All Mozilla needs is Lisp (or Guile) embedded interpreter linked to XPCOM - then people around the world will create a feature flood, like it happened to (X)Emacs in its days. As for JavaScript... That's not the real language - many programmers abondon it.
Or maybe more general idea can work better: language extension mechanism, like PostgreSQL has got. So, developer will add their languages as they like.
BSD has a monolytic-type kernel. Darwin and OS X are micro-kernel (Mach) based ones. That's a very big difference, don't you think?
If the OS has ported some (less than 50%) of BSD utils to work on atop of non-BSD kernel - no way it can be called BSD OS.
I use some (in fact a lot, including TeX and X11) of BSD and Linux tools under Cygwin on my NT office box - will you call that box BSD or Linux?
Besides, I don't like either of them: neither BSD nor Mac OS. Linux is the way to go on my servers, do you like it or not. It supports more (and better!) hardware device models and CPUs than BSD does. Not on the paper - on real desktops and servers. Linux is more than BSD (again BSD, not Mac OS!) recognized by 3rd party software vendors. You cannot call Apple as a 3rd party vendor b/c Apple did not port Aqua to FreeBSD - they have stolen BSD tools to work in Mac OS X, and that is not 3rd-party vendorship. Besides, people prefer to Beowulf on Linux rather than on BSD - that's also important for me.
As for Mac OS... I guess you can keep using it for your Photoshop - that's what it is designed for, isn't it?
What's wrong with Gentoo? Don't you like it? Then take another Linux distro. I am choosing Gentoo b/c I can clone it to my own distro, specialized for my server needs. And I can do it easier and better than with any other distros.
And don't even try to remind me about BSD. First of all, I've never heard any positive feedbacks from any PPC (IBM, Apple or embedded) users triing BSD on PPC (I don't count any official press releases - only real user feedbacks). Second, we all slashdotters know that BSD is dead, right? No need to explain why I don't count any Mac OS, AIX (anything proprietary) at all:)
I wonder if it's possible to create database clients similar to MS Access based on Mozilla/Gecko/XPCOM code. And if the answer is positive then why I can not find anything like that?
The book is good and interesting, but it reminds me another book, Programming Jabber: a lot of examples in the book, and no available examples in real life (besides Jabberd itself).
Apple won't be talking about the 970 anytime soon.
I hope IBM will make 970-based PPC with price affordable for programmers. And I hope it will be compatible with Linux/PPC kernel. Then we may buy such thing to run Gentoo on it for server-side applications.
I wonder whether IBM will sell it with Linux or AIX pre-installed.
it is very hard to find a Mac OS X capable system without an included license for some Mac OS
I am not sure what you mean in the sentence above.
I think it is unfair to compare Mac OS X to non-Mac OSes. Instead, it is fair to compare Linux/PPC to MAC OS X - if we compare Operating Systems.
However, if you want to compare hardware, x86 vs Mac then let's compare the combination of x86 with all OSes compatible with it (Win*, Linux, *BSD, BeOS, OS/2) versus the combination of Mac PPC with all OSes compatible with it (Mac OS 9, Mac OS X, Linux, BeOS, *BSD).
First comparison will give us real OS versus whistles. Second one - poorly-designed free-of-vendor-choice low-price easy-to-upgrade-hardware plaform versus well-designed very-expensive hard-to-upgrade-hardware from-monopolist-vendor platform.
Now compbine both results and see what you have on Mac OS X: broken BSD, broken Next, full of whistles, proprietary API, fee for each update, hard to maintain etc etc etc
I think that in a year or so software vendors (Adobe, Macromedia and others) will recognize Linux/PPC and then you can count days left for Mac OS X. Add here the potential pressure from IBM, another Linux lover. The only choice Steve Jobs has is to give up and to disclose that Apple is a bad OS vendor/developer.
Since 1957 Lisp stays as the most innovative and the most inspiring language. Other languages come and die. Some stay, but don't inspire. Some were born dead or old. Lisp stays. It evolves but it still same Lisp. It might not be able to say "spoon", only "thpoon", but at least it discovers that there is no spoon. And it doesn't matter how to say about something that does not exist!
In 12 months people will finally realize that Java is not the way to go. IBM will push primary Python, while other companies will try to move from Java to Erlang.
In fact, IBM, begin pushing Python as a primary language, will try Haskell in same way as they try Python today secondary after Java. IBM will even create a partnership with Microsoft to develop Haskel#. At the same, time Ericson competitors will try Oz and Mercury.
