Firstly, not true. You need xcodebuild to build some of the components, and xcodebuild isn't free. Secondly, you appear to have deliberately avoided any point this thread contained: you require Darwin to build Darwin, as you require Solaris to build Solaris. The OP was stating that the latter was a shame and would probably be fixed, the FCP observed that it ain't necessarily so.
Unfortuntely, all builds require a system to bootstrap the build. At the moment, the only option is Solaris Community Edition, a non-Torrented download. (Boo!) That being said, I don't think we'll have to wait too long for the OSS community to fix that little issue.
Sadly, that might not necessarily be the case - OpenDarwin has been around for five years and you still need a Darwin system [including some non-free tools] to bootstrap the build. But as Solaris Express is free as in non-costworthy, I don't see any problem with downloading SE to bootstrap an OpenSlowlaris kernel build. Me, I'm still using Solaris 2,7, and 8;-)
Yes, in that case it is your compiler's fault if that segfaults, it shouldn't even bloody compile. For a start, the function main() is of type int, not void. Its arguments are (const int c, const char **argv). Your program's control reaches the end of a non-void function, which is naughty although IIRC not strictly an error condition. But you also have (an implicit) typing problem in the test=malloc(10); line, because you have not declared the malloc() function - return type defaults to int when you want int * (in fact, malloc() returns a void * so you should be casting anyway). You haven't declared free() either, but as it's used in void context that's not too much of a problem
I think that was more the absurd licensing costs, and the fact that it didn't offer enough over other existing formats. I all but switched to MD from CD about seven years ago. The fact that you have two different media for RO and RW discs is ridiculous.
My experience would suggest that there are a lot of people using distributed RPC mechanisms to achieve utility computing and distributed work. However, my experience also encompasses Grid computing, and the sad fact is that cluster computing in the Grid seems to be limited to running 'nohup' on multiple machines.
I agree with your sentiment; using the Apple developer tools and environment as standard would be sweet as. Even back in the mid 1990s the NeXT developer environment was absolute luxury. The problem is most heads of IT (and most IT support staff) depend on Windows for their livelihood so aren't about to endorse a switch to Mac, Linux, OpenVMS or anything else.
The first time I really got annoyed by this sort of thing was back when I was learning NeXTSTEP programming. Their developer reference docs were largely a hardcopy set of API regurgitations, but this was OK because they were well indexed (as they bloody well should have been, I expect they were just printouts of the NeXT Librarian format online documentation). I also don't mind this because for really ploughing through a book, I still prefer dead tree to online format[*]. However, the printed documentation for the compiler/linker/debugger toolset was just a troff-formatted pretty version of the GNU manpages (remember this was back when RMS didn't mind man too much). That really annoyed me. They were badly indexed, containing the same information as a manpage in the same linear style and without the utility of a content search.
[*]This is why I own a number of books in both online (for quick reference) and hardcopy (for actually studying) formats, even though the online version is free (c.f. Numerical Recipes, the X manuals, some of the Perl books etc.)
One part of the reg's article was certainly correct - that a number of these sites act as echo chambers for people who want their own beliefs reaffirmed. So far in this Slashdot thread I've seen "well it's obvious, Groklaw gets it right" or similar sentiment displayed in multiple comments, without one statement of factual evidence used to corroborate the claim. Me, I don't know anything about Monterey, but I'd be happy to consider any evidence that people are willing to post:-)
It's good when that's possible, but it isn't always so. Easier on HURD-based systems than on Linux, especially. I remember a local DoS exploit which (I think) involved the SIGALRM handler, that was a kernel vulnerability that needed a reboot to fix.
Or, It Already Works For Someone Else So We'll Pinch It:
auto-defragmenting in the background HFS+
ability to have files in more than one folder simultaneously symlinks, Smart Folders
the new ad campaign Microsoft is running to get people excited about Windows Maybe that does indeed Just Work. No-one ever got fired for choosing a Microsoft (although there are places where that's beginning to change).
You are Andy Tanenbaum and I claim my five copies of Operating Systems: Design and Implementation. Having said that I agree, microkernel = sweet. Especially Pistachio.
