/*
* Structure of a directory entry
*/
#define EXT2_NAME_LEN 255
Copright Remy Card, Linus Torvalds
Now that all the MS engineers are gone, I'll continue...:-)
Hmm...it would depend. We are forbidden to look at any GPL'ed code. BSD code is a grey area, and we need to consult with legal. If the indic language support is truly in the public domain, then we can certainly look at it.
Don't you feel that this is a ridiculous rule? Ok - if you read GPL'd code, you are now in the position that anything you write of similar functionality is "at risk" of being contaminated by this knowledge. I have to be careful not to view, say, any code in Postgresql as it might affect any coding decisions I might make in the future (I can feel someone about to post some Postgresql code in reply to this...). That said, almost every piece of code contains the same ideas - maths, caching strategies, data transport. Coming from a scientific background, being able to build on the vast store of knowledge of those who have gone before me is a natural process. Having to walk the boundaries of copyright law, patents and other legal straight-jackets is a confining and ultimately unproductive method.
Just where would we be today if we could treat source code in the same way we treat mathematics?
"It is, however, nice to know that IBM's fire-breathing legion of IP lawyers is on the side of the GPL."
No, they are not on the side of the GPL; they are on IBM's side. It just so happens that, currently, that IBM and the GPL have a common interest.
But consider this. Having placed themselves on the side of the GPL and by litigating against another company using the GPL license, it makes it nearly impossible for IBM (or any other company) to ever turn around and disregard the GPL. This will set legal precedent in stone. If IBM didn't want that precedent, it would not be pursuing this line of action.
With X, there is no easy way (xmove was impractical) to leave an application running, and move it between desktops.
In what way was Xmove impractical? I run it on the remote server and have xmovectl clients to jump me around when I need to drag X applications around with me as I move IP address. It works fine.
On the other hand, if anyone can point me at a way to secure Xmove so that I am the only user who can muck with my apps, that would be good... I'm not convinced that Xmove was built with security in mind.
The average user learns by doing, just like you did, you arrogant twit.
I'm sick of hearing how regular folk are just too dumb to use a computer like an elite genius such as yourself.
Must... not... feed... the... trolls....
***failed***
Look, my opinions are the result of working with normal people trying to use their machines. They don't result from some ivory tower view of the world. Computers are HARD TO USE. Period. Most people find their computers baffling at least some of the time, and downright frustrating and impossible when things stop working.
And yes, the average user does learn by doing. But doing what? Most people get better and more experienced at using their systems by being able to ask someone for help when the proverbial hits the fan. If the family in question was not previously technical and they are given a computer as part of their high tech make-over, they will need help a lot in the first month, fairly regularly for the next 6 months and after that most things will be okay and fixable by themselves. If you are trying to introduce the family to computer games (which was the point of this thread if you hadn't noticed) why burden them with a bunch of system admin? They don't need the hassle.
[AverageGuy] Awesome, thanks! So what games are on here?
xBill & Nibbles. 'Nuff said:-)
Quite frankly, if I was kitting out a family with a games machine, it would be a PS2 or similar console system. The idea of any family not familiar with keeping a PC up and running would use it for gaming is pretty funny. Wait until they have to install their first Windows SP, update Direct X or their box gets overrun by Sven, Welchia and Blaster. It would actually be easier to install UT2k3, Quake3, Savage, Orbz, Thinktanks and half a dozen other native Linux games and LOCK DOWN THE BOX. The average user (note: not the average PC gamer) has little enough clue about these issues in the first place. As has been postulated elsewhere, maintenance is the tricky part. At least with a console, the effective locking down has already been done and the average person can find games in future shop or wherever.
Also...what do people do when they have to browse an IE only web site? I know...web sites SHOULD be standards compliant, but fact is they're not.
They should contact Sun to find out why that web page doesn't work in their environment. After all, this is a supported product. If there is a business need that can justify this complaint, opening a ticket for this problem will start the chain of events needed to fix this problem.
Sun then has the same problem that the Mozilla Evangelism team has to deal with - convincing the site in question that they can still make IE users happy while allowing everyone access to the information. Now that Sun is reliant on the Mozilla project for its browser, Sun now has a vested interest in ensuring that all companies who are likely to be supplying web pages accessed by the Java DS users are abiding by the web standards. In the long run, this should lead to more and more sites realising that obeying the standards is actually in their interest.
To be honest, I've never understood why ActiveX controls in a browser was necessary. I suspect it was done primarly to support Windows Update and help tie IE down as a critical part of the Windows desktop that can't be removed. Having said that, my dual boot RH8.0/ Win2k laptop generally only visits microsoft.com to download the latest service pack and critical fixes. That I download those fixes in Linux and then apply them to the Win2k side when my laptop is disconnected from the network just goes to prove that even that tying of OS function to web activities isn't needed.
Beside the link being wrong (it points to the top requested servers, not the top uptime, select that statistic in the menu on the left hand)...
