Um, if you're working on a commercial project, be careful if you look at worldforge beyond architecture documentation, without at least talking to your legal department (or lawyer) first. WF is GPL and you are no more safe against copyright infringement suits if you copy GPL code than you would be if you got your hands on the code to EverQuest, Anarchy Online, or any other MMORPG.
Looking at code is tricky buisness. You can do it in some ways, and you're probably safe as long as you're not creating a direct competitor with a commercial product, or a work-alike of something. If you do a clean-room implementation of something there cannot be even a chance that you've seen the competitors code. If you're doing your own thing, and just glance at the code to get a basic feel for how they've done things, then write your own, different code, you can probably consider it fair use. Maybe. But it's tricky territory.
Stable systems run without a crash for longer than Windows 2000 has even been released.
To anyone upgrading from a dll-rotted Win 95 I'm sure Windows 2000 or XP seem really stable. To anyone who hasnt had their expectations quite so lowered, it still isnt near good enough.
Not quite. Microsoft isnt giving anyone anything, they're submitting a standard to a standards body. Ximian and others want to implement the good parts of that standard, to increase development speed and portability (and probably slow down executable speed (there's a reason all those 'lets make desktop apps in java' initiatives went boom)).
The point of it is that even if (well, rather when, in my opinion) Microsoft starts corrupting the standard it doesnt really matter because the point isnt really interoperability with MS, it is getting another useful development platform that isnt as tied to languages as the ones we have now.
Think of it this way; the entire world is coding in assembler. Along comes the evil empire and says 'hey, we've made up this C thing'. We can build a C compiler too, and use it for what it's worth. Sure enough, the Evil Empire went along and built incompatible API's outside their C reference which made the portability idea useless, but the language and the associated standards would still be a good thing.
Of course, I dont think Miguel is even near critical enough of MS, and heck, the whole.NET thing might actually only be a plot to trick the rest of the world to waste resources trying to implement something that wont work in practice, and when everyone is all worn out and tired, along they come and reveal what they've *really* been working on meanwhile... MS BOB 2.0.
Yep, and in sane patenting laws you cannot patent algorithms any more than you can patent any other form of basic mathematics, or discoveries. Patents should apply to inventions as originally intended.
True. Or, well, large datacenters hopefully (ha) have the sense to consolidate their DB's onto DB servers with 20-500 databases each.
Still, there are piles of machines with loads that could easily be served with a 5 yearold low to midrange machine... which will be the death of the large machines, it's getting hard to find applications of any but the most estoteric use to corporations that actually need 'serious' processing power anymore. Which, of course, is the reason that all the iron vendors are adding partitioning capabilities to their systems like mad so they can sell on the consolidation capability.
Uhm, QC is usually done for the separate components, and those are pretty much the same wether you build the computer yourself or a company builds it.
I have yet to see any computer assembly QC that exceeds 'it boots. cool.'. And I can do that very well myself, thankyou. Me (or rather an assembly line worker), my (well, his) screwdriver, and a poweron is what you get, despite your requirements.
You dont get reliable and predictable from anyone. Apple does not burn in the computers for several weeks to months, nor does anyone else, and that is the time you'd need to ensure that no components are gonna go poof due to miniscule and by QC undetectable chip flaws. Even then you can get faults from thermal stress, shipping, handling, etc, and eventually random faults and after that physical wearing out of things like disks, fans, etc.
And let me tell you, if you have a scientific cluster or any other cluster doing critical calculations and your software isnt doing savepoints you need drag the programmer out into the parking lot and flog him in public (or visit the supplier with a squad of goons with big sticks), because you _are_ going to have outages (and man do you have a software quality problem).
Well, hey, I can write a massively parallel distributed computing manual in _one slashdot comment_!
Requirements: Any machine supported by the seti-at-home client and some form of internet connection.
Installation: Download and install the client.
Select a parallel application: Easy! Already done for you!
Select nodes: Not your problem!
Congratulations, your computer is now part of a massively paralell computing network.
What? You want to do something else? Well... there's a 560 page manual... you know, parallel computing isnt that easy.
There is a difference between a quickstart guide and reference manuals. Comments such as that are just silly.
