Well, I'm an employee for a highly successful company, and we had no end of problems with MySQL. Deleting large numbers of rows from MyISAM tables with full text indexing would corrupt them. Running an ptimize on large tables (MyISAM and InnoDB) would corrupt them. Replication quietly fails rather than generating warnings if the syntax of the failover commands is not exactly right. The MySQL JDBC driver is terrible. The C++ code used in the admin tools is shockingly bad. Date and time columns cause all manner of problems. Need I go on?
We switched to PostgreSQL. Performance is better, replication works reliably and the JDBC driver is great.
It's not hard because there are dozens of Debian maintainers feverishly working day and night to keep it from breaking.
If it's true that it needs dozens of maintainers on Debian then the package management is poor. In the NetBSD pkgsrc tree, GNOME is maintained largley by one individual. It's much more likely that there's more people willing and able to work on the Debian packages, but I very much doubt that it absolutely requires dozens of them to keep up with GNOME releases. GNOME code has become very clean and portable since the release of version 2.0.
I also have serious doubts that NetBSD are still using XFree86
Yes, the version of X in the NetBSD source tree is still XFree86. This is because many platform specific changes have been made to the X code in the NetBSD tree, many of which had not found their way back into XFree86 before the X.Org fork. Getting the changes back into X.Org will become increasingly difficult as the code evolves, although much of it is localised into X server code that's unlikely to be receiving much attention from the X.Org guys, so fingers crossed...
That said, many people are running X.Org on NetBSD, installing it via pkgsrc. I've done just that on my main machine, an i386 laptop, while my "fun" machines such as a Vax, SGI Indy and SparcStation run the default XFree86 (or no X at all in the case of the Vax).
There is some economic sides to it as well with HIGHER taxes (definitely not lower, as you said), and pro-welfare state.
Not true. Liberal political theory promotes a "rolling back" of the state so that it interferes in the life of citizens as little as possible. This doesn't preclude a welfare state, far from it, however it does suggest having one that does not rely on excessive taxation. the reason the conservatism of Thatcher and Bush has become known as "neo-liberalism" is because it has adopted the principle of rolling back the state. The difference is that this rolling back is geared towards the benefit of business rather than the individual.
For those puzzled, RMS's Liberal comment is in reference to Canada's Liberal party.
Whoever transcribed the interview, or RMS himself if the interview was conducted via email, should have written "liberal" with a small "l". In political writing a capitalised word like Liberal indicates a party, while the lowercase form indicates a theory or dogma. For example, Conservative would imply the political party when speaking about UK politics, whereas conservative would imply the political theory. Exceptions include political theories such as Marxism, which are named after an individual and should of course be capitalised.
I'd be happy to run "one of the busiest websites on earth" using Opterons, which are basically 64bit x86 boxes. I wouldn't use Windows as the OS though...
Re:GCC is important, but what about progress in C+
on
GCC 4.1 Released
·
· Score: 1
GCC is a very important piece of open source software, but it is time to see some real progress in C++
(list of features snipped)
Why not use a toolkit that gives you much more than a plain STL, something like Qt or Boost. Alternatively, why not use a language and environment that already has those features, like Java?
I assume you mean "uses it instead of Linux", what with this being Slashdot. How about people who've benchmarked it against Linux and found Solaris to scale better and more smoothly? Some of us like having beefy Sparc or Opteron SMP machines that perform predictably with Solaris, rather than the erratic behaviour we've seen with Linux on SMP Intel hardware. The 2.6.x Linux kernel has also been a serious disappointment in terms of reliability, a definite step back from 2.4.x.
In my experience the Linux binary compatiblity in NetBSD is perfect. Until I finally got around to building a native JDK, I was running the Sun JDK for Linux under NetBSD 2.0.2. It performed flawlessly, and I only switched to a native JDK because I was bored one afternoon and decided to see if it would compile. I have also run Linux binaries of Mozilla and Firefox to get support for the Flash plugin, and these also performed flawlessly under various versions of NetBSD.
I'm subscribed to the NetBSD-current mailing list, and I see regular messages about tweaks and improvements to the Linux syscall compatability in the development version of NetBSD. This gives me considerable confidence that changes to system calls in Linux are quickly picked up on by the NetBSD developers.
You could try NetBSD instead: http://www.netbsd.org/Ports/sparc64/. I'd recommend NetBSD 3.0, which has just gone from a beta to a release candidate, and includes support for all of the graphics cards found in Ultra workstations. A prebuilt release candidate can be found on the FTP mirrors under/pub/NetBSD-daily/
Hyundai is giving american car manufactures a run for the money, not only in quality but service, performance and reliability.