No - the choice of link was quite deliberate. I wanted a cross section of most requested sites. Top uptimes come from systems not seeing a ton of changes in hardware so they are a little less interesting. For example, Slashdot could not have handled the traffic of this site on 5 year old hardware like the top BSD sites. And Windows 2000 would have no chance to appear on a list of 5 year uptimes through no fault of MS (at least in this case...).
And I totally agree with your point about Linux admins being more likely to recompile the kernel to pick up the latest kernel tweaks. One of the reasons I enjoy mucking around on these Linux boxes that now litter my existence is the fast changing improvements and refinements that continually arrive. The idea of tying a box down for 5 years is a bit of an anathema for me...
With perfect honesty, as someone who has used and programmed various OSes and hardware, the differences between Linux and Windows are few. Both are hugely complex from the user's point of view, and both are arguably incomprehensible on the source level. (Remember, you can't just talk about the Linux kernel, but the entire package including XFree86, drivers, the window manager, KDE, etc.) So it's not like either one is a clear winner in terms of ease-of-use or architectural cleanliness. It used to be that Linux was more stable, but with Windows 2000 that's no longer true.
Well, my system uptimes tell a different story. But you don't have to trust my figures - take a look at Netcraft uptimes. These aren't the highest uptimes of all - almost without exception, the best uptimes are held by BSD derivatives (around 5 years). For the most popular sites, Linux tends to knock in around 100+ days. Windows 2000 knocks around in the high-20's, low 30's.
To get the most ease of use out of your computer, it really needs to interact with you on your level. This means a user interface that is based on faulty human communication (verbal, facial expressions, gestures, posture, etc...) We're not there yet. However, this is what the Linux crowd should focus on instead of trying to imitate what's already been done.
It's a great idea. And maybe for the general public it would be superb. But for someone like me (techincal, detail orientated, good at scripting) it's a nightmare.
Hell why do I like working with computers? For me, working on a computer system is akin to the 'Infinite Fun' envisaged by Iain M. Banks in the Culture novels. Its the possibility of a nearly endless variety of pursuits and combinations sitting underneath my figuretips.
But put a fuzzy 'did you want to raise that window over there that you glanced at' interface over it and it's no longer completely under my control. Some finite state machine with fuzzy logic is going to try and interpret what were unambiguous statements and are now some analogue input or gesture? Maybe there are niceties out there to be done for this but I happen to like the current state of the UI. For the most part, it's simple, I can control all the aspects I need to and leave the rest on automatic (but deterministic) control.
So please start up a company to build this interface for the masses. I'm sure it will do well. But I don't want any part of it.
There is a GPL'd version of QT available for Linux. So if you are a commercial developer you have two choices:
Use the GPL'd QT libraries and GPL your own software too
Pay Trolltech for a commercial license to use QT and keep your software closed source.
This is not a Windows/Linux split. It is a GPL/commercial license split. I seem to remember an effort to port the GPL'd QT libraries to Win32 starting up somewhere... no idea whether it got anywhere.
[bigjocker@anacreon tmp]$ gcc test.c -Wall -o test
test.c: In function `main':
test.c:5: warning: control reaches end of non-void function
[bigjocker@anacreon tmp]$
Good thing it didn't compile. If it had, we'd discover whether you have . in your PATH. If you do (why?) you might find that a lot of scripts now throw a wobbly and do interesting things to your system.
Don't EVER create a program called 'test' in your PATH.
Never attribute to malice what you can explain by stupidity.
That has to be my number one mantra...
Indian search on Windows. Only 10th result is good. First one is Windows Media Player, 2-6 are about Tablet PC with the same page for WinXP for Tablets appearing twice. There are also such wonders of relevance as Windows into the Soul of Satyajit Ray and Windows to the Universe - Jupiter.
However, you're not comparing Apples with Oranges. Or Apples with Windows. Ahem.
The term 'Linux' is variously used to describe:
An open source kernel
A complete operating system
There are no other regular uses of this term.
The term 'Windows' is applied to:
An operating system sold by MS.
An item comprising a allegory used for interacting with a computer (as in Windows, Icons, Menus and Pointers)
That thing which lets light in through the hole in the wall
An interval of data (such as in a sliding windowed average)
Something through which an item can be seen or pictured.
...
some other stuff I've missed by not bothering to dig out my dictionary.
Searching just for Windows is likely to pull up a wide range of possible answers. Searching for 'Microsoft Windows' might be expected to pull up the same sorts of results as 'Linux' on its own. Searching for 'Linux operating system' or 'Windows operating system' would maybe be fairer.
So while I think you are onto something by pointing out that MS search technology is 'crap', a blind search for Linux should stand a decent chance of getting something about some OS code. A blind search for 'Apple' or 'Windows' is much less likely to pull up something computer related.
As I read this, SCO has terminated their right to distribute their Linux-based OS (and anything using Samba...) by attempting to sublicense to others under a non-GPL license
See that little bit in Section 4 that says 'the Program'? That means that each GPL license applies to the code it is with, rather than a blanket provision on all GPL software. So SCO has almost certainly violated the Linux kernel GPL license by attempting to sublicense the kernel but that does not remove their rights to distribute Samba 3 under the GPL.
