... well, If you are that advanced user, I guess you really should try Linux, its command line and scripts.
I was that kind of user of Windows, knowing every shortcut, reading every.txt in %WINDIR%. Then I found that on Linux that all is much easier to access. Even more: you can put all that stuff into script and use it on infinitely easier. And no need to dig forums for tricks - it is all in man pages.
Vista still provide neither macro recorder nor simple scripting facility. They improved a lot on the front - cmd.exe alredy in XP was somewhat usable - yet, all the new stuff in Vista is way too radical and way too complicated. Writing scripts, for every second line googling or searching local help, is the dead end. *nix automation is much simpler and yet more powerful.
With only difference that you will not get Ubuntu on your PC unless especially asked. And you would pay $0 for it to be installed. And thus you have no contemplations about installing some other Linux (you might like better) on it. Or even buying Windows for it.
If you do not like 8.10, you can go on using 8.04 or 7.10 or even earlier. You might be surprised how many people do not upgrade their *buntus since it works for them already.
The point of RTFA that people prefer XP to Vista, yet are forced to pay for it hefty extra, which is btw much much higher than cost of OEM XP itself. If it's not extortion, I do not know what it is. IMNSHO M$ is shooting itself in foot. People want to buy the OS, yet, they will not sell it to them: $150 tag is precisely that: message from Redmond "we want you get Vista instead". Absurd. Second most absurd business decision I have ever seen. (First - Sony suing out of business company for selling PSPs)
Besides CPAN, Perl was one of the first languages to integrate advanced data structures - lists and hashes - directly into language itself. And not some half-assed implementation - e.g. C++'s STL or Lisp' lists - but really really good implementation, supported by many standard functions and (most importantly) internal optimizer.
Last, but not least, Perl is quite well performing. Compromise fitting most tasks: scripts are loaded relatively fast (e.g. compared to Python), yet if you use structures intelligently, it will run very fast.
All that together, with Perl's pragmatical approach, you have a tool which easily scales from irreplaceable "perl -pe" one-liners to relatively huge projects. And in many cases, huge projects start as one liner scripts. That's where I'm addicted to Perl: if you know what you do, you can write short but powerful scripts in few seconds. And if you need, you can easily improve the one liner into some good tool, usable by other too.
As noted by many Perl fans (like I am) you do not write in Perl - you think in Perl. It is language without any artificial barriers between you and resources you need to accomplish your task. That's why it is so hard to get off the Perl.
P.S. Can't compare to Ruby, since I haven't used it. Few examples I have seen before hadn't stroke me as anything radically new or more useful/practical than Perl.
It must be humiliating to have to schedule regular downtime just to amortise the cost of Windows' sheer stupidity.
You seem never had to deal with usual corporate IT.
They take special pride for their own incompetence. After all, bigger impact their actions have, more important they feel themselves in company.
And management compatible "sweet talk" seems to be first thing they are taught at the "IT inCompetence Courses" so they always have ready nice plain explanation to management, sometimes with PowerPoint slides.
It's kind of funny, because we also have bunch of Unix and hordes of Linux boxes. All Windows servers go down periodically, mostly scheduled downtimes, sometimes unscheduled. Unix servers mostly rebooted due to heavy abuse of resources and fact that our IT still doesn't know how to kill a process. Linux servers... in past year they rebooted two of them to replace faulty RAM modules.
Compared to background indexers in Vista and Mac OS X, Linux and Win2k run nothing in background. Even more most indexers I have seen on Linux allow to suspend or schedule the indexing process. On Vista and Mac OS X you can't do that - it is just there doing all the time something with your disk...
You raise valid point, though/me sarcastic pushes me to rather write something like "probably they couldn't boot it on the hardware"...
Also note that Sun isn't really advertising their OpenSolaris well. Lots of people who use Solaris on daily basis have absolutely no clue about its open source cousin for Intel processors - and heard first about OpenSolaris from me, Linuxoid.
But it would be really interesting to see whether the Solaris' optimizations bring something or not.
I had in past run a Windows server for about 2 years. And it was terribly slow, despite the fact that in the beginning it was blazingly fast. I believe it was O&O Defrag which actually returned the server to life: downtime on Sunday with boot time defragmentation did the miracle.