It's not difficult to give American car manufacturers a run for their money. I owned a 1999 Mustang coupe that was the shittiest car I've ever been lumbered with. My MGB GT was more reliable (yup, the Mustang was that bad), while my Jaguar XJS had better performance. As for servicing, San Francisco streets meant the Mustang's suspension was fucked within a month so the servicing needed to be cheap.
It's a pity that GNOME was written way back when instead of GNUstep being the free desktop of choice - had all that effort gone into GNUstep, it would have been pretty easy to target both Mac OS X and Linux/*BSD instead of having to write separate UI code for each.
Miguel de Icaza and the other founders of the GNOME project did consider GNUstep. Their conclusion was that the amount of work needed to finish GNUstep, and the added burden of staying compliant with future OpenStep specifications weren't worth the effort. They weren't to know that NeXT would find a massive new audience as the new MacOS...
Hah! Java generics are laughable, especially if you ever used C++ templates.
Thanks I needed a laugh. Templates in C++ use a disgustingly unreadable syntax, and don't compensate for the bloated mess that the rest of the language represents. If I ever need to develop another server side system in C++ rather than Java I'll be shocked.
More importantly, "write once, run anywhere" really should be qualified as "write once, run anywhere the exact same JVM is installed and a decent script/command/launcher exists to properly start your application with all the flags, classpaths, memory management, and other configuration needed for this platform"
With all due respect, I feel you need to read up on how to package your applications better. If that fails I think you need to look at the quality of your code. I have a 200000 line Java codebase which, thanks to an MVC architecture, builds into a web application and a number of standalone applications (some are GUI tools, others are daemons). Thanks to a decent Ant build system along with a good understanding of Jar files, Log4J and webstart, I can deploy onto Windows, MacOS X and Solaris. No platform dependent tweaks at all, and no plaform specific problems encountered.
That reminds me of a story about how Cuba is filled with classic American cars from the 1940s and 1950s that are (for the most part) still well maintained and still on the road after all these years.
Getting a bit off topic, but many of those American cars are still running because the engines have been transplanted from a Russian Lada or Volga. If the same principle is applied in South Korea, then something like Linux will replace Windows when the localised version is too out of date.
This is an operating system that runs on more CPU architectures than any other in the world
I didn't realise we were talking about NetBSD;-) Linux has gradually become more portable, but still has large chunks of machine dependent code that would have been better implemented had Linus set out to write a portable kernel. Linux allegededly runs on a number of platforms where it was only ever able to boot diskless or single user at best, and where the support has disappeared thanks to no centralised build or test environment.
Linux throughout its history has almost always been the most performant in comparison with its contemporaries in terms of primitive operations like fork(), syscalls, context switches and the like.
Bullshit. FreeBSD wiped the floor with Linux on i386 until a few years ago. Solaris still wipes the floor with Linux on anything more sophisticated than Joe Internet Startup Inc and their Dell servers. Try benchmarking Linux and Solaris on an SMP Opteron sometime and you'll find out what I mean.
It was good enough for the likes of IBM, Intel, SGI, HP, NSA, NEC, Sony, etc to justify putting significant resources into, and most of those organisations continue to do so.
All that happened was that PC technology became just about good enough to perform serious server tasks, and the likes of IBM et. al. saw that pushing Linux was a cheaper option than porting their existing enterprise operating systems.
No, Linus gets an "A" in operating systems design. He is the chief designer and architect of this system.
Nope, Linus hasn't been heavily involved in the design of many of the subsystems that make up the kernel. VM? Rik van Riel and Andrea Achanegeli. Schedulers? Con Kolivo (sp?). VFS? Al Viro. And that's just a few examples.
Linus would have deserved that "F" in operating system design, but he wasn't writing his kernel to get grades on a computer course. If he had been then he probably wouldn't have written a crude, monolithic kernel that was totally unportable. Apart from the crudity of it, those were his explicit goals - to write a monolithic kernel that would run optimally on his 80386. (Bear in mind that the Linux kernel we know today is pretty far removed from that early version in design and implementation).
As for AT, he's a very smart guy. He writes books on operating system deign and networking that clearly describe quite complex topics. Even if you don't like the idea of microkernels, the "Operating Systems..." book that describes the Minix kernel is an excellent read.
I recently finished recording a full length album at a top London studio (Flood was mixing the latest Sophie Ellis-Bexter album next door). The standard setup in that studio unless you insist on analogue, is Protools running on a Mac. If it's the industry standard, then the industry better start looking for better software, as the fucker crashed frequently. We lost several takes of one particular track, which meant the studio had to give us some extra time for free to make up for it. The plugins in Protools maybe quite neat, but the core application is not that sophisticated - Ardour covers these areas very well.