Still does nothing to remove the hypocrisy of their situation, but quite frankly if the court of public opinion counts for anything, SCO is going to be feeling the draft of departing VARs and customers for a long time. Last one out switch out the lights. Although "better take off and nuke the site from orbit. It's the only way to be sure" has a nice ring to it.:-)
If you want to do research, a Ph.D is a must. If not, it is a waste.
I wouldn't go that far. A PhD builds a new skill set that a first degree does not cover:
improves independence - you have to be able to go from idea to solution in almost any technical PhD.
strong technical writing skills - if you write a lousy thesis, your chances of getting your PhD are virtually nil.
strong technical presentation skills - most PhDs have to present their material in person to technical audiences, some of whom will be strongly critical of your work.
Having worked for four years to get my PhD (in observational astronomy as it happens) and then left academia to go and work in software development, the tricks and tips I learnt during that time have proved extremely useful and I have no regrets about spending that time.
Basicly, the code they've showed goes as far back as 1992 from BSD 2.11, perhaps even further?
Looking through the UnixTree sources on Minnie, it goes back to 1973 from the Fifth edition Unix sources. It's in the malloc code.
So for SCO to claim this is interesting. Especially as it is out in the open for everyone to look at. And thanks to Bell labs - their copyright notice is dated 1973.
Having works in a games company, I can at least tell you what may opinions of why that wouldn't work are. Most games do start with a nice seperated interface, but the it gets destroyed for 3 main reasons.
I work on DB2 UDB (which runs on more platforms than I care to remember). While I don't have to worry overly much about graphics or sound libraries, everything else requires similar care to make sure that I write once and it must work everywhere.
1) Performance - you want to squeeze every last bit out of the latest shiny card that has come out and you suspect the reviewers will be using.
Performance is pretty key for any program out there. If you are using function pointers to glue your external libraries onto the base code and you are calling those function pointers a lot, you are probably hosed using this approach. Of course, this isn't the only solution. You can have each seperate library use the same function call names and make the dynamic library linker pick up the one you actually want for your current config. Or you can get the C/C++ preprocessor rewrite your code for you using each distinct target. Given that the code on a CD/DVD is a small fraction of the actual product, this approach is not a problem. Any extra performance you might be able to wring out of any one interface by hard coding is probably only a few percent and not worth the worry.
2) Driver problems - similarily, you have to work around the problems that there are in most directX drivers (and don't suggest doing openGL only, because the openGL drivers are often worse!)
Industry support for OpenGL isn't as good as it should be - NVidia does a decent enough job, ATI sucks some of the time... Others are various. Having never worked with Direct X, I can't comment. But only one console implements Direct X. I have no idea how many of the consoles (if any) have OpenGL libraries (PS2/Xbox running linux excepted...).
3) Deadlines - in that last month, it's a case of slap in anything to make it work.. and if it's a choice between slipping a couple of weeks or dumping linux (which it almost certainly will be), the dumping linux it will be
Deadlines are THE KILLER when it comes to product quality. They basically force things to be patched to some 'reasonable' level and the product is booted out the door. If the user is lucky, patches and fixes will follow.
4) connected to 3, with windows "all" you have to do is check windows 95,98,ME,2000,XP (which is still a lot!) With linux there isn't really a limit to the number of versions you could have to test on...
I don't really buy the 'There are less versions of Windows than Linux' argument. In either case, it boils down to kernel versions, central system library versions and support library versions. Fundamentally, Windows is no different to linux in this regard. It would be nice to see some framework emerge where you could download a system analysis tool (maybe like autoconf) onto a clients computer, test out the key libraries to see that they behave in the expected manner, and then either allow the customer to download the correct product or give the customer the 'this won't work on your system' message. Be nice to see this for both Windows and Linux - I've had a couple of Windows games that never worked on Win 98 despite the protestations on the side of the box that this was a supported platform. DLL hell exists anywhere on the planet. At least with Linux, you can quickly tell what the dynamic linker is going to throw in your direction.
So, I'm not totally disagreeing with you. But I'm not convinced by your points either. At the end of the day, managers often seem to decide that the short runway to the deadlines in the gaming industry means one product on one platform That id software manages to release a Linux version of a well tested, carefully crafted engine and make a lot of Linux gamers happy is something I am profoundly grateful for. I also hope that one day it actually pays the folks at Id some cash back for their efforts and support.
Linux is here to stay and it's presence on the desktop will only increase from here. As people get used to the idea that Windows is not the only option, Mac OS may also pick up new converts.
You are an optimist. People buy home machines because they look like what they are used to at work/school/whatever. Hence, most purchases are Windows boxes.