Same thing in the company I work for right now: IT recently took off net two file servers and exchange server to defragment file systems, because FS performance went considerably down. Folks have said that they "lose to fragmentation 30% of FS performance," meaning that system works about twice faster after defragmentation. That's why they schedule at least one down time for every windows server in company.
Whatever synthetic benchmarks people perform - it is irrelevant.
Actually few games are truly random at low level. On many occasions I was faced with trivial matter that "pseudo" random numbers are not really all that random.
The mentioned above problem is easily visible in aforementioned Tetris example and is direct result of poorness of random numbers.
On one occasion, analyzing one source code, I have found clever trick with premade chains of "random" numbers: applying the chains twice (two level indexing), overall the random generator had for numbers in 0..127 range loop far above 32k and all delivered numbers were well distributed: as per definition there were no same number streaks longer than 3.
As gamer though, I'd say, too much randomsness is often very distracting. Some console RPGs (e.g. mentioned above FF) are especially affected. If you look at real life, few events around us are really random...
As fan of SRPG and strategical games in general, I prefer games where random events are normally consequences of your past actions. Or even games like Heroes of Might and Magic where there are essentially no randomness in battles (or it is always in known bounds) so that outcome largely depends on player's capabilities.
To me, better feature would be to save history for only specified sites and/or to never save anything from specified sites.
There are some sites I visit only to check something periodically - and I do not want them to be in history. Or sites which force you for every little thing to go to new page. I simply do not need them to clog my history.
I do not think that after C# learning Python would bring anything new. I'd say one need to know at least two orthogonal in their concepts languages. That's why I actually brought up C & Perl in another comment.
Or probably my thought would be clearer this way: Python and C# have lots of common concepts and "learning" Python would amount to nothing more than learning new syntax of doing stuff the same way. C or Perl are quite orthogonal and force you to think and approach problems differently.
For language complementing C# I would advise Perl, mainly because Perl is well supported across all platforms, thus your knowledge will not be locked to Linux, but also might benefit Windows side of work.
I second. I'm pretty appalled on how clueless new young programmers on resource management.
N.B. As I was taught 15 years ago, programming is art of resource management. That was my teacher's way of saying the famous "algorithms + data structures == programmes," but pitching that all programs all the time use system resources (CPU, memory, files, etc) and your programs will always be as good as your understanding of the resources.
C, if nothing else, teaches how to manage resources and resource management skills are pretty much what defines a good programmer.
After you have grasped on how all stuff is managed and how it really works under the hood, Perl becomes irreplaceable tool at helping managing them very easily. Perl documentation is excellent on telling you precisely what Perl does and how much of what overhead the given code has.
I bring up resource management because it feels stupid that every time pure Java/C# programmers have problems they come to me: most questions amount to "we do this but whole system explodes" (because e.g. they have tried to allocate and object for 0... MAX_INT). And when you tell them that you have to eventually free the memory, they act surprised since they have thought that Java/C# handles that all for them and they expect it to e.g. automagically clear the list which keeps references to all the object which code doesn't intend to use anymore. This is primitive template example, but happened more than once to me.
Understanding resource management is quite tiresome and for smaller projects is overkill. But for anything large enough you have to know it and you have to understand it. I haven't seen a single large Java project written solely by pure Java programmers: very very likely that they have number of ex-C folks on the team. I guess the same goes for C# and Python.
Knowing C and Perl - not as languages - but as tools to work with system resources is invaluable knowledge. Learning another way to express your thoughts gives you nothing (I know probably more than 20 languages), but the very and most important is the knowledge of the system resources and how to work with them properly. Some standard libraries (e.g. Java) had put abstraction level way too high to allow to learn about system resources and to write efficient programs.
1. Install Google Toolbar
2. Google Toolbar > Settings > Options > Buttons > Add Button > Enter in search field "Pirate Bay" > pick one of the buttons to add to your google tool bar.
3. Go to Amazon
4. Select the name of product
5. Right mouse click to call pop up menu > More Search Types > middle mouse click on added in step #2 button.
IP's just need to be charged for on a early basis. Start with $1 per year per ip to EVERYONE who owns an IP's and you'll see the "IP Shortage" vanish overnight.
You miss the point.
Internet now has much much more participants than 10 years ago.