Maybe the fact that I've never even heard of anything other than MySQL until today will give you an idea as to why I picked that first.
Maybe if you'd done some research before settling on MySQL you might have picked a better RDBMS. There again, given the sheer amount of MySQL books and apologists maybe not. As for "having all the jobs to your money-grubbing selves" I'd quite happily see more people coming into the industry with a modicum of knowledge. Rather than muppets like the bloke I dealt with today who considers an XML document valid if IE6 treats it as XML rather than plain text!
Here I thought I've been actually learning something
You may be learning the rudiments of SQL, relational databases and a web templating system, but you need to be aware of the limitations and quirks of MySQL. Too many people are leaving college having only used MySQL, and end up perpetuating the myth that MySQL is particularily fast in all but the simplest applications and that the missing features are somehow "unnecessary" in all circumstances. That last thing is what fucks me off about MySQL. For years Monty and co. said that foreign keys were a misfeature, implying that a couple of dodgy[1] coders at MySQL AB knew better than all the other companies and researchers who have decades of database analysis to back them up. When MySQL belatedly added foreign key support with the InnoDB table type, they suddenly stopped saying that database integrity should be handled at the application level. Cunts. Oh well, at least it keeps me in work, replacing shitty systems written by novices like you with something more robust and scalable.
[1] I say dodgy, because they release dubious changes at micro patch levels that break existing functionality.
Wikipedia is becoming the Restatement of the English Constitution, eh?
Not a restatement of the English constitution, as there isn't a constitution to restate. In Britain, the country is run using long standing conventions that are not necessarily codified anywhere. It's been a loooong time since I sat a politics exam, but there is a Victorian political writer who described really well the principles on which Britain is governed. The resulting system is very flexible, even the post of Prime Minister has no official definition, and governments have been run on a more collective basis in the past. Most conventions are nowadays codified in laws, and then tested in court - the independence of the judiciary is supposed to enforce a sanity filter on government and parliament. Slightly messy really.
Slashdot Mirror
Well, I'm an employee for a highly successful company, and we had no end of problems with MySQL. Deleting large numbers of rows from MyISAM tables with full text indexing would corrupt them. Running an ptimize on large tables (MyISAM and InnoDB) would corrupt them. Replication quietly fails rather than generating warnings if the syntax of the failover commands is not exactly right. The MySQL JDBC driver is terrible. The C++ code used in the admin tools is shockingly bad. Date and time columns cause all manner of problems. Need I go on?
We switched to PostgreSQL. Performance is better, replication works reliably and the JDBC driver is great.
It's not hard because there are dozens of Debian maintainers feverishly working day and night to keep it from breaking.
If it's true that it needs dozens of maintainers on Debian then the package management is poor. In the NetBSD pkgsrc tree, GNOME is maintained largley by one individual. It's much more likely that there's more people willing and able to work on the Debian packages, but I very much doubt that it absolutely requires dozens of them to keep up with GNOME releases. GNOME code has become very clean and portable since the release of version 2.0.
I also have serious doubts that NetBSD are still using XFree86
Yes, the version of X in the NetBSD source tree is still XFree86. This is because many platform specific changes have been made to the X code in the NetBSD tree, many of which had not found their way back into XFree86 before the X.Org fork. Getting the changes back into X.Org will become increasingly difficult as the code evolves, although much of it is localised into X server code that's unlikely to be receiving much attention from the X.Org guys, so fingers crossed ...
That said, many people are running X.Org on NetBSD, installing it via pkgsrc. I've done just that on my main machine, an i386 laptop, while my "fun" machines such as a Vax, SGI Indy and SparcStation run the default XFree86 (or no X at all in the case of the Vax).
Not true. Liberal political theory promotes a "rolling back" of the state so that it interferes in the life of citizens as little as possible. This doesn't preclude a welfare state, far from it, however it does suggest having one that does not rely on excessive taxation. the reason the conservatism of Thatcher and Bush has become known as "neo-liberalism" is because it has adopted the principle of rolling back the state. The difference is that this rolling back is geared towards the benefit of business rather than the individual.
For those puzzled, RMS's Liberal comment is in reference to Canada's Liberal party.
Whoever transcribed the interview, or RMS himself if the interview was conducted via email, should have written "liberal" with a small "l". In political writing a capitalised word like Liberal indicates a party, while the lowercase form indicates a theory or dogma. For example, Conservative would imply the political party when speaking about UK politics, whereas conservative would imply the political theory. Exceptions include political theories such as Marxism, which are named after an individual and should of course be capitalised.