I'm actually a realist. And a realist with access to a large amount of business information through various sales channels. For a large segment of the population, what you say is currently absolutely correct, at least in the US or Canada. In Germany however, the user is more and more likely to be working on a Linux platform AT WORK. Across Europe and Asia, the chances of working on a Linux desktop are going up pretty much day-on-day. That will translate pretty directly to more linux boxes at home. There is also another important side effect of this increase in Linux desktops in a work environment - people become aware that Microsoft is not the only provider of operating systems. Most Joe/Jane User people out there have little or no idea that there are any alternatives. A lot of people would not consider buying Apple systems because thats 'not a PC'.
Breaking that requires breaking into schools and businesses with Linux on the desktop, and that isn't happening anywhere significant in the US. Cited issue is 'the users are used to what we have, why change'. You might have a better shot in Europe. Even the specter of huge licensing costs doesn't seem to have made much of a dent here.
Really, it doesn't matter to me where it starts. For many reasons, the US looks like it will be playing catch-up further down the line as it's margins get undercut by other IT players not funnelling a large chunk of their revenue via a certain company in Redmond. There will still be IT costs - those won't magically disappear just because Linux is widely deployed, but they will be a smaller percentage of the financial capital being laid out for expenses. And the base cost of the OS really is mostly irrelevant to companies - the BIG dollars come from paying for staff or support companies to keep tabs on the servers. Downtime, crashes and unfixable behaviour are BIG factors in working out the costs of an operating system and Linux is generally more stable, has better uptimes, has a proven track record in business , is easier to maintain and manage remotely and is a consistent environment for any UNIX admin to come in and work with. RedHat charges considerably less for it's Advanced Server support than comparable Windows-supporting companies, and for companies with decent UNIX admins, Linux is a comfortable fit into the system without external support.
But to come back to your point about education. Schools probably come top of the table for people who are strapped for cash. Projects like K12LTSP are a godsend for schools if they are prepared to bite the bullet. Many students entering university have Linux skills (from surveys I've seen, those figures are increasing year-on-year).
Linux gaming will be about the last area for Linux users to see even 10% of the games market support Linux from release. But it will happen, and probably within the next 5 years at most.
Why is it so hard to port games from Windows to Linux?? Is it just the connections to interfaces (graphics cards, keyboard, joystick mouse) or does it have more to do with compilers and incompatibilities??
Partly. But if that were all of it, it still wouldn't take so long. The real reason is that most software companies writing Windows software never consider any other platform at all during the design stages and so the code is written in the idiom of the target platform.
If you want to write cross-platform (or even just plain good simple) code, you probably want to:
use definite sized types such as unsigned 32bit integer rather than native int.
segregate all platform distinct functions in separate layers
have a general graphics interface through which all graphics are done - all graphics interface (direct 3D, open GL, funky personal 3D lib as found in various consoles) work is done on in a library implementing this general interface.
have a general audio interface so that you can plug A3D, EAX, OpenAL or whatever in on the far side according to the eventual target.
have all input devices run through a general interface so that you can plug Directplay, libSDL or whatever into the far side as needed.
choose a specific endianess for all saved files and convert if necessary when saving/loading (this allows you to switch files between Windows and Max, or Linux Intel to Linux Alpha, etc.
Given all that, and certain other caveats about threading, forking and other process model distinctions, you'll find that you can have multiple platforms up and running fairly easily. Your code won't be making specific assumptions about its platform architecture and most nasty bugs will either be shallow for one of the platform/lib combinations or limited to one of the platform specific libraries (such as the Direct 3D driver, etc) where you can hack/fix/workaround it in that library alone.
Of course, robust code isn't necessarily always a priority for Windows games and at the moment, a Windows-only company is only cutting off 7% or so of it's possible market by being Windows only. That will change given time. Linux is here to stay and it's presence on the desktop will only increase from here. As people get used to the idea that Windows is not the only option, Mac OS may also pick up new converts.
With sufficient patience you can rip Metacity out and run the most recent version of Sawfish. I'm doing that on RH 9, albeit with GARNOME rather than the RH 9 GNOME stuff.
Sufficient patience? You install Sawfish, librep, rep-gtk and any other missing dependencies and then the final blow:
I don't know what flavor of Windows they run in the office you work at, but I'm yet to see a WinNT/2k/XP crash at my office, with over 500 systems.
Application crashes, yes. Operating system failures, no.
Err. You must have a good supply of magic pixie dust, because I work with considerably less than 500 NT/2000 machines (no XP yet) and I see system crashes day in, day out. From laptops bluescreening on wake-up from sleep mode, to explorer.exe hangs. My NT boxen had a nasty habit of surviving 7-14 days of moderate work before slowing down to molasses and poping a few VxDs.
On the other hand, my Linux, AIX, Solaris and HP-UX boxes all run full time between scheduled power downs pretty much all the time. Admittedly, the AIX, Solaris and HP-UX boxes generally run headless, but the desktop linux boxes around are rock-solid stable. I never saw uptimes above 100 days for any NT/2000 box, whereas that is minimum uptime for the Unix/Linux crowd.
So be glad of your supply of magic pixie dust for Windows.