IPv4 are running out because compromise was made on management side. Essentially, Internet management is loose because people want to encourage participation. Do not forget that Internet is merely set of connected proprietary networks, operated by all possible entities. And every of the entities wants to have some address space to grow.
I frankly do not see any problem with going to IPv6 as most of the network equipment was ready for several years now.
There's no reason every person on earth needs an IP. Nat+uPNP is perfectly capable and 100% backwords compatible.
Had you ever tried to operate middle to large sized network, you would understand that adding extra layer at every interconnect would make Internet infrastructure at least twice more expensive. I'd say that costs would at least quadruple since such smart equipment, supporting dynamic resource management, costs at least 4 times more if compared to its plain version. (e.g. compare prices on "switch" vs. "managed switch" to see the difference).
Making Internet infrastructure expensive => killing Internet.
P.S. And do not get me started on the "uPNP" bogosity which doesn't even belong to the discussion...
I don't see any of this changing in UNIX any time soon, until it's not actually UNIX any more; the best anyone is going to be able to do is namespace folding hacks, for which they will need shell (and anything else that does globbing) cooperating in the versioning "conspiracy". This is not really something doable at the kernel level without some serious namespace escape trickery.
I do not think that there is a real challenge implementing versioning file system. The problems are well known and their solutions also present in distributed revision control systems.
The problem like always is that "one size never fits all."
Revision control evolves much faster than file systems and putting it into a file system would establish a barrier to amount of customizations and improvements which could be made. Essentially version control and file systems solve two different orthogonal tasks: former is responsible for preserving modification history, later - for reliably storing information.
As I have witnessed many times in business, revision control is very crucial to business process and often the two are tightly coupled (think distribution and demarcation of responsibilities). As result, business must have the flexibility to change revision control according to new market conditions at any moment of time. Putting that stuff into kernel would essentially force to cede control over software development process to some 3rd party. Nobody would ever do it.
Exemplary is my actual employer where process tightly integrates revision control with issue tracking system but also essentially formalizes whole process: from development to testing to customer delivery to maintenance. Everybody has own place in the process and rarely you have the question who is responsible for what. With few gotchas it works pretty well and allows to track what/when/how was done. (I called it already on couple of occasions "todo list on steroids on steroids.")
I agree with that. OS is quite performer and their compiler is state of the art. I'd say they have good scheduler and pthreads implementation is fastest I have ever seen.
Yet. Before, I was swearing at Solaris user tools. Until I haven't started working with HP-UX. Their user tools are even more primitive and counter-intuitive. sed/grep/awk/sort/uniq/friends often fails with fancy meaningless (or "dead end" type of) messages. (e.g. "line is too long").
Spending 1 hour every day writing some scripts to do something what under Linux is available for free or working around some decade old bogosity vs. very fast CPU/OS/compiler - is tough choice. I tend to choose system with better user tools because people do less stupid mistakes there and overall business process then flows smoother.
From that point of view HP-UX is quite poor performer.
P.S. And they were last (even later than Mac OS X!!) to become UNIX'03 certified.
Also, they have used several openly pro-Intel applications: Cinebench and M$.Net.
Cinebench never hid the fact that they optimize for Intel and if you want to have best performance you need to buy Intel CPUs.
M$.Net XML benchmark - M$ C/C++ compiler and libraries in many parts use Intel's hand written asm code. And it always produced code optimized for Intel architectures.
You didn't because you have gotten so used to it that you don't see it any more.
I didn't notice because I do not care about that.
I can't check what you say, as I do not have M$Office at hand - only portable OO.o.
The point here is that for the small things you really do not care about all the bogosities of VBA. And I can recall many of them all too well: I had a course on VBA in versity. I can't (nor want to) write in VB, yet I remember many of those bumps one had to jump over to make something working in VBA.
What I'm really pissed about OO.o is that in M$Office I didn't have to be bothered by all the bogosities of VBA but it just worked for trivial things with some amount of copy-pasting. As some expert had showed me before, StarBasic is much much more faster, yet to archive that level of performance simplicity was sacrificed, making StarBasic unacceptable for all those who do not want to invest weeks (or do not have that time) in digging up documentation and examples.
