Well, that image has been animated as such for as long as there has been a Gimp icon on Slashdot. With your million+ UID, it has been here longer then you have. I'm sure it doesn't welcome you.
So you are in a field that no one cares about.. You had a good run for 40 years, and that ended. And now you are mad? Sorry buddy, you should be grateful for what you got when you got it.
Very few games themselves have ever triggered me to do much hardware purchasing. Sure, after playing through _all_ of Doom 3, I got a new video card. But this, this is just amazing. If it runs under Cedega, I'm all over that. If not, fuck.. It'll be the first time.. First time this century.. that I'll slap Windows on my personal home system.
Well, any given line of code might be < 10 years old (and thus, not from Netscape Inc), but it wasn't like they started one day from 0 and went forward. A Ship of Theseus kinda thing.
Things like the build system, and the dialect (Well, maybe, "very specific and strange conventions") of C++ remain,
Its a question of what is "core framework" type stuff, and what is "actual application". Things like UI layout, interaction, networking, security, caching, rendering - as well as executing run time JS - is "core" functionality. I'd wager that >90% of binary code of Fx and, say, Thunderbird is the same.
And most of that core stuff is written in C++. Well, actually, its written in an obscure dialect of C++, developed when Netscape ran on a dozen various platforms, with mutually incomparable cpp compilers.
But that 10% that makes an application engine a web browser, or a mail client.. Most of that is written in Javascript. And most of it is "leaf" code, with not much cross calling, or dependencies that don't go through the underlying engine. Stuff that the just about total lack of "programming in the large" that Javascript has doesn't much matter.
The current legitimate requirement for anonymity today is an indication that we don't have free expression. Or to put thing another way, if you can think of a system where there is legitimate need for anonymity, then that system isn't totally free.
I recognize OSGi as "some Java thing", but don't much really understand exactly its benefits. Well, thats not quite true, it vaguely seems to play the same kind of role as say the Linux standards base, plus package management. I know that is pretty high level, but its close enough.
And Java (the VM part) plays a relatively similar role as OS kernels; memory, device management, security, etc.
Well, you know what? We have very good systems in place for "container" standards, and device and memory management. We call them operating systems. And they have decades of tuning and optimization and management tools (and trained people) behind them. Despite the "just run this war" mentality of most/all Java people, it is never that simple. And since it usually "just works" for them, the documentation around installation, dependencies, and such is poor. Generally, culturally, either (configure;make;make install) or (rpm -Uvh ) "just works" more frequently then Java stuff.
As an example of this, just about all the Atlassian products (Jira, confluence, bamboo, etc) come in both "war" and "Standalone" versions, the later of which is packaged including Tomcat (or Jetty), because (apparently) documenting how to get a "drop in" war working in an existing tomcat install is just too difficult.
The Java paradigm may be theoretically better then historical OSs, but the reality that we live in is that Java hasn't delivered on this front.
Because all the various "easy" Java management stuff ultimately still requires a host OS - with its on management stuff - Java only adds to the burden.
We have a way of abstracting out hardware. And modern operating systems (distributions) have an excellent ways of managing applications.
Hardware virtualization technologies such as the ESX version of VMWare (or say, IBM 360s from 1965), which leverage hardware support (which mainframes do better then PC class stuff, I grant) is fundamentally better then putting a VM on top of an OS.
The overhead on modern hardware with virtualization capabilities is marginal. The additional ability to manage that hardware efficiently outweighs that overhead alone. Add to that the fault tolerance and resource balancing you can get with a cluster of real servers, and wow.
Ill call that bluf. You mean the partitioning that MySQL has had for a couple of years? That partitioning? http://dev.mysql.com/doc/refman/5.1/en/partitioning.html
I'm getting mighty tired of your "just do it yourself" attitude on on all these disro stories.
If you want to just do it yourself, feel free. Start from scratch and compile everything yourself. You have that option.
But in a debate about the relative merits of distributions, it goes with out saying that the people involved merit others doing some kind of work for them.
They value some kind of integrated configuration / package management. They value some kind of coherent system layout. They value the ability to get most - if not all - of their system updates from one place. They value the community around that distro, and it being somewhat aligned with quite possibly arbitrary (and even not perfect) technical decisions of that distribution.