To think that you can slashdot Microsofts site because of a puny WMV-file is naive.
It's really naive when you consider the that Microsoft don't host their own website. Akamai does.
It shouldn't even be on x86i64.
I'd be happy to run "one of the busiest websites on earth" using Opterons, which are basically 64bit x86 boxes. I wouldn't use Windows as the OS though ...
GCC is a very important piece of open source software, but it is time to see some real progress in C++
(list of features snipped)
Why not use a toolkit that gives you much more than a plain STL, something like Qt or Boost. Alternatively, why not use a language and environment that already has those features, like Java?
Who uses Solaris 10?
I assume you mean "uses it instead of Linux", what with this being Slashdot. How about people who've benchmarked it against Linux and found Solaris to scale better and more smoothly? Some of us like having beefy Sparc or Opteron SMP machines that perform predictably with Solaris, rather than the erratic behaviour we've seen with Linux on SMP Intel hardware. The 2.6.x Linux kernel has also been a serious disappointment in terms of reliability, a definite step back from 2.4.x.
In my experience the Linux binary compatiblity in NetBSD is perfect. Until I finally got around to building a native JDK, I was running the Sun JDK for Linux under NetBSD 2.0.2. It performed flawlessly, and I only switched to a native JDK because I was bored one afternoon and decided to see if it would compile. I have also run Linux binaries of Mozilla and Firefox to get support for the Flash plugin, and these also performed flawlessly under various versions of NetBSD.
I'm subscribed to the NetBSD-current mailing list, and I see regular messages about tweaks and improvements to the Linux syscall compatability in the development version of NetBSD. This gives me considerable confidence that changes to system calls in Linux are quickly picked up on by the NetBSD developers.
You could try NetBSD instead: http://www.netbsd.org/Ports/sparc64/. I'd recommend NetBSD 3.0, which has just gone from a beta to a release candidate, and includes support for all of the graphics cards found in Ultra workstations. A prebuilt release candidate can be found on the FTP mirrors under /pub/NetBSD-daily/
Hyundai is giving american car manufactures a run for the money, not only in quality but service, performance and reliability.
It's not difficult to give American car manufacturers a run for their money. I owned a 1999 Mustang coupe that was the shittiest car I've ever been lumbered with. My MGB GT was more reliable (yup, the Mustang was that bad), while my Jaguar XJS had better performance. As for servicing, San Francisco streets meant the Mustang's suspension was fucked within a month so the servicing needed to be cheap.
The fitness race has already picked a winner, IMHO: Solaris kernel with Debian user-space
Debian user-space wins the fitness race?!? You obvisouly haven't taken a look at the code for the bloated pig called glibc.
It's a pity that GNOME was written way back when instead of GNUstep being the free desktop of choice - had all that effort gone into GNUstep, it would have been pretty easy to target both Mac OS X and Linux/*BSD instead of having to write separate UI code for each.
Miguel de Icaza and the other founders of the GNOME project did consider GNUstep. Their conclusion was that the amount of work needed to finish GNUstep, and the added burden of staying compliant with future OpenStep specifications weren't worth the effort. They weren't to know that NeXT would find a massive new audience as the new MacOS ...
Hah! Java generics are laughable, especially if you ever used C++ templates.
Thanks I needed a laugh. Templates in C++ use a disgustingly unreadable syntax, and don't compensate for the bloated mess that the rest of the language represents. If I ever need to develop another server side system in C++ rather than Java I'll be shocked.
More importantly, "write once, run anywhere" really should be qualified as "write once, run anywhere the exact same JVM is installed and a decent script/command/launcher exists to properly start your application with all the flags, classpaths, memory management, and other configuration needed for this platform"
With all due respect, I feel you need to read up on how to package your applications better. If that fails I think you need to look at the quality of your code. I have a 200000 line Java codebase which, thanks to an MVC architecture, builds into a web application and a number of standalone applications (some are GUI tools, others are daemons). Thanks to a decent Ant build system along with a good understanding of Jar files, Log4J and webstart, I can deploy onto Windows, MacOS X and Solaris. No platform dependent tweaks at all, and no plaform specific problems encountered.
That reminds me of a story about how Cuba is filled with classic American cars from the 1940s and 1950s that are (for the most part) still well maintained and still on the road after all these years.
Getting a bit off topic, but many of those American cars are still running because the engines have been transplanted from a Russian Lada or Volga. If the same principle is applied in South Korea, then something like Linux will replace Windows when the localised version is too out of date.