OO.o, to compete against VBA, needs something comparable in simplicity. After all, most of the automation done in word processors is really really trivial: replace that with that; set that attribute on selection or count that and put it in here. This are one/two liner in VBA (what is top of my VB capabilities), yet in StarBasic this is Java-like Object Oriented monstrosity.
the kernel application binary interfaces are a moving target.
That's why we have glibc, which abstracts that ABI from applications.
Kernel driver interface - the horse was already beaten to death many times ( see here ).
a consistent configuration system, to enable distribution;
Windows tried that with Registry - and it didn't worked. And it will never work since "one size never fits all" requirements of all applications.
native file versioning;
Was tried many times before and failed miserably. As long as majority of files are blobs, versioning on level of file system makes no sense. Versioning on level of applications is implemented already more or less everywhere it was needed and SVN/git is there for the rest of applications.
[...] I fail to understand the "several minutes" issue when opening up docs [...]
Dunno what gp meant, but can speak from my experience.
It happens in some configurations when opening documents from network shares. First I thought that it was some server problem, but then OO.o opened the same documents instantly. And M$Office was actually hanging hard for 1.5-2.5 minutes. "Hanging hard" means that Windows wasn't allowing to kill the application leading to worst case of document being locked on server for some time.
Was one of the last drops, which triggered one of my past employers to migrate completely to OO.o internally, keeping M$Office for communication with partners only.
Slashdot Mirror
I was that kind of user of Windows, knowing every shortcut, reading every .txt in %WINDIR%. Then I found that on Linux that all is much easier to access. Even more: you can put all that stuff into script and use it on infinitely easier. And no need to dig forums for tricks - it is all in man pages.
Vista still provide neither macro recorder nor simple scripting facility. They improved a lot on the front - cmd.exe alredy in XP was somewhat usable - yet, all the new stuff in Vista is way too radical and way too complicated. Writing scripts, for every second line googling or searching local help, is the dead end. *nix automation is much simpler and yet more powerful.
With only difference that you will not get Ubuntu on your PC unless especially asked. And you would pay $0 for it to be installed. And thus you have no contemplations about installing some other Linux (you might like better) on it. Or even buying Windows for it.
If you do not like 8.10, you can go on using 8.04 or 7.10 or even earlier. You might be surprised how many people do not upgrade their *buntus since it works for them already.
The point of RTFA that people prefer XP to Vista, yet are forced to pay for it hefty extra, which is btw much much higher than cost of OEM XP itself. If it's not extortion, I do not know what it is. IMNSHO M$ is shooting itself in foot. People want to buy the OS, yet, they will not sell it to them: $150 tag is precisely that: message from Redmond "we want you get Vista instead". Absurd. Second most absurd business decision I have ever seen. (First - Sony suing out of business company for selling PSPs)
Check here.
Besides CPAN, Perl was one of the first languages to integrate advanced data structures - lists and hashes - directly into language itself. And not some half-assed implementation - e.g. C++'s STL or Lisp' lists - but really really good implementation, supported by many standard functions and (most importantly) internal optimizer.
Last, but not least, Perl is quite well performing. Compromise fitting most tasks: scripts are loaded relatively fast (e.g. compared to Python), yet if you use structures intelligently, it will run very fast.
All that together, with Perl's pragmatical approach, you have a tool which easily scales from irreplaceable "perl -pe" one-liners to relatively huge projects. And in many cases, huge projects start as one liner scripts. That's where I'm addicted to Perl: if you know what you do, you can write short but powerful scripts in few seconds. And if you need, you can easily improve the one liner into some good tool, usable by other too.
As noted by many Perl fans (like I am) you do not write in Perl - you think in Perl. It is language without any artificial barriers between you and resources you need to accomplish your task. That's why it is so hard to get off the Perl.
P.S. Can't compare to Ruby, since I haven't used it. Few examples I have seen before hadn't stroke me as anything radically new or more useful/practical than Perl.
It must be humiliating to have to schedule regular downtime just to amortise the cost of Windows' sheer stupidity.
You seem never had to deal with usual corporate IT.
They take special pride for their own incompetence. After all, bigger impact their actions have, more important they feel themselves in company.
And management compatible "sweet talk" seems to be first thing they are taught at the "IT inCompetence Courses" so they always have ready nice plain explanation to management, sometimes with PowerPoint slides.