Red Hat and Fedora come configured against SELinux. Their packages include the metadata to actually drive SELinux and make it useful. Open/SUSE comes configured against AppArmour. Their packages include the metadata to actually drive AppArmor.
Choosing to hitch your wagon to SELinux/RH or AppArmor/SUSE would be a reasonable choice. Going with either will save you weeks - if not years - of hand tuning the various sets of security metadata. Either choosing RH and then adding AppArmor, or SUSE and adding SELinux would be nothing short of insanity.
Again.. People choose to use distributions partially because they already want a bunch of pedantic, arbitrary and tedious decisions made for them. In many cases, these decisions need not be "100%" right, and I doubt that anyone agrees 100% with their distro-of-choice choices. But, it would be stupid to choose to go one way, but then from day 0 fight with that distros design decisions.
Your assertion of a absolute truth is plainly false. While it is possible for other distros do do the same thing that they do, that is not at all the same thing as saying that they do do the same thing as SUSE.
Clearly, if all distros did the same thing, then there would be no reason for multiple distos. All distributions aim to be best at something special, and make some effort to differentiate themselves.
Distributions are designed around different goals, different sets of target users, built by different developers, working for different companies, to different ends. Quite possibly zero distributions modify none of their included packages, and all distributions modify heavily especially their "core" package and system management tools, at least adding on distinct high level tools. YOU vs RHN comes to mind here.
SLES, during installation, prompts you to join a Windows domain. I don't remember seeing this option the last time I tried Fedora. I do specifically remember seeing it in SUSE at about the same time as I last installed Fedora though. The Fedora people could do the same thing. But they don't (or didn't, at the same time as SUSE)
SUSE installs the Novell version of OpenOffice. This version exists specifically to better interoperate with MS office (and this polished version existed long before any formal Novell/MS partnerships). It is available for anyone to download - source included. There is nothing stopping other distributions from including it - but they don't.
How well these two specific examples work is a point I won't argue. But they are two things, off the top of my head, that are distinct about SUSE and their commitment to better interoperate with MS stuff.
An application/web/game/blah programmers job is to write programs to do that. A compiler programmers job is to right good compilers. They are separate and distinct tasks... And really are distinct jobs/trades.
If an application programmer needs top think about what the compiler is doing, then the compiler isn't doing its job well.
In any case, with processors (and good compilers) from the last decade, trying to second guess what the hardware will do and how the compiler will try and optimize things.. Its just as likely you will make things worse then better. With SMP machines seeming to be the default moving forward, it is even more critical that compilers are less shitty. Writing cross platform apps? Different OSs, compilers, hardware? You should be able to ignore this, iff the low level tools/os/lib developers did their job well.
In short: If "every" programmer needs to know about assembler, then some programmers are not doing their jobs.
-In all the discussions/debates I've seen about PG/MySQL, stability has never been brought up. And I've never had stability issues with MySQL, nor do I know of any one else that has.
-Standards compliance with database servers is next to pointless.. Its an interesting target, but all vendors provide not quite the same compliance level. Missing things, or extensions.
-Current, recent, and not-so-recent version of MySQL also have transaction and subselect support. I wouldn't install a 5 year old version today, and I don't expect to tomorrow.
And finally: your opinion counts very little (for me), and I doubt it counts much more for the general population.
But you never really my real question (which wasn't stated explicitly, I grant).. How is Suns purchase of MySQL AB going to make PG be used more?
Really, I still have not heard a good reason why one should use - in 2005, let alone today - PostgreSQL over MySQL. Technology wise I've never heard a compelling reason to use PostgreSQL that applies to recent (well, no so recent) versions.
Or are you now concerned about Sun breaking MySQL, or only having Sun as an option for support?
Perhaps it is true that Sun may be the only "official" support provider for MySQL, but there will still be the plethora of quality consultants and other large integrators who can still provide support. PostgreSQL continues to have exactly 0 official support providers - and I suspect a slightly smaller plethora of consultants. So how is this different then yesterday?
If it took exactly 8 hours and I knew that I would be away from a given computer for 8 or more hours.. No.,. That would still be unacceptable.