A last hope before I throw my hands in the air crying, "Why George, why?"
OK, George, you've done some rubbish, but I'm going to give you one more go. I'll watch episode 3, and if that's bad then by heck I'll not watch another Star Wars film as long as I live! Erm....
...or get a laptop that takes multiple batteries. Mine does - it's a PowerBook G3 Series (a.k.a. wallstreet). With a battery in each of the two bays, I can get a good 7+hrs of console-based emacs/mutt goodness. I don't know what it's like in X11, because I don't use it enough...
most Linux users seem to favour computers with proprietary locked-in BIOS
Now, unless you can prove that statement to be incorrect by stating usage statistics for LinuxBIOS, I think the point still stands. Anyway, to use and support LinuxBIOS is merely to fragment the user and developer communities for an open BIOS project - seeing as one (IEEE1275) already exists, is very stable and widely adopted.
Yes, I have. However, an open source update client on its own is useless to actually maintain a source repository. Linus is using a BitKeeper server (and presumably, a more functional non-Free client) for the Linux sources. That would be not Free, and hence he has chosen pragmatism over politics. Not to mention that but most Linux users seem to favour computers with proprietary locked-in BIOS (he says, patting his IEEE1275 machine).
But because Linus uses what I'm sure we all like to refer to as The Satanic Evilness to maintain the revision control, you'll agree that it can never truly be Free. Again, with the perverse developers and their pragmatism. BitKeeper may be better than Free alternatives but that's no reason to go around using it.
It seems a decision based largely on practical considerations
Silly open source developers - putting practicality and pragmatism above more important things like dyed-in-the-wool political viewpoints. Next you'll be telling me they're all off using these newbie Linux systems, rather than diligently waiting for HURD to stabilise like they're supposed to. Tch.
Slashdot Mirror
Firstly, not true. You need xcodebuild to build some of the components, and xcodebuild isn't free. Secondly, you appear to have deliberately avoided any point this thread contained: you require Darwin to build Darwin, as you require Solaris to build Solaris. The OP was stating that the latter was a shame and would probably be fixed, the FCP observed that it ain't necessarily so.
Sadly, that might not necessarily be the case - OpenDarwin has been around for five years and you still need a Darwin system [including some non-free tools] to bootstrap the build. But as Solaris Express is free as in non-costworthy, I don't see any problem with downloading SE to bootstrap an OpenSlowlaris kernel build. Me, I'm still using Solaris 2,7, and 8 ;-)
* bashes you with python and ruby :-P *
Yes, in that case it is your compiler's fault if that segfaults, it shouldn't even bloody compile. For a start, the function main() is of type int, not void. Its arguments are (const int c, const char **argv). Your program's control reaches the end of a non-void function, which is naughty although IIRC not strictly an error condition. But you also have (an implicit) typing problem in the test=malloc(10); line, because you have not declared the malloc() function - return type defaults to int when you want int * (in fact, malloc() returns a void * so you should be casting anyway). You haven't declared free() either, but as it's used in void context that's not too much of a problem
I think that was more the absurd licensing costs, and the fact that it didn't offer enough over other existing formats. I all but switched to MD from CD about seven years ago. The fact that you have two different media for RO and RW discs is ridiculous.
My experience would suggest that there are a lot of people using distributed RPC mechanisms to achieve utility computing and distributed work. However, my experience also encompasses Grid computing, and the sad fact is that cluster computing in the Grid seems to be limited to running 'nohup' on multiple machines.
;-)
Oh, and there so is Oracle on OpenVMS
What's the problem with commodity software? POSIX and X11 are both widely-adopted standards, yes?
Someone who likes measuring their uptimes in decades might consider OpenVMS.
I agree with your sentiment; using the Apple developer tools and environment as standard would be sweet as. Even back in the mid 1990s the NeXT developer environment was absolute luxury. The problem is most heads of IT (and most IT support staff) depend on Windows for their livelihood so aren't about to endorse a switch to Mac, Linux, OpenVMS or anything else.
