The article also describes a FreeBSD 7.0 pre-release from October last year. This still had debugging code turned on in the builds, as mentioned on the NetBSD lists when Andrew Doran was comparing NetBSD -current SMP performance.
Ousterhout never worked for Sun, they simply provided him with office space and equipment to work on Tcl/Tk away from his regular job as a university professor. I'm pretty sure the original Tcl book makes this clear in the preface, but my copy is in storage at the moment so I can't check.
GNU screen is the equivalent of window(1), which has been in the BSD userland since the mid 1980's. Glibc is something I detest, as its footprint is huge, its maintainer is a notorious jerk and compared to porting from a BSD to Solaris I find myself having to sprinkle my code with a lot of #ifdefs to port to Linux. The bloat of glibc may not be an issue for desktop or server machines, but is a pain in the backside for the embedded stuff I've worked on.
The Wine breakage is most likely down to how Linux implements pthreads - there are some grey areas in the POSIX threads spec, where things could be more strictly enforced such as double locking or freeing of mutexes. Linux takes a less stringent approach than FreeBSD and NetBSD, accepting such common coding mistakes, whereas the kernel and libc threading code in the BSD's will print an error and dump core. Being a fan of such things as rigidly type safe languages and compilers that offer a high degree of warnings - having seen the mess less exacting languages can cause (*cough* Perl *cough*) - I prefer the BSD approach over the Linux one.
Better performance than Linux, that degrades under load much more predictably than Linux (as does Solaris, but FreeBSD is better on commodity hardware). A better written C library (just look at the source code to glibc - it's shockingly bad, unreadable macro soup as though its maintainer hates C). A better documented userland than Linux with complete and accurate manpages.
FreeBSD is popular amongst hosting companies (the tools for security are easier to use and more mature than Linux), and is also used by companies like Yahoo! because of it's reliability and performance. Linux has outperformed FreeBSD for a while, as the fine grained locking introduced in version 5 matured, but the pain getting it right is beginnng to pay off now.
Re:Good reporting there, submitter
on
LLVM 2.2 Released
·
· Score: 1
If you're going to go completely off the deep end and call the GPL commie, at least get your commies straight. Stalin isn't marxist.
Stalin wasn't a Marxist? He described himself as Marxist-Leninist, and attempted to drag the Soviet Union into the industrial age with the five year plans ostensibly to fit with Marxist preconditions for a Communist state. Marxist apologists like to counter the argument that their doctrine is fundamentally flawed by claiming that Stalin, Mao, et. al. aren't Marxists, while many still admire Trotsky and Lenin. The facts are that regardless of Lenin's alleged deathbed warning about Stalin, it was Lenin and Trotsky who laid the groundwork for the Stalin regime with the Cheka and Red Terror. Just because the inevitable results, in line with Michel's Iron Law of Oligarchy, were authoritarian dictatorships does not mean that their creators weren't working from Marxist precepts. Somewhat ironic when you consider how much selective use of facts Marx made when writing Das Kapital, and the downright misleading way he portrayed the Paris Commune as Communist.
Re:Good reporting there, submitter
on
LLVM 2.2 Released
·
· Score: 1
We're talking about freedom not GNU Freedom*. If you want to draw politic analogies, then it's a little like comparing BSD liberalism to GNU Marxism with Stallman as Stalin, changing the consitution - sorry, license - as it suits him.
* A registered trademark of the Free Software Foundation. (I jest).
Re:Good reporting there, submitter
on
LLVM 2.2 Released
·
· Score: 2, Informative
That's why Apple used a BSD base as opposed to a Linux base for Mac OS X.
No, it's because MacOS X is a continuation of NeXTstep, which was based on a BSD userland and Mach kernel (itself originally developed from the BSD Unix kernel circa version 4.3). At the time NeXTstep was originally developed, Linux didn't exist and the GNU userland was still in its infancy. In addition, I've never seen any evidence that Apple considered a move to a Linux base for MacOS X, and considering how solid NeXTstep and OpenStep were I'd be surprised if they did. Licensing would have been an unlikely issue anyway, as NeXTstep always shipped with GCC as part of the development tools and there are a number of GNU tools in MasOS X.