Sometimes FF needs inter-day restarts. Sometimes it crashes.. Or the OS does. Or something locks up and it needs to be killed. And ATI driver crash might cost me a minute a week.. But an 8 hour restart of my browser would really suck. In any case... Having to restart an application on a regular basis is insanity.
On fragmentation: I don't care. Mozilla is built on some fairly old and obscure code, I grant. The story of FF memory usage is one of developers trying very, very, hard to avoid doing any work. Sure they solved some very specific cases - after getting more users then necessary to produce very detailed error reports - and these small cases were resolved. But now - and this "memory fragmentation" revelation seems to be new - it turns out that perhaps there is a fundamental problem with low level FF/Moz memory handling. Why has it taken years for someone to say "Wow. FF memory management really sucks and there are dozens of small examples of how this is generally true. Perhaps there is a fundamental problem?". It seems to me no one was taking a holistic view of the project. Individual developers may have tweaked very specific issues in their little world.. But why did no leader step up and say "hey, people, what the fuck?"
Because closing it takes more then 0 time and opening it takes more then 0 time. Software is supposed to work for me, not me for it. You are asking the wrong question. The question is why is this insanity of a application requiring to be restarted continues to be acceptable?
Then don't put any settings in (e.g.)/etc/ports, and continue on as you are now. Don't use the option.
In your thought experiment one must become root to to bind to:80, preventing users from launching their own web server. In my idea, one must become apache (or wwwrun) (or also root) to bind to:80. "Getting wwwrun" may be slightly easier then getting root, but I think about the same. Slight loss in security there. On the flip side, Apache now never needs to be root. Slight increase in security there. Apache might not be a good test, its reasonably trusted. But what about Apache with local users running CGI or PHP scripts? Or one of your alternate http servers? Or a written-in-perl spam proxy on:25?
I consider not running things as root a more significant risk then the possibility of "users" breaking a nologin user account. On any "important" system, the only users would be administrators anyway, wouldn't it?
The argument against changing is simple: change takes effort. This invariably means money, be it small, or huge. Also, the "solution" may not actually be any better. Ethanol, say.. Yay, no more non-renewable oil. Well, pollution wise, its worse off. Biodiesel? Again, marginally worse off pollution wise. Tidal power? Excellent, unless you are a fish. If everyone in North America switches to hybrid vehicles, we are still fucked, because developing nations (e.g. China) will consume that much oil and more in the near future.
In your thought experiment of drink/dont drink, what if the guy who said "don't drink, you will die" has been saying that for years, yet people never seem to die when they do drink? Then you would start to think that buddy is a wack job.
The reality is that scientists and pundents on both side just don't know what the fuck is going on. If in 1981 you produce a nice mathematical model that given the stats from 1400-1970 correctly predicts from 1971-1980, that isn't bad. But then in 1991 you see that 1981-1990 is wrong. So you update the model, and now 1400-1990 is right. Better. Then in 2000, you see that it was wrong again. So you update the model. And in 2007 you see that it was wrong for 2001-2006. Perhaps your fundamental assumptions about reality are flawed, and while you can tweak all you want, your model is fundamentally broken. How about that?
Climate scientists can even effectively predict if we will have a "good" or "bad" winter. And you want to fundamentally change the modern economy based on what their 200 year predictions are? I'm not suggesting that we all be big environmental dicks, neither however is doomsday coming tomorrow.
You assertions are invalid. First off, not all OSs have this restriction. It assumes that only smart and non-evil people are host administrators. And for the majority of hosts on the internet, neither of theses is true.
The can is open, there are worms everywhere. Its has been at least 25 years (if ever) since you could make the assumption that remote hosts were run by smart people and against a policy that was similar to yours. Your assertion might only be valid in some magical fairy land, which I doubt has ever existed.
Slashdot Mirror
This was back in the days that computers were flaky and needed the occasional blood sacrifice to keep them happy.
Computer manufactures knew this. Those hard to reach things behind lots of sharp edges... Features!
Well, that image has been animated as such for as long as there has been a Gimp icon on Slashdot. With your million+ UID, it has been here longer then you have. I'm sure it doesn't welcome you.