The first time I really got annoyed by this sort of thing was back when I was learning NeXTSTEP programming. Their developer reference docs were largely a hardcopy set of API regurgitations, but this was OK because they were well indexed (as they bloody well should have been, I expect they were just printouts of the NeXT Librarian format online documentation). I also don't mind this because for really ploughing through a book, I still prefer dead tree to online format[*]. However, the printed documentation for the compiler/linker/debugger toolset was just a troff-formatted pretty version of the GNU manpages (remember this was back when RMS didn't mind man too much). That really annoyed me. They were badly indexed, containing the same information as a manpage in the same linear style and without the utility of a content search.
[*]This is why I own a number of books in both online (for quick reference) and hardcopy (for actually studying) formats, even though the online version is free (c.f. Numerical Recipes, the X manuals, some of the Perl books etc.)
One part of the reg's article was certainly correct - that a number of these sites act as echo chambers for people who want their own beliefs reaffirmed. So far in this Slashdot thread I've seen "well it's obvious, Groklaw gets it right" or similar sentiment displayed in multiple comments, without one statement of factual evidence used to corroborate the claim. Me, I don't know anything about Monterey, but I'd be happy to consider any evidence that people are willing to post :-)
It's good when that's possible, but it isn't always so. Easier on HURD-based systems than on Linux, especially. I remember a local DoS exploit which (I think) involved the SIGALRM handler, that was a kernel vulnerability that needed a reboot to fix.
In reference to his "over a year of uptime". My servers tend to get uptimes of a few months, because I like to patch security issues.
Which of the Linux kernels are you using that hasn't had a security hole fixed in the previous year? Thanks.
That's such a good idea, it's already been done. One example is:
Password Helper
Use the Password Helper panel to pick a secure password.
From mac os X 10.4.
My point is the same one that you picked up on; that Microsoft's new features for Longhaul are as innovative as we've come to expect from them.
Or, It Already Works For Someone Else So We'll Pinch It:
auto-defragmenting in the background HFS+
ability to have files in more than one folder simultaneously symlinks, Smart Folders
the new ad campaign Microsoft is running to get people excited about Windows Maybe that does indeed Just Work. No-one ever got fired for choosing a Microsoft (although there are places where that's beginning to change).
You are Andy Tanenbaum and I claim my five copies of Operating Systems: Design and Implementation. Having said that I agree, microkernel = sweet. Especially Pistachio.
OK, George, you've done some rubbish, but I'm going to give you one more go. I'll watch episode 3, and if that's bad then by heck I'll not watch another Star Wars film as long as I live! Erm....
Emendation or amendment? Either way, it should really read: The next Slashdot story will be the same as the last one.
...or get a laptop that takes multiple batteries. Mine does - it's a PowerBook G3 Series (a.k.a. wallstreet). With a battery in each of the two bays, I can get a good 7+hrs of console-based emacs/mutt goodness. I don't know what it's like in X11, because I don't use it enough...
I said:
Now, unless you can prove that statement to be incorrect by stating usage statistics for LinuxBIOS, I think the point still stands. Anyway, to use and support LinuxBIOS is merely to fragment the user and developer communities for an open BIOS project - seeing as one (IEEE1275) already exists, is very stable and widely adopted.
Yes, I have. However, an open source update client on its own is useless to actually maintain a source repository. Linus is using a BitKeeper server (and presumably, a more functional non-Free client) for the Linux sources. That would be not Free, and hence he has chosen pragmatism over politics. Not to mention that but most Linux users seem to favour computers with proprietary locked-in BIOS (he says, patting his IEEE1275 machine).
But because Linus uses what I'm sure we all like to refer to as The Satanic Evilness to maintain the revision control, you'll agree that it can never truly be Free. Again, with the perverse developers and their pragmatism. BitKeeper may be better than Free alternatives but that's no reason to go around using it.
Silly open source developers - putting practicality and pragmatism above more important things like dyed-in-the-wool political viewpoints. Next you'll be telling me they're all off using these newbie Linux systems, rather than diligently waiting for HURD to stabilise like they're supposed to. Tch.