Of course, Web Sphere will die in its current Java- based shape - perhaps IBM will adapt the best Web Sphere ideas and libraries on atop of Zope. Will it be called Zope Sphere? I don't know yet, but I sure IBM will design either EPC (Enterprise Python Containers, sort of EJB) or PINE (Python Inter-Network Environment, sort of JINI). Jython will be very popular in a path of migration from EJB and JINI.
Both Oracle and IBM, suffuring from over-patched proprietary source code, will try to adapt PostgreSQL implementation ideas in Oracle whatever-i and in DB/2. Will it be the war for PostgreSQL? I am not sure. But I know that Microsoft will not be there until they repeate after Apple and adapt BSD for the next generation of Windows (will it be named TNG?), breaking out from Intel boundaries and turning more Linux users to MS BSD campus.
By the way, IBM will anonunce "no new AIX releases any more" and release IBM Linux/PPC based on Gentoo. Red Hat will be de-listed from the stock exchange. Sun will adapt Suse and consider it instead of Solaris. HPQ will adapt Debian for all HP-Compaq-DEC zoo. The war between Linux and BSD fanatics will become the war between IBM/Sun/HPQ camp and MS/Apple camp. Did I mention that Apple will establish a consortium "Proprietary BSD"?
DOJ will require to open the documentation for all file formats which are used to exchange documents between customers of public available software products. In its own turn Microsoft will anounce full support of Open Office [XML] file format.
After AOL will switch most of users to Mozilla, the top fun in Open Source Community will shift to X-Smile - new legacy-free XML browser. Mozilla developers will disclose that their code is too heavy and too hardcoded to support pure XML content different from HTML based scenarios.
By the way, alternately, Python people may forget about being lazy and keep better backward compatibility of the interpreter. Python is used now around the world for serious stuff - let's stop experiments (even I like new stuff in 2.2) and let's think about users!
I guess all software, which depends on Python 1.5, is dead in terms that no one update it anymore (otherwise - why not update the first line?). Red Hat people should forget about being lazy and change first line in all such obsolete stuff.
That's not a big deal to create a script, which will go through all filesystem, find all "#!/usr/bin/python", change them to #!/usr/bin/python1.5. Once it's done - install Python 2.2 as defult python.
Untill it's done - I am moving to Gentoo. Well, I begin to love Gentoo and I may stay with it anyway:)
First of all, the set of predifined users and groups should be updated.
I think it's time to include groups into group member lists - let's play trees. Well, with multiple group assignment it will be actually a DAG.
Let's use profiles with pre-assigned user parameters (and scripts) even on group level. Of course such parameters should be inherited (another question - overriden?) by users traversing their DAGs. At run-time or at new user constructor - it is up-to implementations. That's not difficult even with bash scipts and will be just singing with Python as a profile scripting language.
Filesystem permission should be assigned to roles in a way similar like today user-group-others. Of course, "user" becomes a role "owner", "group" goes for list of roles the user "belongs", "others" becomes a role "guess" or "anonymous".
Besides grouping users and their permissions, roles can be used to group user profile fragments. Again, user's profile inherits all environment variables and login scripts from all roles it belongs.
Of course the relationsip "belongs" (in its own turn either "part-of" or "sort-of") is not the only type of relationships between nodes. But different applications (like FS driver above or login script) may use (interpretate) different types of relationships with different type of inheritance inference.
We don't live without networks, remember? RBAC DAG should be distrbuted. User profiles should roamed. Domains (aka NT domains) should be nested.
bla-bla-bla. We all were reading about RBAC at some time. Now it might be time to discuss how make the system where security level is a subject of configuration. You want a simple home desktop - most of such RBAC configuration goes as default. You need a regular corporate desktop - choose the configuration type "aka NT domain" on the server and workstation. You are required to install a box with C2 level of security - you need RBAC (at least) for it.
back to "Linux TNG" subject. There are lots of theoretical research done. There is (are) some stand with reference implementation form NIST. There is SE-Linux project (still alive?) from NSA. there is no Linux standard for RBAC.
That should be something like LSB, defining default user roles, their default profiles and permissions. Such standard will help to accelerate Linux RBAC implementation in the kernel, encourage Linux application developers to design keeping Linux RBAC in mind and assure corporate decision makers that Linux is targeting corporate networks seriously.