Slashdot Mirror
/*
* Structure of a directory entry
*/
#define EXT2_NAME_LEN 255
Copright Remy Card, Linus Torvalds
Now that all the MS engineers are gone, I'll continue... :-)
Hmm...it would depend. We are forbidden to look at any GPL'ed code. BSD code is a grey area, and we need to consult with legal. If the indic language support is truly in the public domain, then we can certainly look at it.
Don't you feel that this is a ridiculous rule? Ok - if you read GPL'd code, you are now in the position that anything you write of similar functionality is "at risk" of being contaminated by this knowledge. I have to be careful not to view, say, any code in Postgresql as it might affect any coding decisions I might make in the future (I can feel someone about to post some Postgresql code in reply to this ...). That said, almost every piece of code contains the same ideas - maths, caching strategies, data transport. Coming from a scientific background, being able to build on the vast store of knowledge of those who have gone before me is a natural process. Having to walk the boundaries of copyright law, patents and other legal straight-jackets is a confining and ultimately unproductive method.
Just where would we be today if we could treat source code in the same way we treat mathematics?
Cheers,
Toby Haynes
"It is, however, nice to know that IBM's fire-breathing legion of IP lawyers is on the side of the GPL."
No, they are not on the side of the GPL; they are on IBM's side. It just so happens that, currently, that IBM and the GPL have a common interest.
But consider this. Having placed themselves on the side of the GPL and by litigating against another company using the GPL license, it makes it nearly impossible for IBM (or any other company) to ever turn around and disregard the GPL. This will set legal precedent in stone. If IBM didn't want that precedent, it would not be pursuing this line of action.
Cheers,
Toby Haynes
With X, there is no easy way (xmove was impractical) to leave an application running, and move it between desktops.
In what way was Xmove impractical? I run it on the remote server and have xmovectl clients to jump me around when I need to drag X applications around with me as I move IP address. It works fine.
On the other hand, if anyone can point me at a way to secure Xmove so that I am the only user who can muck with my apps, that would be good... I'm not convinced that Xmove was built with security in mind.
Cheers,
Toby Haynes
don't understand why i must drink what's in my toothpaste.
I don't understand the fundamental basis for gravity but I'm glad I don't float off into space while I try and work it out :-)
Cheers,
Toby Haynes
The average user learns by doing, just like you did, you arrogant twit.
I'm sick of hearing how regular folk are just too dumb to use a computer like an elite genius such as yourself.
Must ... not ... feed ... the ... trolls ....
***failed***
Look, my opinions are the result of working with normal people trying to use their machines. They don't result from some ivory tower view of the world. Computers are HARD TO USE. Period. Most people find their computers baffling at least some of the time, and downright frustrating and impossible when things stop working.
And yes, the average user does learn by doing. But doing what? Most people get better and more experienced at using their systems by being able to ask someone for help when the proverbial hits the fan. If the family in question was not previously technical and they are given a computer as part of their high tech make-over, they will need help a lot in the first month, fairly regularly for the next 6 months and after that most things will be okay and fixable by themselves. If you are trying to introduce the family to computer games (which was the point of this thread if you hadn't noticed) why burden them with a bunch of system admin? They don't need the hassle.
Cheers,
Toby Haynes
[AverageGuy] Awesome, thanks! So what games are on here?
xBill & Nibbles. 'Nuff said :-)
Quite frankly, if I was kitting out a family with a games machine, it would be a PS2 or similar console system. The idea of any family not familiar with keeping a PC up and running would use it for gaming is pretty funny. Wait until they have to install their first Windows SP, update Direct X or their box gets overrun by Sven, Welchia and Blaster. It would actually be easier to install UT2k3, Quake3, Savage, Orbz, Thinktanks and half a dozen other native Linux games and LOCK DOWN THE BOX. The average user (note: not the average PC gamer) has little enough clue about these issues in the first place. As has been postulated elsewhere, maintenance is the tricky part. At least with a console, the effective locking down has already been done and the average person can find games in future shop or wherever.
Cheers,
Toby Haynes
Also...what do people do when they have to browse an IE only web site? I know...web sites SHOULD be standards compliant, but fact is they're not.
They should contact Sun to find out why that web page doesn't work in their environment. After all, this is a supported product. If there is a business need that can justify this complaint, opening a ticket for this problem will start the chain of events needed to fix this problem.
Sun then has the same problem that the Mozilla Evangelism team has to deal with - convincing the site in question that they can still make IE users happy while allowing everyone access to the information. Now that Sun is reliant on the Mozilla project for its browser, Sun now has a vested interest in ensuring that all companies who are likely to be supplying web pages accessed by the Java DS users are abiding by the web standards. In the long run, this should lead to more and more sites realising that obeying the standards is actually in their interest.
To be honest, I've never understood why ActiveX controls in a browser was necessary. I suspect it was done primarly to support Windows Update and help tie IE down as a critical part of the Windows desktop that can't be removed. Having said that, my dual boot RH8.0/ Win2k laptop generally only visits microsoft.com to download the latest service pack and critical fixes. That I download those fixes in Linux and then apply them to the Win2k side when my laptop is disconnected from the network just goes to prove that even that tying of OS function to web activities isn't needed.