This is an operating system that runs on more CPU architectures than any other in the world
I didn't realise we were talking about NetBSD ;-) Linux has gradually become more portable, but still has large chunks of machine dependent code that would have been better implemented had Linus set out to write a portable kernel. Linux allegededly runs on a number of platforms where it was only ever able to boot diskless or single user at best, and where the support has disappeared thanks to no centralised build or test environment.
Linux throughout its history has almost always been the most performant in comparison with its contemporaries in terms of primitive operations like fork(), syscalls, context switches and the like.
Bullshit. FreeBSD wiped the floor with Linux on i386 until a few years ago. Solaris still wipes the floor with Linux on anything more sophisticated than Joe Internet Startup Inc and their Dell servers. Try benchmarking Linux and Solaris on an SMP Opteron sometime and you'll find out what I mean.
It was good enough for the likes of IBM, Intel, SGI, HP, NSA, NEC, Sony, etc to justify putting significant resources into, and most of those organisations continue to do so.
All that happened was that PC technology became just about good enough to perform serious server tasks, and the likes of IBM et. al. saw that pushing Linux was a cheaper option than porting their existing enterprise operating systems.
No, Linus gets an "A" in operating systems design. He is the chief designer and architect of this system.
Nope, Linus hasn't been heavily involved in the design of many of the subsystems that make up the kernel. VM? Rik van Riel and Andrea Achanegeli. Schedulers? Con Kolivo (sp?). VFS? Al Viro. And that's just a few examples.
Linus would have deserved that "F" in operating system design, but he wasn't writing his kernel to get grades on a computer course. If he had been then he probably wouldn't have written a crude, monolithic kernel that was totally unportable. Apart from the crudity of it, those were his explicit goals - to write a monolithic kernel that would run optimally on his 80386. (Bear in mind that the Linux kernel we know today is pretty far removed from that early version in design and implementation).
As for AT, he's a very smart guy. He writes books on operating system deign and networking that clearly describe quite complex topics. Even if you don't like the idea of microkernels, the "Operating Systems ..." book that describes the Minix kernel is an excellent read.
ProTools is industry standard, period.
I recently finished recording a full length album at a top London studio (Flood was mixing the latest Sophie Ellis-Bexter album next door). The standard setup in that studio unless you insist on analogue, is Protools running on a Mac. If it's the industry standard, then the industry better start looking for better software, as the fucker crashed frequently. We lost several takes of one particular track, which meant the studio had to give us some extra time for free to make up for it. The plugins in Protools maybe quite neat, but the core application is not that sophisticated - Ardour covers these areas very well.
Maybe the fact that I've never even heard of anything other than MySQL until today will give you an idea as to why I picked that first.
Maybe if you'd done some research before settling on MySQL you might have picked a better RDBMS. There again, given the sheer amount of MySQL books and apologists maybe not. As for "having all the jobs to your money-grubbing selves" I'd quite happily see more people coming into the industry with a modicum of knowledge. Rather than muppets like the bloke I dealt with today who considers an XML document valid if IE6 treats it as XML rather than plain text!
Here I thought I've been actually learning something
You may be learning the rudiments of SQL, relational databases and a web templating system, but you need to be aware of the limitations and quirks of MySQL. Too many people are leaving college having only used MySQL, and end up perpetuating the myth that MySQL is particularily fast in all but the simplest applications and that the missing features are somehow "unnecessary" in all circumstances. That last thing is what fucks me off about MySQL. For years Monty and co. said that foreign keys were a misfeature, implying that a couple of dodgy[1] coders at MySQL AB knew better than all the other companies and researchers who have decades of database analysis to back them up. When MySQL belatedly added foreign key support with the InnoDB table type, they suddenly stopped saying that database integrity should be handled at the application level. Cunts. Oh well, at least it keeps me in work, replacing shitty systems written by novices like you with something more robust and scalable.
[1] I say dodgy, because they release dubious changes at micro patch levels that break existing functionality.
But the more you can tell me about your crazy system, the better, mate.
A Wikipedia search for "British Constitution" brings back some very informative articles.
Wikipedia is becoming the Restatement of the English Constitution, eh?
Not a restatement of the English constitution, as there isn't a constitution to restate. In Britain, the country is run using long standing conventions that are not necessarily codified anywhere. It's been a loooong time since I sat a politics exam, but there is a Victorian political writer who described really well the principles on which Britain is governed. The resulting system is very flexible, even the post of Prime Minister has no official definition, and governments have been run on a more collective basis in the past. Most conventions are nowadays codified in laws, and then tested in court - the independence of the judiciary is supposed to enforce a sanity filter on government and parliament. Slightly messy really.
Wow, a Scott Nudds posting.
(this an impossibly obscure comp.lang.c reference that no one will get).