It's kind of funny, because we also have bunch of Unix and hordes of Linux boxes. All Windows servers go down periodically, mostly scheduled downtimes, sometimes unscheduled. Unix servers mostly rebooted due to heavy abuse of resources and fact that our IT still doesn't know how to kill a process. Linux servers ... in past year they rebooted two of them to replace faulty RAM modules.
Compared to background indexers in Vista and Mac OS X, Linux and Win2k run nothing in background. Even more most indexers I have seen on Linux allow to suspend or schedule the indexing process. On Vista and Mac OS X you can't do that - it is just there doing all the time something with your disk...
You raise valid point, though /me sarcastic pushes me to rather write something like "probably they couldn't boot it on the hardware"...
Also note that Sun isn't really advertising their OpenSolaris well. Lots of people who use Solaris on daily basis have absolutely no clue about its open source cousin for Intel processors - and heard first about OpenSolaris from me, Linuxoid.
But it would be really interesting to see whether the Solaris' optimizations bring something or not.
LOL.
I had in past run a Windows server for about 2 years. And it was terribly slow, despite the fact that in the beginning it was blazingly fast. I believe it was O&O Defrag which actually returned the server to life: downtime on Sunday with boot time defragmentation did the miracle.
Same thing in the company I work for right now: IT recently took off net two file servers and exchange server to defragment file systems, because FS performance went considerably down. Folks have said that they "lose to fragmentation 30% of FS performance," meaning that system works about twice faster after defragmentation. That's why they schedule at least one down time for every windows server in company.
Whatever synthetic benchmarks people perform - it is irrelevant.
Long term real life experience tells otherwise.
I can only add small tidbit from developer pov.
Actually few games are truly random at low level. On many occasions I was faced with trivial matter that "pseudo" random numbers are not really all that random.
The mentioned above problem is easily visible in aforementioned Tetris example and is direct result of poorness of random numbers.
On one occasion, analyzing one source code, I have found clever trick with premade chains of "random" numbers: applying the chains twice (two level indexing), overall the random generator had for numbers in 0..127 range loop far above 32k and all delivered numbers were well distributed: as per definition there were no same number streaks longer than 3.
As gamer though, I'd say, too much randomsness is often very distracting. Some console RPGs (e.g. mentioned above FF) are especially affected. If you look at real life, few events around us are really random...
As fan of SRPG and strategical games in general, I prefer games where random events are normally consequences of your past actions. Or even games like Heroes of Might and Magic where there are essentially no randomness in battles (or it is always in known bounds) so that outcome largely depends on player's capabilities.
Access to resources of video card has to be synchronized.
Unless of course you are going to "init 1" just to run you suppa-puppa OpenCL application.
You can't avoid X.org - because they are the driver of video card.
You will stuck into Mesa folks - because they are closer than anybody else to what Kronos groups is doing.
To me, better feature would be to save history for only specified sites and/or to never save anything from specified sites.
There are some sites I visit only to check something periodically - and I do not want them to be in history. Or sites which force you for every little thing to go to new page. I simply do not need them to clog my history.
Now, if only they could do the same for OpenGL.
Care to elaborate?
OpenGL is quite well supported on both Mac OS X and Linux. So with OpenGL you are already free.
Well, it doesn't mean a thing. You know, M$ was on ODF OASIS board quite some time too...
To me litmus test of OpenCL would be independent (from video card vendor) portable implementation which runs on Linux.
Participation doesn't mean a lot to how the spec would develop.
To put it in another words: I'm waiting for reaction from Mesa and X.org folks. Then, if reaction would be positive, the news would get me excited.
Real lawyers on Graklaw use something like "I Am A Lawyer, But Not Your Lawyer."
I do not think that after C# learning Python would bring anything new. I'd say one need to know at least two orthogonal in their concepts languages. That's why I actually brought up C & Perl in another comment.
Or probably my thought would be clearer this way: Python and C# have lots of common concepts and "learning" Python would amount to nothing more than learning new syntax of doing stuff the same way. C or Perl are quite orthogonal and force you to think and approach problems differently.
For language complementing C# I would advise Perl, mainly because Perl is well supported across all platforms, thus your knowledge will not be locked to Linux, but also might benefit Windows side of work.