What you are saying is the direct way to anarchy and chaos.
US has the biggest military forces and it uses them to change the govt in other country or to make a pressure to get immunity from International Crime Court, which is supposed to deal with genocide, what is exactly what US afraid of being presecuted for (why?).
The other countries should shut up even when their neighbours are more dangerous than Afganistan. Just shut up and do not interfere with big games of big guys.
Moreover, those big guys in US govt make everything possible to assure that ALL other coutries are weaker than US. Thus, it would be possible to dictate them everything and to dominate the world.
Of course any interests of local peoples are ignored. Even the life of own american people (as well europeans) is ignored and ready for being sacraficed - it is no problem that Afganistan increased its drag export (major part of it is going to US and EU) - it is even good for the business of some big guys who are behind US govt.
As for washing the brain - no problem at all. At the time when US refused to keep peacekeepers further on Balkans, all European newspapers where full of headlines about it - CNN has been keeping a silence in 4 days. I guess 4 days american people did not know how ugly its govt is again.
I guess the biggest chaos is in conflict of interests of big US corps and the face the US govt must leep badly protecting such interests in international hidden games.
It is too hard to enforce
It will much easier if US will show a good example of ebeying international laws.
Creating international "bully forces" to impose Western values on other countries isn't going to do anything but cause more people worldwide to become "terrorists" who are out to "kill the West."
If you, my friend, will leave for awhile you warm seat and see other people outside of US, you will hear the opposite point - people are angry that US ignore international laws.
By the way, the goal of peacekeepers is not to bring western values, but to bring the peace - the function is known from the history of all continents. I mean to bring the peace to people first all, only after that thinking about the peace for govts.
International law should be stripped down to cover war crimes like genocide. It shouldn't have any place outside the of things like that.
A genocide of the govt against its own people must be the subject of such curt as well as war crimes.
It will take better international laws, better international police and better respect of international laws by ALL goverments to get such pirates, narko-business, terrorism and other internation crime nailed down. All == including US govt.
Unfortunately - forget it for awhile. US goverment questioned: "What International Crime Court? There is no such thing if US soldgers do not have any immunity from it.". And then claimed: "I am the international law!". And actually proved it: "I was the International law for Afganistan. I am the International law for Iraq. I will be the International law for all evil nations."
Unfortunately, that won't work. It did not work in Afganistan - the production of heroin and opium has been trippled compared to what it was with taliban. I doubt it will work with Iraq and Saudi. Bad example creates another bad example. Malasyan govt ecentually will give up. As Latin America govts gave up against narko-business. Why should they respect international laws if US does not?
Fortunately, it won't last forever. Switching the attention of american people from economy problems to the war cannot last forever and creates even stronger internal economy problems. Sooner or later USA will be back in 1929 and the history will repeat itself. I hope, after all USA will have some internal law forcing their goverment to respect international laws. And actually improve it - if US disagree with UN they should suggest how to reform UN, instead of boykoting it (as it was before) or ignoring (as it is now).
And only after that the war against CD pirates will bring any positive results.
NetBSD runs on more architectures than Linux, and they are often more actively maintained
PowerPC is the second (after Intel x86) on desktops and I don't see lot of PCC users enjoying Net(or other)BSD - all Macers, who hate Mac OS (9 or X), install Linux/PPC (YDL, gentoo or Debian or others). You are wrong here.
About USB - BSD might support USB as a bus, but what about devices connected to the bus? The list is short and it grows slowly in BSD. Unlike Linux.
the configuration of X is identical under BSD and Linux
Configuration files are the same. The process of hardware type and parameters discovery is what matters.
Porting applications from commercial Unix to Linux is much more of a headache than porting to BSD.
Again - wrong. Oracle, Sun and other prefer to port to Linux at furst.
Firstly, there's no standard Linux distribution, and all the distributions have subtly different patched kernel versions.
Even this is wrong: There is one tree of Linux kernel against several BSD ones. And noone, besides you, mentioned any problem with different kernel patches in different distributions. I cannot say same about the constant headache: "Which BSD should I choose? FreeBSD? NetBSD? OpenBSD?"
Actually I like the fact of such interruptions - it may convience more and more managers, architects and developers to use asyncronous transactional messaging protocols (like JMS or SOAP or even SMTP if with confirmation) vs decent client server ones (CORBA, proprietary).
