Hey, let's cut this guy some slack: sure he's got all his facts wrong, but his conclusion is actually correct!
1) kernel design: As everyone who actually has studied system architecture knows, MACH is not a retrograde version of unix, it the most advanced superset of unix design principles (indeed, it is linux which is widely regarded as being DOA from an engineering perspective). And as for being dead - well, like it or not, NT is based on a mK: can't get more maindtream than that!
As a result, OS/X is widely parised by conaisseurs as representing the epitome of kernel elegance & extensibility:
2) programming model: As everyone who has studied language design knows, ObjectiveC is not some retrograde version of C++, which is ultimately little more than a statically typed set of pseudo-object extentions bolted onto a procedural language; ObjectiveC has a front-end that looks & feels like a pure OOP, and a back-end which emits C. Therefore it does not have make all the comprimises that C++ must in order to attain OOP, yet retains all the 'virtues' of still producing low-level code that lots of programmers were originally trained in.
As a result, NextStep/Cocoa is widely regarded by conaisseurs as the preeminent development platform ever created.
Here is the key document - a small gem, which could easily serve as self-contained textbook for an Progamming 101:
Noneltheless, despite getting all his facts severly wrong, this fellows's consclusions have a scary foundation.
Apple has indeed before taken developers down the garden path before (and discarding truly amazing stuff like Taligent).
And they appear to be doing it all over again!!!
http://www.stepwise.com/articles/2000-06-21.01.htm l
So just as Apple finally has the definitive environment... including the Opensource implementations - Darwin & http://www.gnustep.org... it starts shooting itself in the foot again.
Once bitten, twice shy!!
So our ignorant friend should not necessarily be ignored.
A word to the wise!
ps: if the url's are not precisely correct, the root is:
http://developer.apple.com/techpubs/macosx/macosx. html
An Intel port does indeed depend upon firmware issues. But contrary to AC's assertion, it is entirely plausible to imagine x86 firmware that fits into a rational business model for Apple. The firmware used on PowerPC motherboards is called OpenFirware. OpenFirmware (based on Sun technology) is a really elegant & very powerful bootime environment for what is called BIOS in Inteland. It actually runs its own, full-fledged OS-independent interpreter (in Forth, which is a stack-based language ideally suited for creating device-drivers). Indeed, just to show the power of OpenFirmware, some guy hacked up a version of the game PONG which runs entirely in the OpenFirware console! (ie no "OS" is loaded). Try getting the measely, brittle Intel BIOS to be a programmable environment!! Anyways, "all" that is required for an Intel-based product (not just a "port") from Apple is to create x86 bindings for OpenFirware (along with re-bindings for the standard Apple control logic chipset, aka UMA). Because OpenFirmware isolates hardware dependencies, Apple could -if it wanted to- have its own Intel motherboard without being dragged down into the mire of supporting all the low-quality motherboards which proliferate in Inteland. Moreover, it could chose to license this motherboard implementation to AMD (who is emerging a key, stealth partner in manfacturing G4's). As to the ability of a UMA/x86 motherboard to boot an "alternate" OS (eg native Windows instead of Darwin), it would be just be a business model decision - not a technical decsion- as to whether UMA/x86 drivers for Windoze would be created... which gives Apple some interesting leverage with Microsoft in the post-DOJ world;) Since OpenFirware *per se* is not proprietary (just the bindings, ie implementation, to particular chipsets), it is entirely conceivable that other charter OpenFirmware OEM's (ie SUN & IBM) may also decide that OpenFirmware/x86 might be a controlled way of supporting/differentiating OpenSource on x86. And because Darwin is based on a modern microkernel, it can support other (unix) personalities besides BSD... yup, that means Linux. So OpenFirware/x86 kills two birds with one stone - it offers tremendous OEM finesse in the business model. In conslusion: firware is not an "obstacle" for Darwin; it is an opportunity for Apple.
Slashdot Mirror
1) kernel design: As everyone who actually has studied system architecture knows, MACH is not a retrograde version of unix, it the most advanced superset of unix design principles (indeed, it is linux which is widely regarded as being DOA from an engineering perspective). And as for being dead - well, like it or not, NT is based on a mK: can't get more maindtream than that!