Cheers,
Toby Haynes
The directors of the firm hired to continue the posting after the other people had been sacked, wish it to be known that they have just been sacked.
This posting has been completed in an entirely different style at great expense and at the last minute.
Yours sincerely,
The Great Llama
We apologise again for the fault in the posting. Those responsible for sacking the people who have just been sacked have been sacked.
Damn this 2 moose Slashdot timeout...
We apologise for the fault in the posting. Those responsible have been sacked.
Beside the link being wrong (it points to the top requested servers, not the top uptime, select that statistic in the menu on the left hand)...
No - the choice of link was quite deliberate. I wanted a cross section of most requested sites. Top uptimes come from systems not seeing a ton of changes in hardware so they are a little less interesting. For example, Slashdot could not have handled the traffic of this site on 5 year old hardware like the top BSD sites. And Windows 2000 would have no chance to appear on a list of 5 year uptimes through no fault of MS (at least in this case...).
And I totally agree with your point about Linux admins being more likely to recompile the kernel to pick up the latest kernel tweaks. One of the reasons I enjoy mucking around on these Linux boxes that now litter my existence is the fast changing improvements and refinements that continually arrive. The idea of tying a box down for 5 years is a bit of an anathema for me...
Cheers,
Toby Haynes
With perfect honesty, as someone who has used and programmed various OSes and hardware, the differences between Linux and Windows are few. Both are hugely complex from the user's point of view, and both are arguably incomprehensible on the source level. (Remember, you can't just talk about the Linux kernel, but the entire package including XFree86, drivers, the window manager, KDE, etc.) So it's not like either one is a clear winner in terms of ease-of-use or architectural cleanliness. It used to be that Linux was more stable, but with Windows 2000 that's no longer true.
Well, my system uptimes tell a different story. But you don't have to trust my figures - take a look at Netcraft uptimes. These aren't the highest uptimes of all - almost without exception, the best uptimes are held by BSD derivatives (around 5 years). For the most popular sites, Linux tends to knock in around 100+ days. Windows 2000 knocks around in the high-20's, low 30's.
Have a nice day.
Toby Haynes
Conversely:
To get the most ease of use out of your computer, it really needs to interact with you on your level. This means a user interface that is based on faulty human communication (verbal, facial expressions, gestures, posture, etc...) We're not there yet. However, this is what the Linux crowd should focus on instead of trying to imitate what's already been done.
It's a great idea. And maybe for the general public it would be superb. But for someone like me (techincal, detail orientated, good at scripting) it's a nightmare.
Hell why do I like working with computers? For me, working on a computer system is akin to the 'Infinite Fun' envisaged by Iain M. Banks in the Culture novels. Its the possibility of a nearly endless variety of pursuits and combinations sitting underneath my figuretips.
But put a fuzzy 'did you want to raise that window over there that you glanced at' interface over it and it's no longer completely under my control. Some finite state machine with fuzzy logic is going to try and interpret what were unambiguous statements and are now some analogue input or gesture? Maybe there are niceties out there to be done for this but I happen to like the current state of the UI. For the most part, it's simple, I can control all the aspects I need to and leave the rest on automatic (but deterministic) control.
So please start up a company to build this interface for the masses. I'm sure it will do well. But I don't want any part of it.
Cheers,
Toby Haynes
This is not a Windows/Linux split. It is a GPL/commercial license split. I seem to remember an effort to port the GPL'd QT libraries to Win32 starting up somewhere ... no idea whether it got anywhere.
Cheers,
[bigjocker@anacreon tmp]$ gcc test.c -Wall -o test test.c: In function `main': test.c:5: warning: control reaches end of non-void function [bigjocker@anacreon tmp]$
Good thing it didn't compile. If it had, we'd discover whether you have . in your PATH. If you do (why?) you might find that a lot of scripts now throw a wobbly and do interesting things to your system.
Don't EVER create a program called 'test' in your PATH.
Cheers,
Toby Haynes
Never attribute to malice what you can explain by stupidity.
That has to be my number one mantra ...
Indian search on Windows. Only 10th result is good. First one is Windows Media Player, 2-6 are about Tablet PC with the same page for WinXP for Tablets appearing twice. There are also such wonders of relevance as Windows into the Soul of Satyajit Ray and Windows to the Universe - Jupiter.
However, you're not comparing Apples with Oranges. Or Apples with Windows. Ahem.
The term 'Linux' is variously used to describe:
- An open source kernel
- A complete operating system
There are no other regular uses of this term.The term 'Windows' is applied to:
Searching just for Windows is likely to pull up a wide range of possible answers. Searching for 'Microsoft Windows' might be expected to pull up the same sorts of results as 'Linux' on its own. Searching for 'Linux operating system' or 'Windows operating system' would maybe be fairer.
So while I think you are onto something by pointing out that MS search technology is 'crap', a blind search for Linux should stand a decent chance of getting something about some OS code. A blind search for 'Apple' or 'Windows' is much less likely to pull up something computer related.