I guess it will be much faster to complete the port of Linux to SGI/MIPS. By the time it will be done SGI will have same OS on both platforms, namely Linux:)
Slashdot Mirror
Maybe it's time to rename it to PPCworld and remind the world that Mac is just one of PPC-based computers. Even not the best, IMHO.
I've got it - there is not public transport, no taxi, no sidewalks in Boston. All execs from USA should drive themselves crossing all country.
All execs from the rest-of-the-world must rent-a-car. All Mac-World is organized as "drive-through".
People without cars are not accepted to the exposition.
I wonder if GTK Mozilla is GPLized or not. Any idea?
Or maybe more general idea can work better: language extension mechanism, like PostgreSQL has got. So, developer will add their languages as they like.
BSD has a monolytic-type kernel. Darwin and OS X are micro-kernel (Mach) based ones. That's a very big difference, don't you think?
If the OS has ported some (less than 50%) of BSD utils to work on atop of non-BSD kernel - no way it can be called BSD OS.
I use some (in fact a lot, including TeX and X11) of BSD and Linux tools under Cygwin on my NT office box - will you call that box BSD or Linux?
Besides, I don't like either of them: neither BSD nor Mac OS. Linux is the way to go on my servers, do you like it or not. It supports more (and better!) hardware device models and CPUs than BSD does. Not on the paper - on real desktops and servers. Linux is more than BSD (again BSD, not Mac OS!) recognized by 3rd party software vendors. You cannot call Apple as a 3rd party vendor b/c Apple did not port Aqua to FreeBSD - they have stolen BSD tools to work in Mac OS X, and that is not 3rd-party vendorship. Besides, people prefer to Beowulf on Linux rather than on BSD - that's also important for me.
As for Mac OS... I guess you can keep using it for your Photoshop - that's what it is designed for, isn't it?
And don't even try to remind me about BSD. First of all, I've never heard any positive feedbacks from any PPC (IBM, Apple or embedded) users triing BSD on PPC (I don't count any official press releases - only real user feedbacks). Second, we all slashdotters know that BSD is dead, right? No need to explain why I don't count any Mac OS, AIX (anything proprietary) at all :)
The book is good and interesting, but it reminds me another book, Programming Jabber: a lot of examples in the book, and no available examples in real life (besides Jabberd itself).
I hope IBM will make 970-based PPC with price affordable for programmers. And I hope it will be compatible with Linux/PPC kernel. Then we may buy such thing to run Gentoo on it for server-side applications.
I wonder whether IBM will sell it with Linux or AIX pre-installed.
I am not sure what you mean in the sentence above.
I think it is unfair to compare Mac OS X to non-Mac OSes. Instead, it is fair to compare Linux/PPC to MAC OS X - if we compare Operating Systems.
However, if you want to compare hardware, x86 vs Mac then let's compare the combination of x86 with all OSes compatible with it (Win*, Linux, *BSD, BeOS, OS/2) versus the combination of Mac PPC with all OSes compatible with it (Mac OS 9, Mac OS X, Linux, BeOS, *BSD).
First comparison will give us real OS versus whistles. Second one - poorly-designed free-of-vendor-choice low-price easy-to-upgrade-hardware plaform versus well-designed very-expensive hard-to-upgrade-hardware from-monopolist-vendor platform.
Now compbine both results and see what you have on Mac OS X: broken BSD, broken Next, full of whistles, proprietary API, fee for each update, hard to maintain etc etc etc
I think that in a year or so software vendors (Adobe, Macromedia and others) will recognize Linux/PPC and then you can count days left for Mac OS X. Add here the potential pressure from IBM, another Linux lover. The only choice Steve Jobs has is to give up and to disclose that Apple is a bad OS vendor/developer.
Does any one know what is that Linux for Power4 from IBM based on? Is it Gentoo, some clone of YDL (or RH) or Debian or what?
Writing it from Gentoo installed on my G4 I wonder: what do you mean?
Since 1957 Lisp stays as the most innovative and the most inspiring language. Other languages come and die. Some stay, but don't inspire. Some were born dead or old. Lisp stays. It evolves but it still same Lisp. It might not be able to say "spoon", only "thpoon", but at least it discovers that there is no spoon. And it doesn't matter how to say about something that does not exist!
- Lisp - catapult launching catapults, each launching catpults...
- Scheme - nice b/w picture of the catapult inspiring other artists to create similar pictures.