As a result, OS/X is widely parised by conaisseurs as representing the epitome of kernel elegance & extensibility:
http://www.arstechnica.com/reviews/2q00/macos-x-dp 4/macos-x-dp4-1.html
And here are the key documents:
http://developer.apple.com/techpubs/macosx/system/ documentation/developer/systemoverview.p df
http://developer.apple.com/techpubs/macosx/system/ documentattion/developer/kernel/kernelen vironment.pdf
2) programming model: As everyone who has studied language design knows, ObjectiveC is not some retrograde version of C++, which is ultimately little more than a statically typed set of pseudo-object extentions bolted onto a procedural language; ObjectiveC has a front-end that looks & feels like a pure OOP, and a back-end which emits C. Therefore it does not have make all the comprimises that C++ must in order to attain OOP, yet retains all the 'virtues' of still producing low-level code that lots of programmers were originally trained in.
As a result, NextStep/Cocoa is widely regarded by conaisseurs as the preeminent development platform ever created.
Here is the key document - a small gem, which could easily serve as self-contained textbook for an Progamming 101:
http://developer.apple.com/techpubs/macosx/system/ documentation/developer/cocoa/objectivec .pdf
Noneltheless, despite getting all his facts severly wrong, this fellows's consclusions have a scary foundation.
Apple has indeed before taken developers down the garden path before (and discarding truly amazing stuff like Taligent).
And they appear to be doing it all over again!!!
http://www.stepwise.com/articles/2000-06-21.01.htm l
So just as Apple finally has the definitive environment ... including the Opensource implementations - Darwin & http://www.gnustep.org ... it starts shooting itself in the foot again.
Once bitten, twice shy!!
So our ignorant friend should not necessarily be ignored.
A word to the wise!
ps: if the url's are not precisely correct, the root is:
http://developer.apple.com/techpubs/macosx/macosx. html
An Intel port does indeed depend upon firmware issues. But contrary to AC's assertion, it is entirely plausible to imagine x86 firmware that fits into a rational business model for Apple. The firmware used on PowerPC motherboards is called OpenFirware. OpenFirmware (based on Sun technology) is a really elegant & very powerful bootime environment for what is called BIOS in Inteland. It actually runs its own, full-fledged OS-independent interpreter (in Forth, which is a stack-based language ideally suited for creating device-drivers). Indeed, just to show the power of OpenFirmware, some guy hacked up a version of the game PONG which runs entirely in the OpenFirware console! (ie no "OS" is loaded). Try getting the measely, brittle Intel BIOS to be a programmable environment!! Anyways, "all" that is required for an Intel-based product (not just a "port") from Apple is to create x86 bindings for OpenFirware (along with re-bindings for the standard Apple control logic chipset, aka UMA). Because OpenFirmware isolates hardware dependencies, Apple could -if it wanted to- have its own Intel motherboard without being dragged down into the mire of supporting all the low-quality motherboards which proliferate in Inteland. Moreover, it could chose to license this motherboard implementation to AMD (who is emerging a key, stealth partner in manfacturing G4's). As to the ability of a UMA/x86 motherboard to boot an "alternate" OS (eg native Windows instead of Darwin), it would be just be a business model decision - not a technical decsion- as to whether UMA/x86 drivers for Windoze would be created ... which gives Apple some interesting leverage with Microsoft in the post-DOJ world ;) Since OpenFirware *per se* is not proprietary (just the bindings, ie implementation, to particular chipsets), it is entirely conceivable that other charter OpenFirmware OEM's (ie SUN & IBM) may also decide that OpenFirmware/x86 might be a controlled way of supporting/differentiating OpenSource on x86. And because Darwin is based on a modern microkernel, it can support other (unix) personalities besides BSD ... yup, that means Linux. So OpenFirware/x86 kills two birds with one stone - it offers tremendous OEM finesse in the business model. In conslusion: firware is not an "obstacle" for Darwin; it is an opportunity for Apple.