So you are in a field that no one cares about.. You had a good run for 40 years, and that ended. And now you are mad? Sorry buddy, you should be grateful for what you got when you got it.
Yeah. And if you define 2 as being 2=17, then 2+2 is 34. Sofuckingwhat?
Wow. Just wow.
Very few games themselves have ever triggered me to do much hardware purchasing. Sure, after playing through _all_ of Doom 3, I got a new video card. But this, this is just amazing. If it runs under Cedega, I'm all over that. If not, fuck.. It'll be the first time.. First time this century.. that I'll slap Windows on my personal home system.
Well, any given line of code might be < 10 years old (and thus, not from Netscape Inc), but it wasn't like they started one day from 0 and went forward. A Ship of Theseus kinda thing.
Things like the build system, and the dialect (Well, maybe, "very specific and strange conventions") of C++ remain,
Its a question of what is "core framework" type stuff, and what is "actual application". Things like UI layout, interaction, networking, security, caching, rendering - as well as executing run time JS - is "core" functionality. I'd wager that >90% of binary code of Fx and, say, Thunderbird is the same.
And most of that core stuff is written in C++. Well, actually, its written in an obscure dialect of C++, developed when Netscape ran on a dozen various platforms, with mutually incomparable cpp compilers.
But that 10% that makes an application engine a web browser, or a mail client.. Most of that is written in Javascript. And most of it is "leaf" code, with not much cross calling, or dependencies that don't go through the underlying engine. Stuff that the just about total lack of "programming in the large" that Javascript has doesn't much matter.
The current legitimate requirement for anonymity today is an indication that we don't have free expression. Or to put thing another way, if you can think of a system where there is legitimate need for anonymity, then that system isn't totally free.
I recognize OSGi as "some Java thing", but don't much really understand exactly its benefits. Well, thats not quite true, it vaguely seems to play the same kind of role as say the Linux standards base, plus package management. I know that is pretty high level, but its close enough.
And Java (the VM part) plays a relatively similar role as OS kernels; memory, device management, security, etc.
Well, you know what? We have very good systems in place for "container" standards, and device and memory management. We call them operating systems. And they have decades of tuning and optimization and management tools (and trained people) behind them. Despite the "just run this war" mentality of most/all Java people, it is never that simple. And since it usually "just works" for them, the documentation around installation, dependencies, and such is poor. Generally, culturally, either (configure;make;make install) or (rpm -Uvh ) "just works" more frequently then Java stuff.
As an example of this, just about all the Atlassian products (Jira, confluence, bamboo, etc) come in both "war" and "Standalone" versions, the later of which is packaged including Tomcat (or Jetty), because (apparently) documenting how to get a "drop in" war working in an existing tomcat install is just too difficult.
The Java paradigm may be theoretically better then historical OSs, but the reality that we live in is that Java hasn't delivered on this front.
Because all the various "easy" Java management stuff ultimately still requires a host OS - with its on management stuff - Java only adds to the burden.
We have a way of abstracting out hardware. And modern operating systems (distributions) have an excellent ways of managing applications.
Hardware virtualization technologies such as the ESX version of VMWare (or say, IBM 360s from 1965), which leverage hardware support (which mainframes do better then PC class stuff, I grant) is fundamentally better then putting a VM on top of an OS.
The overhead on modern hardware with virtualization capabilities is marginal. The additional ability to manage that hardware efficiently outweighs that overhead alone. Add to that the fault tolerance and resource balancing you can get with a cluster of real servers, and wow.
I'm confused. You seem to have foiled my ability to parse numbers.
80,000 what per hour?
Ill call that bluf. You mean the partitioning that MySQL has had for a couple of years? That partitioning?
http://dev.mysql.com/doc/refman/5.1/en/partitioning.html
I'm getting mighty tired of your "just do it yourself" attitude on on all these disro stories.
If you want to just do it yourself, feel free. Start from scratch and compile everything yourself. You have that option.
But in a debate about the relative merits of distributions, it goes with out saying that the people involved merit others doing some kind of work for them.
They value some kind of integrated configuration / package management. They value some kind of coherent system layout. They value the ability to get most - if not all - of their system updates from one place. They value the community around that distro, and it being somewhat aligned with quite possibly arbitrary (and even not perfect) technical decisions of that distribution.