Re:Good reporting there, submitter
on
LLVM 2.2 Released
·
· Score: 1
There's some concern in the BSD camps and elsewhere about the switch to GPLv3 for GCC and the GNU binutils. Some of the impetus for the changes in GPLv3 is the use of GCC in the embedded programming world, where the development and target platforms are closed source (usually Windows using Cygwin in the case of the development platform). As embedded applications is where BSD is quite heavily used, without the fanfare that Linux often gets simply because you don't have to make the source code available, companies who prefer BSD operating systems are worried about the implications of their reliance on the GNU toolchain. I don't know about Apple, but they may have similar concerns about the FSF agenda with regard to GPLv3, hence them putting resource into a new frontend for LLVM. Regardless of the politics, I'd very much like to see a better open source compiler than GCC whether it's under the BSD license or GPLv2 (without the "later version" wording).
No, I was one of the people who chose to go to that college in the mistaken belief that as it was part of London University I'd be able to pick and choose interesting course modules at all the London colleges. It didn't quite work out like that as all the interesting modules were massively over subscribed. As for my dad's firm, the whole reason I was at university was to avoid having to work on a building site (he was a self-employed floor and wall tiler).
I went to a college that was infested with Oxbridge rejects, rich kids who were so stupid that Mummy and Daddy couldn't get them into Oxford or Cambridge. Instead they ended up at the college I was at, as it was close to where most of the chinless wonders came from (Windsor, Ascot, etc). What really annoyed me about them was how "right on" they were, clad in Che Guevera t-shirts, a copy of the Socialist Worker newspaper tucked under their arm and spouting Trotskyite dogma in the student bar. Of course, nepotism meant they were destined to work at Pater's firm in the city once they'd flunked their exams, hypocritical twats. Anyway, to wind them up I painted "This machine kills Marxists" on my bass guitar before playing the battle of the bands - inspired of course by Woody Guthrie.
Come on now, that simply assigns shat to bricks (and that's some nasty use of the comma operator to separate statements). I think you meant:
while (exploitable) { Bricks *bricks = malloc(sizeof(Bricks)); shit(bricks); sleep(1); }
Note that we don't have to dispose of the bricks we shit, as that's taken care of elsewhere. And of course, if we all still wrote VAX assembler we would be able to optimise this by using the SHTBRCKS instruction.
My experience with MySQL replication is that when it works, it's OK. However, it is very fragile, with none of the robustness that Informix or Sybase have in this area (those two being the other RDBMS I've used replication with extensively). This fits in with the whole cheap and cheerful philosophy of MySQL, but along with the lack of features in MyISAM, the poor performance of InnoDB and the regular corruptions of data on any sizable (20GB plus) database it means I'd rather struggle with the configuration of Slony and use PostgreSQL instead.
Novell doesn't have as many developers working on GNOME as RedHat, and Imendio have more developers working on Glib and GTK+ than Novell. The developers that Novell does have working on GNOME related applications and development tools are focused on Evolution and Mono respectively, and even Evolution has no dependencies on Mono.
Nope, Windows NT hasn't been a microkernel based OS since 1996. Version 3 (3.1 according to the article, but I think it applies to the entire 3 series) was microkernelish, but version 4.0 removed the microkernel aspects. This change was to make performance, particularly for graphics, better by allowing drivers direct access to the hardware, but it buggered up the stability no end.
I wonder if the move to WebKit for the rendering engine used by Epiphany will prompt other GNOME projects to transition from the various gtkhtml versions that are currently used. The maintenance of gtkhtml seems to be sporadic, and the API changes drastically between versions. For example, on a Fedora 8 install at work there's two versions of the gtkhtml library required by different apps in the basic GNOME desktop...