I second. I'm pretty appalled on how clueless new young programmers on resource management.
N.B. As I was taught 15 years ago, programming is art of resource management. That was my teacher's way of saying the famous "algorithms + data structures == programmes," but pitching that all programs all the time use system resources (CPU, memory, files, etc) and your programs will always be as good as your understanding of the resources.
C, if nothing else, teaches how to manage resources and resource management skills are pretty much what defines a good programmer.
After you have grasped on how all stuff is managed and how it really works under the hood, Perl becomes irreplaceable tool at helping managing them very easily. Perl documentation is excellent on telling you precisely what Perl does and how much of what overhead the given code has.
I bring up resource management because it feels stupid that every time pure Java/C# programmers have problems they come to me: most questions amount to "we do this but whole system explodes" (because e.g. they have tried to allocate and object for 0 ... MAX_INT). And when you tell them that you have to eventually free the memory, they act surprised since they have thought that Java/C# handles that all for them and they expect it to e.g. automagically clear the list which keeps references to all the object which code doesn't intend to use anymore. This is primitive template example, but happened more than once to me.
Understanding resource management is quite tiresome and for smaller projects is overkill. But for anything large enough you have to know it and you have to understand it. I haven't seen a single large Java project written solely by pure Java programmers: very very likely that they have number of ex-C folks on the team. I guess the same goes for C# and Python.
Knowing C and Perl - not as languages - but as tools to work with system resources is invaluable knowledge. Learning another way to express your thoughts gives you nothing (I know probably more than 20 languages), but the very and most important is the knowledge of the system resources and how to work with them properly. Some standard libraries (e.g. Java) had put abstraction level way too high to allow to learn about system resources and to write efficient programs.
1. Install Google Toolbar
2. Google Toolbar > Settings > Options > Buttons > Add Button > Enter in search field "Pirate Bay" > pick one of the buttons to add to your google tool bar.
3. Go to Amazon
4. Select the name of product
5. Right mouse click to call pop up menu > More Search Types > middle mouse click on added in step #2 button.
That's what I use for anime and mininova.
IP's just need to be charged for on a early basis. Start with $1 per year per ip to EVERYONE who owns an IP's and you'll see the "IP Shortage" vanish overnight.
You miss the point.
Internet now has much much more participants than 10 years ago.
IPv4 are running out because compromise was made on management side. Essentially, Internet management is loose because people want to encourage participation. Do not forget that Internet is merely set of connected proprietary networks, operated by all possible entities. And every of the entities wants to have some address space to grow.
I frankly do not see any problem with going to IPv6 as most of the network equipment was ready for several years now.
There's no reason every person on earth needs an IP. Nat+uPNP is perfectly capable and 100% backwords compatible.
Had you ever tried to operate middle to large sized network, you would understand that adding extra layer at every interconnect would make Internet infrastructure at least twice more expensive. I'd say that costs would at least quadruple since such smart equipment, supporting dynamic resource management, costs at least 4 times more if compared to its plain version. (e.g. compare prices on "switch" vs. "managed switch" to see the difference).
Making Internet infrastructure expensive => killing Internet.
P.S. And do not get me started on the "uPNP" bogosity which doesn't even belong to the discussion...
I don't see any of this changing in UNIX any time soon, until it's not actually UNIX any more; the best anyone is going to be able to do is namespace folding hacks, for which they will need shell (and anything else that does globbing) cooperating in the versioning "conspiracy". This is not really something doable at the kernel level without some serious namespace escape trickery.
I do not think that there is a real challenge implementing versioning file system. The problems are well known and their solutions also present in distributed revision control systems.
The problem like always is that "one size never fits all."
Revision control evolves much faster than file systems and putting it into a file system would establish a barrier to amount of customizations and improvements which could be made. Essentially version control and file systems solve two different orthogonal tasks: former is responsible for preserving modification history, later - for reliably storing information.
As I have witnessed many times in business, revision control is very crucial to business process and often the two are tightly coupled (think distribution and demarcation of responsibilities). As result, business must have the flexibility to change revision control according to new market conditions at any moment of time. Putting that stuff into kernel would essentially force to cede control over software development process to some 3rd party. Nobody would ever do it.