Cheers,
Toby Haynes
As I read this, SCO has terminated their right to distribute their Linux-based OS (and anything using Samba...) by attempting to sublicense to others under a non-GPL license
See that little bit in Section 4 that says 'the Program'? That means that each GPL license applies to the code it is with, rather than a blanket provision on all GPL software. So SCO has almost certainly violated the Linux kernel GPL license by attempting to sublicense the kernel but that does not remove their rights to distribute Samba 3 under the GPL.
Still does nothing to remove the hypocrisy of their situation, but quite frankly if the court of public opinion counts for anything, SCO is going to be feeling the draft of departing VARs and customers for a long time. Last one out switch out the lights. Although "better take off and nuke the site from orbit. It's the only way to be sure" has a nice ring to it. :-)
Cheers,
Toby Haynes
If you want to do research, a Ph.D is a must. If not, it is a waste.
I wouldn't go that far. A PhD builds a new skill set that a first degree does not cover:
Having worked for four years to get my PhD (in observational astronomy as it happens) and then left academia to go and work in software development, the tricks and tips I learnt during that time have proved extremely useful and I have no regrets about spending that time.
Cheers,
Toby Haynes
Basicly, the code they've showed goes as far back as 1992 from BSD 2.11, perhaps even further?
Looking through the UnixTree sources on Minnie, it goes back to 1973 from the Fifth edition Unix sources. It's in the malloc code.
So for SCO to claim this is interesting. Especially as it is out in the open for everyone to look at. And thanks to Bell labs - their copyright notice is dated 1973.
Cheers,
Toby Haynes
Having works in a games company, I can at least tell you what may opinions of why that wouldn't work are. Most games do start with a nice seperated interface, but the it gets destroyed for 3 main reasons.
I work on DB2 UDB (which runs on more platforms than I care to remember). While I don't have to worry overly much about graphics or sound libraries, everything else requires similar care to make sure that I write once and it must work everywhere.
1) Performance - you want to squeeze every last bit out of the latest shiny card that has come out and you suspect the reviewers will be using.
Performance is pretty key for any program out there. If you are using function pointers to glue your external libraries onto the base code and you are calling those function pointers a lot, you are probably hosed using this approach. Of course, this isn't the only solution. You can have each seperate library use the same function call names and make the dynamic library linker pick up the one you actually want for your current config. Or you can get the C/C++ preprocessor rewrite your code for you using each distinct target. Given that the code on a CD/DVD is a small fraction of the actual product, this approach is not a problem. Any extra performance you might be able to wring out of any one interface by hard coding is probably only a few percent and not worth the worry.
2) Driver problems - similarily, you have to work around the problems that there are in most directX drivers (and don't suggest doing openGL only, because the openGL drivers are often worse!)
Industry support for OpenGL isn't as good as it should be - NVidia does a decent enough job, ATI sucks some of the time... Others are various. Having never worked with Direct X, I can't comment. But only one console implements Direct X. I have no idea how many of the consoles (if any) have OpenGL libraries (PS2/Xbox running linux excepted...).
3) Deadlines - in that last month, it's a case of slap in anything to make it work.. and if it's a choice between slipping a couple of weeks or dumping linux (which it almost certainly will be), the dumping linux it will be
Deadlines are THE KILLER when it comes to product quality. They basically force things to be patched to some 'reasonable' level and the product is booted out the door. If the user is lucky, patches and fixes will follow.
4) connected to 3, with windows "all" you have to do is check windows 95,98,ME,2000,XP (which is still a lot!) With linux there isn't really a limit to the number of versions you could have to test on...
I don't really buy the 'There are less versions of Windows than Linux' argument. In either case, it boils down to kernel versions, central system library versions and support library versions. Fundamentally, Windows is no different to linux in this regard. It would be nice to see some framework emerge where you could download a system analysis tool (maybe like autoconf) onto a clients computer, test out the key libraries to see that they behave in the expected manner, and then either allow the customer to download the correct product or give the customer the 'this won't work on your system' message. Be nice to see this for both Windows and Linux - I've had a couple of Windows games that never worked on Win 98 despite the protestations on the side of the box that this was a supported platform. DLL hell exists anywhere on the planet. At least with Linux, you can quickly tell what the dynamic linker is going to throw in your direction.
So, I'm not totally disagreeing with you. But I'm not convinced by your points either. At the end of the day, managers often seem to decide that the short runway to the deadlines in the gaming industry means one product on one platform That id software manages to release a Linux version of a well tested, carefully crafted engine and make a lot of Linux gamers happy is something I am profoundly grateful for. I also hope that one day it actually pays the folks at Id some cash back for their efforts and support.
Cheers,
Toby Haynes
Linux is here to stay and it's presence on the desktop will only increase from here. As people get used to the idea that Windows is not the only option, Mac OS may also pick up new converts.
You are an optimist. People buy home machines because they look like what they are used to at work/school/whatever. Hence, most purchases are Windows boxes.