I expect there to still be a descendant of Java in widespread use come 2040. As a language and set of libraries, the Java environment is a fine combination of OO abstractions on top of a C like syntax while avoiding the complexity of C++ and the STL. Smalltalk fans may grumble that Java's not totally object oriented because of the fundamental types like int, double, and so on, but they strike me as bitter that Java got the success that Smalltalk could have had. And Java is undoubtedly inspired by Smalltalk, either directly or via Objective-C and the NeXTSTEP Foundation classes (which Sun were very interested in when Java was being conceived). Is Java perfect? Far from it, however it has improved my productivity and the robustness of the applications I work on more than any other language I've used (including scripting languages such as Perl and Tcl). I've dabbled in Scheme, and read up on Common LISP, but while they are both very "pure" in concept, the former strikes me as too limited to get anything useful done and the latter too unwieldy (and I don't mean the parentheses laden syntax).
If you consider how long AI courses have been going, where LISP is (was?) the de-facto programming language, then there may actually be more lines of code written in Lisp than in Java. However, in the non-academic world it's pretty obvious that there's far more Java than LISP code in production.
1st year students are rarely taught debugging skills
Why? In the first week of a HNC course I attended (a vocational qualification that's considered less of an achievement than a full degree) we were taught the rudiments of debugging and profiling our C code. That was only eleven years ago, so I find it hard to understand why an emphasis on understanding what high level code does when it's been compiled and run has been lost. To me this smacks more of laziness on the part of the lecturers, as debugging and profiling should be taught regardless of the programming language and th tools are provided with the Java SDK.
As for the article, this guy's got a real hard on for Pascal. Sorry, but Java is a much better choice for an educational programming language - all you need to do is exclude the use of classes from the java.awt.* and javax.swing.* packages if you want students to avoid getting diverted by graphics programming. Likewise, the presence of the Java Collections doesn't preclude an understanding of data structures and algorithms - if anything they provide convenient examples of usage, as demonstrated by Weiss book "Data Structures and Algorithm Analysis in Java" which is a standard textbook on many courses.
Java seems to have survived the lack of a mod_java
That's because Java web apps are typically deployed on a app container behind either Apache using mod_jk, or a firewall using port forwarding rules. Both methods allow the graceful restart of a web app the app. The first scenario avoids the need for a "mod_java", as Apache is just proxying requests. In the second scenario, you typically run the app container on Unix using an unprivileged port such as 8080, and use port forwarding from port 80.
Slashdot Mirror
The article also describes a FreeBSD 7.0 pre-release from October last year. This still had debugging code turned on in the builds, as mentioned on the NetBSD lists when Andrew Doran was comparing NetBSD -current SMP performance.
Good points, and it's worth noting that NT was a new kernel with the Win16 and Win32 APIs on top (the former running in an emulation of DOS).
Ousterhout never worked for Sun, they simply provided him with office space and equipment to work on Tcl/Tk away from his regular job as a university professor. I'm pretty sure the original Tcl book makes this clear in the preface, but my copy is in storage at the moment so I can't check.
GNU screen is the equivalent of window(1), which has been in the BSD userland since the mid 1980's. Glibc is something I detest, as its footprint is huge, its maintainer is a notorious jerk and compared to porting from a BSD to Solaris I find myself having to sprinkle my code with a lot of #ifdefs to port to Linux. The bloat of glibc may not be an issue for desktop or server machines, but is a pain in the backside for the embedded stuff I've worked on.
The Wine breakage is most likely down to how Linux implements pthreads - there are some grey areas in the POSIX threads spec, where things could be more strictly enforced such as double locking or freeing of mutexes. Linux takes a less stringent approach than FreeBSD and NetBSD, accepting such common coding mistakes, whereas the kernel and libc threading code in the BSD's will print an error and dump core. Being a fan of such things as rigidly type safe languages and compilers that offer a high degree of warnings - having seen the mess less exacting languages can cause (*cough* Perl *cough*) - I prefer the BSD approach over the Linux one.
what for?
Better performance than Linux, that degrades under load much more predictably than Linux (as does Solaris, but FreeBSD is better on commodity hardware). A better written C library (just look at the source code to glibc - it's shockingly bad, unreadable macro soup as though its maintainer hates C). A better documented userland than Linux with complete and accurate manpages.
FreeBSD is popular amongst hosting companies (the tools for security are easier to use and more mature than Linux), and is also used by companies like Yahoo! because of it's reliability and performance. Linux has outperformed FreeBSD for a while, as the fine grained locking introduced in version 5 matured, but the pain getting it right is beginnng to pay off now.