Red Hat and Fedora come configured against SELinux. Their packages include the metadata to actually drive SELinux and make it useful. Open/SUSE comes configured against AppArmour. Their packages include the metadata to actually drive AppArmor.
Choosing to hitch your wagon to SELinux/RH or AppArmor/SUSE would be a reasonable choice. Going with either will save you weeks - if not years - of hand tuning the various sets of security metadata. Either choosing RH and then adding AppArmor, or SUSE and adding SELinux would be nothing short of insanity.
Again.. People choose to use distributions partially because they already want a bunch of pedantic, arbitrary and tedious decisions made for them. In many cases, these decisions need not be "100%" right, and I doubt that anyone agrees 100% with their distro-of-choice choices. But, it would be stupid to choose to go one way, but then from day 0 fight with that distros design decisions.
Cynical Idealist, indeed. Politics and government are not the same thing.
Because forcing an election that you are not prepared to fight, or are prepared to fight, but know you cant win, its politically stupid.
Your assertion of a absolute truth is plainly false. While it is possible for other distros do do the same thing that they do, that is not at all the same thing as saying that they do do the same thing as SUSE.
Clearly, if all distros did the same thing, then there would be no reason for multiple distos. All distributions aim to be best at something special, and make some effort to differentiate themselves.
Distributions are designed around different goals, different sets of target users, built by different developers, working for different companies, to different ends. Quite possibly zero distributions modify none of their included packages, and all distributions modify heavily especially their "core" package and system management tools, at least adding on distinct high level tools. YOU vs RHN comes to mind here.
SLES, during installation, prompts you to join a Windows domain. I don't remember seeing this option the last time I tried Fedora. I do specifically remember seeing it in SUSE at about the same time as I last installed Fedora though. The Fedora people could do the same thing. But they don't (or didn't, at the same time as SUSE)
SUSE installs the Novell version of OpenOffice. This version exists specifically to better interoperate with MS office (and this polished version existed long before any formal Novell/MS partnerships). It is available for anyone to download - source included. There is nothing stopping other distributions from including it - but they don't.
How well these two specific examples work is a point I won't argue. But they are two things, off the top of my head, that are distinct about SUSE and their commitment to better interoperate with MS stuff.
To get horses to jump over fences, and they want to do it with elemental carbon? Craziness, I say!
I could not disagree more so.
An application/web/game/blah programmers job is to write programs to do that. A compiler programmers job is to right good compilers. They are separate and distinct tasks... And really are distinct jobs/trades.
If an application programmer needs top think about what the compiler is doing, then the compiler isn't doing its job well.
In any case, with processors (and good compilers) from the last decade, trying to second guess what the hardware will do and how the compiler will try and optimize things.. Its just as likely you will make things worse then better. With SMP machines seeming to be the default moving forward, it is even more critical that compilers are less shitty. Writing cross platform apps? Different OSs, compilers, hardware? You should be able to ignore this, iff the low level tools/os/lib developers did their job well.
In short: If "every" programmer needs to know about assembler, then some programmers are not doing their jobs.
In reverse order:
-In all the discussions/debates I've seen about PG/MySQL, stability has never been brought up. And I've never had stability issues with MySQL, nor do I know of any one else that has.
-Standards compliance with database servers is next to pointless.. Its an interesting target, but all vendors provide not quite the same compliance level. Missing things, or extensions.
-Current, recent, and not-so-recent version of MySQL also have transaction and subselect support. I wouldn't install a 5 year old version today, and I don't expect to tomorrow.
And finally: your opinion counts very little (for me), and I doubt it counts much more for the general population.
But you never really my real question (which wasn't stated explicitly, I grant).. How is Suns purchase of MySQL AB going to make PG be used more?
PostgreSQL adoption: why?
Really, I still have not heard a good reason why one should use - in 2005, let alone today - PostgreSQL over MySQL. Technology wise I've never heard a compelling reason to use PostgreSQL that applies to recent (well, no so recent) versions.
Or are you now concerned about Sun breaking MySQL, or only having Sun as an option for support?