And some missed ones:Since XSL, XQuery and FlatCurry.
In fact, IBM, begin pushing Python as a primary language, will try Haskell in same way as they try Python today secondary after Java. IBM will even create a partnership with Microsoft to develop Haskel#. At the same, time Ericson competitors will try Oz and Mercury.
Of course, Web Sphere will die in its current Java- based shape - perhaps IBM will adapt the best Web Sphere ideas and libraries on atop of Zope. Will it be called Zope Sphere? I don't know yet, but I sure IBM will design either EPC (Enterprise Python Containers, sort of EJB) or PINE (Python Inter-Network Environment, sort of JINI). Jython will be very popular in a path of migration from EJB and JINI.
Both Oracle and IBM, suffuring from over-patched proprietary source code, will try to adapt PostgreSQL implementation ideas in Oracle whatever-i and in DB/2. Will it be the war for PostgreSQL? I am not sure. But I know that Microsoft will not be there until they repeate after Apple and adapt BSD for the next generation of Windows (will it be named TNG?), breaking out from Intel boundaries and turning more Linux users to MS BSD campus.
By the way, IBM will anonunce "no new AIX releases any more" and release IBM Linux/PPC based on Gentoo. Red Hat will be de-listed from the stock exchange. Sun will adapt Suse and consider it instead of Solaris. HPQ will adapt Debian for all HP-Compaq-DEC zoo. The war between Linux and BSD fanatics will become the war between IBM/Sun/HPQ camp and MS/Apple camp. Did I mention that Apple will establish a consortium "Proprietary BSD"?
DOJ will require to open the documentation for all file formats which are used to exchange documents between customers of public available software products. In its own turn Microsoft will anounce full support of Open Office [XML] file format.
After AOL will switch most of users to Mozilla, the top fun in Open Source Community will shift to X-Smile - new legacy-free XML browser. Mozilla developers will disclose that their code is too heavy and too hardcoded to support pure XML content different from HTML based scenarios.
PGP will be saved by governments. Perhaps.
The solution is to check 3rd party installed scripts and fix them right after installation. Something like applying a pacth for ports in FreeBSD.
By the way, alternately, Python people may forget about being lazy and keep better backward compatibility of the interpreter. Python is used now around the world for serious stuff - let's stop experiments (even I like new stuff in 2.2) and let's think about users!
That's not a big deal to create a script, which will go through all filesystem, find all "#!/usr/bin/python", change them to #!/usr/bin/python1.5. Once it's done - install Python 2.2 as defult python.
Untill it's done - I am moving to Gentoo. Well, I begin to love Gentoo and I may stay with it anyway :)
I think it's time to include groups into group member lists - let's play trees. Well, with multiple group assignment it will be actually a DAG.
Let's use profiles with pre-assigned user parameters (and scripts) even on group level. Of course such parameters should be inherited (another question - overriden?) by users traversing their DAGs. At run-time or at new user constructor - it is up-to implementations. That's not difficult even with bash scipts and will be just singing with Python as a profile scripting language.
Filesystem permission should be assigned to roles in a way similar like today user-group-others. Of course, "user" becomes a role "owner", "group" goes for list of roles the user "belongs", "others" becomes a role "guess" or "anonymous".
Besides grouping users and their permissions, roles can be used to group user profile fragments. Again, user's profile inherits all environment variables and login scripts from all roles it belongs.
Of course the relationsip "belongs" (in its own turn either "part-of" or "sort-of") is not the only type of relationships between nodes. But different applications (like FS driver above or login script) may use (interpretate) different types of relationships with different type of inheritance inference.
We don't live without networks, remember? RBAC DAG should be distrbuted. User profiles should roamed. Domains (aka NT domains) should be nested.
bla-bla-bla. We all were reading about RBAC at some time. Now it might be time to discuss how make the system where security level is a subject of configuration. You want a simple home desktop - most of such RBAC configuration goes as default. You need a regular corporate desktop - choose the configuration type "aka NT domain" on the server and workstation. You are required to install a box with C2 level of security - you need RBAC (at least) for it.
back to "Linux TNG" subject. There are lots of theoretical research done. There is (are) some stand with reference implementation form NIST. There is SE-Linux project (still alive?) from NSA. there is no Linux standard for RBAC.
That should be something like LSB, defining default user roles, their default profiles and permissions. Such standard will help to accelerate Linux RBAC implementation in the kernel, encourage Linux application developers to design keeping Linux RBAC in mind and assure corporate decision makers that Linux is targeting corporate networks seriously.