If you're going to go completely off the deep end and call the GPL commie, at least get your commies straight. Stalin isn't marxist.
Stalin wasn't a Marxist? He described himself as Marxist-Leninist, and attempted to drag the Soviet Union into the industrial age with the five year plans ostensibly to fit with Marxist preconditions for a Communist state. Marxist apologists like to counter the argument that their doctrine is fundamentally flawed by claiming that Stalin, Mao, et. al. aren't Marxists, while many still admire Trotsky and Lenin. The facts are that regardless of Lenin's alleged deathbed warning about Stalin, it was Lenin and Trotsky who laid the groundwork for the Stalin regime with the Cheka and Red Terror. Just because the inevitable results, in line with Michel's Iron Law of Oligarchy, were authoritarian dictatorships does not mean that their creators weren't working from Marxist precepts. Somewhat ironic when you consider how much selective use of facts Marx made when writing Das Kapital, and the downright misleading way he portrayed the Paris Commune as Communist.
We're talking about freedom not GNU Freedom*. If you want to draw politic analogies, then it's a little like comparing BSD liberalism to GNU Marxism with Stallman as Stalin, changing the consitution - sorry, license - as it suits him.
* A registered trademark of the Free Software Foundation. (I jest).
That's why Apple used a BSD base as opposed to a Linux base for Mac OS X.
No, it's because MacOS X is a continuation of NeXTstep, which was based on a BSD userland and Mach kernel (itself originally developed from the BSD Unix kernel circa version 4.3). At the time NeXTstep was originally developed, Linux didn't exist and the GNU userland was still in its infancy. In addition, I've never seen any evidence that Apple considered a move to a Linux base for MacOS X, and considering how solid NeXTstep and OpenStep were I'd be surprised if they did. Licensing would have been an unlikely issue anyway, as NeXTstep always shipped with GCC as part of the development tools and there are a number of GNU tools in MasOS X.
There's some concern in the BSD camps and elsewhere about the switch to GPLv3 for GCC and the GNU binutils. Some of the impetus for the changes in GPLv3 is the use of GCC in the embedded programming world, where the development and target platforms are closed source (usually Windows using Cygwin in the case of the development platform). As embedded applications is where BSD is quite heavily used, without the fanfare that Linux often gets simply because you don't have to make the source code available, companies who prefer BSD operating systems are worried about the implications of their reliance on the GNU toolchain. I don't know about Apple, but they may have similar concerns about the FSF agenda with regard to GPLv3, hence them putting resource into a new frontend for LLVM. Regardless of the politics, I'd very much like to see a better open source compiler than GCC whether it's under the BSD license or GPLv2 (without the "later version" wording).
so do you now work at your father's firm?
No, I was one of the people who chose to go to that college in the mistaken belief that as it was part of London University I'd be able to pick and choose interesting course modules at all the London colleges. It didn't quite work out like that as all the interesting modules were massively over subscribed. As for my dad's firm, the whole reason I was at university was to avoid having to work on a building site (he was a self-employed floor and wall tiler).
I went to a college that was infested with Oxbridge rejects, rich kids who were so stupid that Mummy and Daddy couldn't get them into Oxford or Cambridge. Instead they ended up at the college I was at, as it was close to where most of the chinless wonders came from (Windsor, Ascot, etc). What really annoyed me about them was how "right on" they were, clad in Che Guevera t-shirts, a copy of the Socialist Worker newspaper tucked under their arm and spouting Trotskyite dogma in the student bar. Of course, nepotism meant they were destined to work at Pater's firm in the city once they'd flunked their exams, hypocritical twats. Anyway, to wind them up I painted "This machine kills Marxists" on my bass guitar before playing the battle of the bands - inspired of course by Woody Guthrie.
However, bricks = shat.
Come on now, that simply assigns shat to bricks (and that's some nasty use of the comma operator to separate statements). I think you meant:
Note that we don't have to dispose of the bricks we shit, as that's taken care of elsewhere. And of course, if we all still wrote VAX assembler we would be able to optimise this by using the SHTBRCKS instruction.
That's because he's so l33t he can pick a Slashdot id at random every time he posts.