Perhaps it is true that Sun may be the only "official" support provider for MySQL, but there will still be the plethora of quality consultants and other large integrators who can still provide support. PostgreSQL continues to have exactly 0 official support providers - and I suspect a slightly smaller plethora of consultants. So how is this different then yesterday?
If it took exactly 8 hours and I knew that I would be away from a given computer for 8 or more hours.. No.,. That would still be unacceptable.
Sometimes FF needs inter-day restarts. Sometimes it crashes.. Or the OS does. Or something locks up and it needs to be killed. And ATI driver crash might cost me a minute a week.. But an 8 hour restart of my browser would really suck. In any case... Having to restart an application on a regular basis is insanity.
On fragmentation: I don't care. Mozilla is built on some fairly old and obscure code, I grant. The story of FF memory usage is one of developers trying very, very, hard to avoid doing any work. Sure they solved some very specific cases - after getting more users then necessary to produce very detailed error reports - and these small cases were resolved. But now - and this "memory fragmentation" revelation seems to be new - it turns out that perhaps there is a fundamental problem with low level FF/Moz memory handling. Why has it taken years for someone to say "Wow. FF memory management really sucks and there are dozens of small examples of how this is generally true. Perhaps there is a fundamental problem?". It seems to me no one was taking a holistic view of the project. Individual developers may have tweaked very specific issues in their little world.. But why did no leader step up and say "hey, people, what the fuck?"
Because closing it takes more then 0 time and opening it takes more then 0 time. Software is supposed to work for me, not me for it. You are asking the wrong question. The question is why is this insanity of a application requiring to be restarted continues to be acceptable?
Most car mechanics do not know how to do metal machining. So fucking what? They are different jobs.
Then don't put any settings in (e.g.) /etc/ports, and continue on as you are now. Don't use the option.
:80, preventing users from launching their own web server. In my idea, one must become apache (or wwwrun) (or also root) to bind to :80. "Getting wwwrun" may be slightly easier then getting root, but I think about the same. Slight loss in security there. On the flip side, Apache now never needs to be root. Slight increase in security there. Apache might not be a good test, its reasonably trusted. But what about Apache with local users running CGI or PHP scripts? Or one of your alternate http servers? Or a written-in-perl spam proxy on :25?
In your thought experiment one must become root to to bind to
I consider not running things as root a more significant risk then the possibility of "users" breaking a nologin user account. On any "important" system, the only users would be administrators anyway, wouldn't it?
The argument against changing is simple: change takes effort. This invariably means money, be it small, or huge. Also, the "solution" may not actually be any better. Ethanol, say.. Yay, no more non-renewable oil. Well, pollution wise, its worse off. Biodiesel? Again, marginally worse off pollution wise. Tidal power? Excellent, unless you are a fish. If everyone in North America switches to hybrid vehicles, we are still fucked, because developing nations (e.g. China) will consume that much oil and more in the near future.
In your thought experiment of drink/dont drink, what if the guy who said "don't drink, you will die" has been saying that for years, yet people never seem to die when they do drink? Then you would start to think that buddy is a wack job.
The reality is that scientists and pundents on both side just don't know what the fuck is going on. If in 1981 you produce a nice mathematical model that given the stats from 1400-1970 correctly predicts from 1971-1980, that isn't bad. But then in 1991 you see that 1981-1990 is wrong. So you update the model, and now 1400-1990 is right. Better. Then in 2000, you see that it was wrong again. So you update the model. And in 2007 you see that it was wrong for 2001-2006. Perhaps your fundamental assumptions about reality are flawed, and while you can tweak all you want, your model is fundamentally broken. How about that?
Climate scientists can even effectively predict if we will have a "good" or "bad" winter. And you want to fundamentally change the modern economy based on what their 200 year predictions are? I'm not suggesting that we all be big environmental dicks, neither however is doomsday coming tomorrow.
You assertions are invalid. First off, not all OSs have this restriction. It assumes that only smart and non-evil people are host administrators. And for the majority of hosts on the internet, neither of theses is true.
The can is open, there are worms everywhere. Its has been at least 25 years (if ever) since you could make the assumption that remote hosts were run by smart people and against a policy that was similar to yours. Your assertion might only be valid in some magical fairy land, which I doubt has ever existed.