I'm actually a realist. And a realist with access to a large amount of business information through various sales channels. For a large segment of the population, what you say is currently absolutely correct, at least in the US or Canada. In Germany however, the user is more and more likely to be working on a Linux platform AT WORK. Across Europe and Asia, the chances of working on a Linux desktop are going up pretty much day-on-day. That will translate pretty directly to more linux boxes at home. There is also another important side effect of this increase in Linux desktops in a work environment - people become aware that Microsoft is not the only provider of operating systems. Most Joe/Jane User people out there have little or no idea that there are any alternatives. A lot of people would not consider buying Apple systems because thats 'not a PC'.
Breaking that requires breaking into schools and businesses with Linux on the desktop, and that isn't happening anywhere significant in the US. Cited issue is 'the users are used to what we have, why change'. You might have a better shot in Europe. Even the specter of huge licensing costs doesn't seem to have made much of a dent here.
Really, it doesn't matter to me where it starts. For many reasons, the US looks like it will be playing catch-up further down the line as it's margins get undercut by other IT players not funnelling a large chunk of their revenue via a certain company in Redmond. There will still be IT costs - those won't magically disappear just because Linux is widely deployed, but they will be a smaller percentage of the financial capital being laid out for expenses. And the base cost of the OS really is mostly irrelevant to companies - the BIG dollars come from paying for staff or support companies to keep tabs on the servers. Downtime, crashes and unfixable behaviour are BIG factors in working out the costs of an operating system and Linux is generally more stable, has better uptimes, has a proven track record in business , is easier to maintain and manage remotely and is a consistent environment for any UNIX admin to come in and work with. RedHat charges considerably less for it's Advanced Server support than comparable Windows-supporting companies, and for companies with decent UNIX admins, Linux is a comfortable fit into the system without external support.
But to come back to your point about education. Schools probably come top of the table for people who are strapped for cash. Projects like K12LTSP are a godsend for schools if they are prepared to bite the bullet. Many students entering university have Linux skills (from surveys I've seen, those figures are increasing year-on-year).
Linux gaming will be about the last area for Linux users to see even 10% of the games market support Linux from release. But it will happen, and probably within the next 5 years at most.
Cheers,
Toby Haynes
Why is it so hard to port games from Windows to Linux?? Is it just the connections to interfaces (graphics cards, keyboard, joystick mouse) or does it have more to do with compilers and incompatibilities??
Partly. But if that were all of it, it still wouldn't take so long. The real reason is that most software companies writing Windows software never consider any other platform at all during the design stages and so the code is written in the idiom of the target platform.
If you want to write cross-platform (or even just plain good simple) code, you probably want to:
Given all that, and certain other caveats about threading, forking and other process model distinctions, you'll find that you can have multiple platforms up and running fairly easily. Your code won't be making specific assumptions about its platform architecture and most nasty bugs will either be shallow for one of the platform/lib combinations or limited to one of the platform specific libraries (such as the Direct 3D driver, etc) where you can hack/fix/workaround it in that library alone.
Of course, robust code isn't necessarily always a priority for Windows games and at the moment, a Windows-only company is only cutting off 7% or so of it's possible market by being Windows only. That will change given time. Linux is here to stay and it's presence on the desktop will only increase from here. As people get used to the idea that Windows is not the only option, Mac OS may also pick up new converts.
Cheers,
Toby Haynes
With sufficient patience you can rip Metacity out and run the most recent version of Sawfish. I'm doing that on RH 9, albeit with GARNOME rather than the RH 9 GNOME stuff.
Sufficient patience? You install Sawfish, librep, rep-gtk and any other missing dependencies and then the final blow:
killall metacity; sleep 5; sawfish; gnome-session-save
All done.
Okay, so there is no nice clicky GUI thing. But hard it ain't.
Cheers,
Toby Haynes
For amusement though, you could always add driftnet into the mix and see what other people in your subnet are looking at.
And no, I have no idea whether this was actually the embarrassing screensaver. Just a word to the wise is all...
Cheers,
Toby Haynes
I don't know what flavor of Windows they run in the office you work at, but I'm yet to see a WinNT/2k/XP crash at my office, with over 500 systems.
Application crashes, yes. Operating system failures, no.
Err. You must have a good supply of magic pixie dust, because I work with considerably less than 500 NT/2000 machines (no XP yet) and I see system crashes day in, day out. From laptops bluescreening on wake-up from sleep mode, to explorer.exe hangs. My NT boxen had a nasty habit of surviving 7-14 days of moderate work before slowing down to molasses and poping a few VxDs.
On the other hand, my Linux, AIX, Solaris and HP-UX boxes all run full time between scheduled power downs pretty much all the time. Admittedly, the AIX, Solaris and HP-UX boxes generally run headless, but the desktop linux boxes around are rock-solid stable. I never saw uptimes above 100 days for any NT/2000 box, whereas that is minimum uptime for the Unix/Linux crowd.
So be glad of your supply of magic pixie dust for Windows.
Cheers,
Toby Haynes