US has the biggest military forces and it uses them to change the govt in other country or to make a pressure to get immunity from International Crime Court, which is supposed to deal with genocide, what is exactly what US afraid of being presecuted for (why?).
The other countries should shut up even when their neighbours are more dangerous than Afganistan. Just shut up and do not interfere with big games of big guys.
Moreover, those big guys in US govt make everything possible to assure that ALL other coutries are weaker than US. Thus, it would be possible to dictate them everything and to dominate the world.
Of course any interests of local peoples are ignored. Even the life of own american people (as well europeans) is ignored and ready for being sacraficed - it is no problem that Afganistan increased its drag export (major part of it is going to US and EU) - it is even good for the business of some big guys who are behind US govt.
As for washing the brain - no problem at all. At the time when US refused to keep peacekeepers further on Balkans, all European newspapers where full of headlines about it - CNN has been keeping a silence in 4 days. I guess 4 days american people did not know how ugly its govt is again.
I guess the biggest chaos is in conflict of interests of big US corps and the face the US govt must leep badly protecting such interests in international hidden games.
It is too hard to enforce
It will much easier if US will show a good example of ebeying international laws.
Creating international "bully forces" to impose Western values on other countries isn't going to do anything but cause more people worldwide to become "terrorists" who are out to "kill the West."
If you, my friend, will leave for awhile you warm seat and see other people outside of US, you will hear the opposite point - people are angry that US ignore international laws.
By the way, the goal of peacekeepers is not to bring western values, but to bring the peace - the function is known from the history of all continents. I mean to bring the peace to people first all, only after that thinking about the peace for govts.
International law should be stripped down to cover war crimes like genocide. It shouldn't have any place outside the of things like that.
A genocide of the govt against its own people must be the subject of such curt as well as war crimes.
Unfortunately - forget it for awhile. US goverment questioned: "What International Crime Court? There is no such thing if US soldgers do not have any immunity from it.". And then claimed: "I am the international law!". And actually proved it: "I was the International law for Afganistan. I am the International law for Iraq. I will be the International law for all evil nations."
Unfortunately, that won't work. It did not work in Afganistan - the production of heroin and opium has been trippled compared to what it was with taliban. I doubt it will work with Iraq and Saudi. Bad example creates another bad example. Malasyan govt ecentually will give up. As Latin America govts gave up against narko-business. Why should they respect international laws if US does not?
Fortunately, it won't last forever. Switching the attention of american people from economy problems to the war cannot last forever and creates even stronger internal economy problems. Sooner or later USA will be back in 1929 and the history will repeat itself. I hope, after all USA will have some internal law forcing their goverment to respect international laws. And actually improve it - if US disagree with UN they should suggest how to reform UN, instead of boykoting it (as it was before) or ignoring (as it is now).
And only after that the war against CD pirates will bring any positive results.
PowerPC is the second (after Intel x86) on desktops and I don't see lot of PCC users enjoying Net(or other)BSD - all Macers, who hate Mac OS (9 or X), install Linux/PPC (YDL, gentoo or Debian or others). You are wrong here.
About USB - BSD might support USB as a bus, but what about devices connected to the bus? The list is short and it grows slowly in BSD. Unlike Linux.
the configuration of X is identical under BSD and Linux
Configuration files are the same. The process of hardware type and parameters discovery is what matters.
Porting applications from commercial Unix to Linux is much more of a headache than porting to BSD.
Again - wrong. Oracle, Sun and other prefer to port to Linux at furst.
Firstly, there's no standard Linux distribution, and all the distributions have subtly different patched kernel versions.
Even this is wrong: There is one tree of Linux kernel against several BSD ones. And noone, besides you, mentioned any problem with different kernel patches in different distributions. I cannot say same about the constant headache: "Which BSD should I choose? FreeBSD? NetBSD? OpenBSD?"
I guess ford has lost less % of its initial cost - it must be in better shape. Isn't it?
Actually I like the fact of such interruptions - it may convience more and more managers, architects and developers to use asyncronous transactional messaging protocols (like JMS or SOAP or even SMTP if with confirmation) vs decent client server ones (CORBA, proprietary).
I guess it will be much faster to complete the port of Linux to SGI/MIPS. By the time it will be done SGI will have same OS on both platforms, namely Linux :)