't know of any that can, however.)
A particular carrier can upgrade to 3G on its network; GSM carriers will probably go to UMTS, and cdmaOne carriers will probably go to CDMA2000.
Here is what happened.. In the majority of the rest of the world.. GSM is the standard for "phone" frequency.. In the US, Qualcomm came up with the CDMA phone frequencies
Yes, although I'd say "protocol stack" rather than "frequency". Qualcomm were, I think, the first people to use Code Division Multiple Access; their "CDMA", however, involves more than just Code Division Multiple Access as the radio layer - it also involves higher-level protocols, which are different from the higher-level protocols used by GSM.
and the US carriers used it.. although AT&T (Cingular) switched to GSM, and the other GSM carrier "Voicestream" was bought by TMobile....
...so that's probably better stated as "some US carriers used it" - others, such as AT&T (who switched from Digital AMPS - which uses TDMA at the lowest layer, as does GSM - to GSM) and Voicestream", didn't.
CDMA "the phone freq technology" did not take off in the rest of the world.. In the rest of the world GSM is still king..
By and large true, although I think cdmaOne and CDMA
Out of curiosity, what does file say if the binary is 64-bit?
On Leopard's libSystem, which is 4-way fat, it says:
so the 64-bit PowerPC slice is "ppc64" and the 64-bit x86 slice is "x86_64".
Apparently for 32-bit i386 it just says i386?
Yes.
OS X apps have had 32-bit and 64-bit executables in their bundles for quite some time now. At least since 10.4, if not some point in 10.3 (it was definitely soon after the arrival of the first G5),
Nope - the output of "file" on the Mail executable on 10.5.8 is:
$ file /Applications/Mail.app/Contents/MacOS/Mail
/Applications/Mail.app/Contents/MacOS/Mail: Mach-O universal binary with 2 architectures
/Applications/Mail.app/Contents/MacOS/Mail (for architecture i386): Mach-O executable i386
/Applications/Mail.app/Contents/MacOS/Mail (for architecture ppc7400): Mach-O executable ppc
No 64-bit code there. Perhaps you're thinking of the libraries, some of which had 32-bit and 64-bit slices in 10.4, and most if not all of which had those slices in 10.5. They would still need their 32-bit PPC slices, even on Snow Leopard, for the benefit of PPC binaries running under Rosetta, although they could lose their 64-bit PPC slice, as Rosetta doesn't support 64-bit PPC binaries.
And, lastly, I know that OS X apps have always utilized Frameworks. But the point is, in Snow Leopard, Apple is utilizing Frameworks more than ever. I mean, how else can Mail.app shrink for 192 MB to 16 MB? It's not just the PPC code being excised.
Where do you get the size figures from? ls -l? size? Activity Monitor? Some other tool? I'm not seeing them.
The only valid way to determine whether, for example, Mail is using more libraries and frameworks is to run otool -L on the binaries and seeing whether it reports the Snow Leopard binary as being linked with more libraries and frameworks. That wouldn't tell you whether a given bit of functionality was moved from Mail to a framework, for use in other applications, but "Mail.app shrunk" won't tell you that, either.
I appreciate your clarifying things, but it is obscuring my main point--apps in Snow Leopard, and the OS itself, are VERY lean compared to any previous version of OS X, and there is a noticeable speed boost.
That wasn't your main point, it was the main point in the posting to which you were responding; the points in your article were claims as to the reason why that was the case:
Yes, because many (all?) of the Apple-supplied apps have been slimmed down from Universal Binaries to Intel-only executables.
And then fattened up again to 32-bit Intel+64-bit Intel executables.
Also, there has been considerable "tightening up" of the Apple-supplied apps in that they use Frameworks (what's known as dynamic libraries or shared libraries on other OS's)
Well, not exactly. There are conventional dynamic shared libraries and there are frameworks. Conventional dynamic shared libraries are pretty much the same as they are on other UN*Xes; frameworks *include* one (or more) such shared libraries, but they also include other items, such as header files, nibs, etc.
as much as possible, instead of having nearly all code stuffed in their .app bundle.
That's not new in SnowLeopard - they've always been linked with shared libraries and frameworks.
Isn't NT Microsoft's CP/M-ish attempt to have a proprietary XENIX-like OS
No. NT was originally Microsoft's project to implement the OS/2 API atop a portable kernel that also supported implementing other APIs atop it as well; they eventually implemented the Win32 API atop it and treated that as the primary API. It also originally supported a character-mode OS/2 API and a character-mode POSIX API atop it as well.
We've never been at a point where we could run a full OS in a mobile profile.
What's your definition of "a full OS"? Presumably you're not counting iPhone OS, even though it's based on much of the same core OS code that Mac OS X is based on (heck, jailbreak it and you can even get a terminal emulator with a shell prompt). (I don't know whether the rumored Nokia N900 would be a phone or not, or whether it'll be Maemo-based.)
The one company to really take a unixish kernel and succeed with it? Apple. Many of your arguments could have been made about OS X and the BSD kernel it is based on. I suspect this will be similarly non recognizable to the other OS's using its kernel, and probably have a similar port ability. Taking all the obvious unix-like parts out of it really is required to get your grandmother to use it on her netbook. Think about explaining /usr /lib /etc to grandma. It requires a complete rewrite.
You are aware that Mac OS X has /usr and /etc and, although it doesn't have /lib, it does have /usr/lib, right? And, yes, they are used by parts of the system. (Heck, just about every process on the system uses /usr/lib/libSystem.dylib.)