My experience with MySQL replication is that when it works, it's OK. However, it is very fragile, with none of the robustness that Informix or Sybase have in this area (those two being the other RDBMS I've used replication with extensively). This fits in with the whole cheap and cheerful philosophy of MySQL, but along with the lack of features in MyISAM, the poor performance of InnoDB and the regular corruptions of data on any sizable (20GB plus) database it means I'd rather struggle with the configuration of Slony and use PostgreSQL instead.
Novell doesn't have as many developers working on GNOME as RedHat, and Imendio have more developers working on Glib and GTK+ than Novell. The developers that Novell does have working on GNOME related applications and development tools are focused on Evolution and Mono respectively, and even Evolution has no dependencies on Mono.
Nope, Windows NT hasn't been a microkernel based OS since 1996. Version 3 (3.1 according to the article, but I think it applies to the entire 3 series) was microkernelish, but version 4.0 removed the microkernel aspects. This change was to make performance, particularly for graphics, better by allowing drivers direct access to the hardware, but it buggered up the stability no end.
WebKit can be used in apps written in C and Objective-C, thanks to the KWQ wrapper, and unlike KHTML it has no dependencies on the Qt toolkit.
I wonder if the move to WebKit for the rendering engine used by Epiphany will prompt other GNOME projects to transition from the various gtkhtml versions that are currently used. The maintenance of gtkhtml seems to be sporadic, and the API changes drastically between versions. For example, on a Fedora 8 install at work there's two versions of the gtkhtml library required by different apps in the basic GNOME desktop ...
Fortran is no longer written all caps, and hasn't been since Fortran 90 came out.
I expect there to still be a descendant of Java in widespread use come 2040. As a language and set of libraries, the Java environment is a fine combination of OO abstractions on top of a C like syntax while avoiding the complexity of C++ and the STL. Smalltalk fans may grumble that Java's not totally object oriented because of the fundamental types like int, double, and so on, but they strike me as bitter that Java got the success that Smalltalk could have had. And Java is undoubtedly inspired by Smalltalk, either directly or via Objective-C and the NeXTSTEP Foundation classes (which Sun were very interested in when Java was being conceived). Is Java perfect? Far from it, however it has improved my productivity and the robustness of the applications I work on more than any other language I've used (including scripting languages such as Perl and Tcl). I've dabbled in Scheme, and read up on Common LISP, but while they are both very "pure" in concept, the former strikes me as too limited to get anything useful done and the latter too unwieldy (and I don't mean the parentheses laden syntax).
If you consider how long AI courses have been going, where LISP is (was?) the de-facto programming language, then there may actually be more lines of code written in Lisp than in Java. However, in the non-academic world it's pretty obvious that there's far more Java than LISP code in production.
1st year students are rarely taught debugging skills
Why? In the first week of a HNC course I attended (a vocational qualification that's considered less of an achievement than a full degree) we were taught the rudiments of debugging and profiling our C code. That was only eleven years ago, so I find it hard to understand why an emphasis on understanding what high level code does when it's been compiled and run has been lost. To me this smacks more of laziness on the part of the lecturers, as debugging and profiling should be taught regardless of the programming language and th tools are provided with the Java SDK.
As for the article, this guy's got a real hard on for Pascal. Sorry, but Java is a much better choice for an educational programming language - all you need to do is exclude the use of classes from the java.awt.* and javax.swing.* packages if you want students to avoid getting diverted by graphics programming. Likewise, the presence of the Java Collections doesn't preclude an understanding of data structures and algorithms - if anything they provide convenient examples of usage, as demonstrated by Weiss book "Data Structures and Algorithm Analysis in Java" which is a standard textbook on many courses.
Java seems to have survived the lack of a mod_java
That's because Java web apps are typically deployed on a app container behind either Apache using mod_jk, or a firewall using port forwarding rules. Both methods allow the graceful restart of a web app the app. The first scenario avoids the need for a "mod_java", as Apache is just proxying requests. In the second scenario, you typically run the app container on Unix using an unprivileged port such as 8080, and use port forwarding from port 80.
Kaz Kylheku? Late of comp.lang.c? My god it's a small world.