Exemplary is my actual employer where process tightly integrates revision control with issue tracking system but also essentially formalizes whole process: from development to testing to customer delivery to maintenance. Everybody has own place in the process and rarely you have the question who is responsible for what. With few gotchas it works pretty well and allows to track what/when/how was done. (I called it already on couple of occasions "todo list on steroids on steroids.")
I agree with that. OS is quite performer and their compiler is state of the art. I'd say they have good scheduler and pthreads implementation is fastest I have ever seen.
Yet. Before, I was swearing at Solaris user tools. Until I haven't started working with HP-UX. Their user tools are even more primitive and counter-intuitive. sed/grep/awk/sort/uniq/friends often fails with fancy meaningless (or "dead end" type of) messages. (e.g. "line is too long").
Spending 1 hour every day writing some scripts to do something what under Linux is available for free or working around some decade old bogosity vs. very fast CPU/OS/compiler - is tough choice. I tend to choose system with better user tools because people do less stupid mistakes there and overall business process then flows smoother.
From that point of view HP-UX is quite poor performer.
P.S. And they were last (even later than Mac OS X!!) to become UNIX'03 certified.
Also, they have used several openly pro-Intel applications: Cinebench and M$ .Net.
Cinebench never hid the fact that they optimize for Intel and if you want to have best performance you need to buy Intel CPUs.
M$ .Net XML benchmark - M$ C/C++ compiler and libraries in many parts use Intel's hand written asm code. And it always produced code optimized for Intel architectures.
Itanium was actually an inferior architecture.
With great sadness, I have to inform you (and remind myself) that IA-64 a.k.a Itanic is still alive.
I'm kind of forced to work on Itanic under HP-UX everyday... :(
You didn't because you have gotten so used to it that you don't see it any more.
I didn't notice because I do not care about that.
I can't check what you say, as I do not have M$Office at hand - only portable OO.o.
The point here is that for the small things you really do not care about all the bogosities of VBA. And I can recall many of them all too well: I had a course on VBA in versity. I can't (nor want to) write in VB, yet I remember many of those bumps one had to jump over to make something working in VBA.
What I'm really pissed about OO.o is that in M$Office I didn't have to be bothered by all the bogosities of VBA but it just worked for trivial things with some amount of copy-pasting. As some expert had showed me before, StarBasic is much much more faster, yet to archive that level of performance simplicity was sacrificed, making StarBasic unacceptable for all those who do not want to invest weeks (or do not have that time) in digging up documentation and examples.
OO.o, to compete against VBA, needs something comparable in simplicity. After all, most of the automation done in word processors is really really trivial: replace that with that; set that attribute on selection or count that and put it in here. This are one/two liner in VBA (what is top of my VB capabilities), yet in StarBasic this is Java-like Object Oriented monstrosity.
the kernel application binary interfaces are a moving target.
That's why we have glibc, which abstracts that ABI from applications.
Kernel driver interface - the horse was already beaten to death many times ( see here ).
a consistent configuration system, to enable distribution;
Windows tried that with Registry - and it didn't worked. And it will never work since "one size never fits all" requirements of all applications.
native file versioning;
Was tried many times before and failed miserably. As long as majority of files are blobs, versioning on level of file system makes no sense. Versioning on level of applications is implemented already more or less everywhere it was needed and SVN/git is there for the rest of applications.
audio APIs;
See ALSA and its user-space libraries.
See SDL.
and the integration of X11 with apps.
As was shown by FreeDesktop initiative not really needed nor X folks want to be bothered by all the end user bells and whistles.
Finally, he argues that Linux needs a committee to insure that all GUIs work consistently and integrate better on the back-end with the kernel.
Committee?? Buahahhahahaha!!!1!!cos(0)!!!!!!!
All what he says was tried before (see (11)) and generally can be described as "failed".
[...] I fail to understand the "several minutes" issue when opening up docs [...]
Dunno what gp meant, but can speak from my experience.
It happens in some configurations when opening documents from network shares. First I thought that it was some server problem, but then OO.o opened the same documents instantly. And M$Office was actually hanging hard for 1.5-2.5 minutes. "Hanging hard" means that Windows wasn't allowing to kill the application leading to worst case of document being locked on server for some time.
Was one of the last drops, which triggered one of my past employers to migrate completely to OO.o internally, keeping M$Office for communication with partners only.