OS X didn't "[take] all the obvious unix-like parts out of [Darwin]", it just didn't require the average user to know about them. (Is Ubuntu "relegated to a niche" because it hasn't done as good a job as OS X does at that? Or is there more to it than that?)
Open Solaris is *much* further away than OS X is -- it even disposes of the normal UNIX permissioning system, along with many other things.
To which "normal UNIX permissioning system" are you referring? Solaris might have ACLs in addition to permission bits, but 1) that's "in addition", not "instead of" and 2) OS X also has ACLs in addition to permission bits, so that's presumably not it.
Simple. Apple, as a content distributor probably have to guarantee that they can control where their itunes content is played.
Which certainly explains why I can burn a CD with music bought from iTunes (yes, even the old protected tracks), and then play the CD on any CD player - or put it into a computer and grab tracks from it.
Your proclamation betrays your ignorance. Provably correct methods are 100% foolproof. Very few things, however, are provably correct.
And, of course, there is the risk of an error in the proof. Possibly a negligible risk, but it can happen.
Finally, big sigs with ASCII art and geek code blocks. The bigger the better. True masters have sigs bigger than their actual post.
Like Death Star, War Lord of the West?
http://web.mit.edu/~simsong/www/ugh.pdf
"Two of the most famous products of Berkeley are LSD and Unix. I don't think that is a coincidence."
Neither of them were, of course, invented at Berkeley; one might, at best, argue that Berkeley perfected both of them. :-)
I think that The Open Group would disagree with your overly broad definition of Unix
Then he needs to put a * in the right place; they can still whine about it, but they can't pretend the notion to which "Un*x" or "Unix-like" refers doesn't exist.
Linux's chroot is actually BSD's chroot. Bill Joy invented it.
BSD's chroot is actually V7's chroot. Ken, Dennis, and company invented it.
Because GTK+ comes with a HIG included... not.
GNOME, on the other hand, does have an HIG.
Jeeze what the hell man, we elected a socialist so you guys would stop hatin' on us.
Presumably you're Icelandic, as Iceland's the most recent country I know of to have elected somebody from a political party somebody could conceive of labeling as "socialist".; you're certainly not from the US, unless you're completely insane, as nobody sane from the US would assert that the US had recently elected a socialist (the last time I know of that a socialist was elected in the US was back in 2006).
What the 2009 Special 301 Report says about Canada is:
Canada will be added to the Priority Watch List in 2009. The United States appreciates the high level of cooperation between our two governments in many important bilateral and multilateral IPR initiatives. The United States also welcomed the Government of Canada's reaffirmation earlier this year of its 2007 and 2008 commitments to improve IPR protection and enforcement. However, the Government of Canada has not delivered on these commitments by promptly and effectively implementing key copyright reforms. The United States continues to have serious concerns with Canada's failure to accede to and implement the WIPO Internet Treaties, which Canada signed in 1997. We urge Canada to enact legislation in the near term to strengthen its copyright laws and implement these treaties. The United States also continues to urge Canada to improve its IPR enforcement system to enable authorities to take effective action against the trade in counterfeit and pirated products within Canada, as well as curb the volume of infringing products transshipped and transiting through Canada. Canada's weak border measures continue to be a serious concern for IP owners. The United States hopes that Canada will implement legislative changes to provide a stronger border enforcement system by giving its customs officers the authority to seize products suspected of being pirated or counterfeit without the need for a court order. The provision of additional resources and training to customs officers and domestic law enforcement personnel would enhance IPR enforcement. The United States will continue to follow Canada's progress toward providing an adequate and effective IPR protection and enforcement regime, including near term accession to and implementation of the WIPO Internet Treaties and improved border enforcement.
It's not clear that it's "[claiming] that Canadian copyright and intellectual property laws are as bad as those found in China and Russia" - or Algeria or Argentina or Chile or India or Israel or Pakistan or Thailand or Venezuela, to give the other countries who appear after China and Russia in the list of Priority Watch countries in the report.
(I'm not saying that the report is justified in thumping Canada - or any of the other Priority Watch or Watch countries; I'm just suggesting that "Canadian copyright and intellectual property laws are as bad as those found in China and Russia" might be an overstatement of what the report is saying.)
Here in Canada we have a Communist party. In the US it is illegal.
O RLY?.
she knows what a monkey does for sex. she has two kids.
A monkey? In the White House? You're probably thinking of the previous occupant.
Google was contemplating compiling JavaScript to pure native code in a story I read here on /. a while back, but how well they would maintain this for both x86 AND ARM remains another story,
Yeah, God knows that nobody who has a browser that runs on x86 and a variant that runs on ARM has ever thought about compiling JavaScript to native code, and that Google haven't done more than just contemplate making a JavaScript engine that translates to machine code, much less made versions for both x86 and ARM, and, of course, the Firefox people haven't done anything like that, either.
Fat binaries are much bigger than binaries targeted for a single platform. (Hence the name). On the PC this may be ok, but on a cell phone, I seriously doubt you'll want to be transferring fat binaries to it.
...which is why the handheld device could tell the machine on the other end of the connection over which the binary is being transferred what instruction set architecture it wants, and the machine on the other end of the connection could thin the binary down and hand the device the thinned-down binary.
The new SDK will allow developers to control accessories attached to the dock adapter. I'm really hopeful someone will make a card reader
Or, at least, an IBM parallel channel adapter, so you can hook up one of these card readers. Unfortunately, the iPhone SDK's terms of service would probably disallow a port of Hercules, so you'd need to jailbreak in order to get that classic mainframe experience on the iPhone....