Of course, the article goes on to describe how they switched from applescript to tcp/ip... followed by the amazing 'they can transfer bigger chunks of data between nodes but their latency is less' which 'balances it out'. I find the conclusion that it is no harder to write for multiple processors (clusters) than it is to write for two processors (SMP) rather interesting too. Apparently he's got his hands on some damn hot networking technology, because last I looked the memory bus in an computer was a bit faster than your average network, which makes the problem quite a bit different.
Oh, well, someone has apparently been overdosing heavily on those one page quickstart guides.
Um, how about 'Microsoft'? Or maybe you missed the upgrade terms for XP? Either you upgrade now or you get to pay full price to upgrade later. Those are the breaks for a lot of companies.
As long as the code is GPL you can go the other way. Fork immediately at license change and convince as many developers as possible to either contribute to both branches or exclusively to the still-GPL branch. You need exclusive copyright to change licenses, but you dont need that to fork the code.
Actually, there are going to be less problems for anyone starting now. When Loki started there was no SDL; they bootstrapped that project. Neither were the distribution tools there.
Above all, when Loki started, Linux was still rather difficult to get working well with X. Some things still arent 'consumer' level easy, like getting 3d acceleration working. But the XFree 4 architecture improvements solves a lot of the problems that were putting a serious cramp on the ease of installation.
The API issue isnt really as big as it seems. The solution is called static linking. It's a disgusting solution but that's how it's usually done. How many games on Windows ship depending on specific versions of general system libraries? Trusting dynamic libraries is death-by-support for any binary-only release program.
The situation is better today. Largely thanks to Loki.
Yep, and USS was, while UNIX '95 compliant, not what your average Unix admin would call 'Unix' when they used it. From PITA EBCDIC problems to lack of things like ftp (did come later tho).
Your average unix sysadmin is willing to live with logfiles in a slightly non-standard place, and can accept that some OSs' have severe problems getting the erase char right, and has probably given up on getting lvms standardized between OSs' but he's not willing to live with a whole new world of imaginative new ways of being entirely different.
USS felt like it'd been forked off from mainstream unix in the early to mid 80's, spent 15 years in a closet somewhere and then had a programmer with a unix 95 spec thrown in with it a year before release. Almost, but not quite, entirely unlike unix.
Not quite right. The diodes _on_ the XP work just fine, there just arent many (any?) motherboards that support them.
The thermal diodes _outside_ the CPU, on the motherboards, are the ones that are slow. It's a case of the motherboard vendors not updating design to use the new thermal sensor in the XP.
Should you find a motherboard that does support the XP thermal diode it wont burn.
Funny, last time I checked, neither authors, nor musicians were making much money from their creativity. Yet they do keep producing it despite that they arent the ones making money off it anymore, dont they...
Your fantasy about copyright is long gone; it's what we _should_ have, but it isnt what we _do_ have. That is exactly why we do need to tear the current legislation up, because neither the common good nor the creative people gain anything from the current situation where the distributors are the only ones making any money.
Entirely true. It is, however, 'illegal' to do so in the US. And it's beyond most people who dont even understand that they cant record their TV shows like they used because it has been intentionally screwed up by the equipment providers, not because they cant 'get their video to work correctly with their digital decoder'.
What makes you think that someone doesnt have the right to do whatever they want with anyones artistic material?
Copyright? Copyright is granted as a means to ensure the maximum availability and production of new material for the common good, by offering a limited exclusive right to copy the material.
Of course, this idea is becoming more and more corrupted, far beyond what is for 'the public good'. IMO, it's time to tear up all copyright (and patent) laws, and start over with the public (rather than the corporate interest) good.
The same thing that happens to anyone who tries to build a non-macrovision compliant device today; they'll get sued by whoever is in charge (JVC, I believe, for VHS).
Of course, a few slip through, but that wont really change anything. Oh, and anyone trying to market and sell such a device in the US would get DMCA'd up their ass.
Actually, the reality is, just because you can see it doesnt mean your recording device does.
Try Macrovision. Screws up most VCR's available today. Screws up a lot of old TV's (if you've tried to play commercial videotapes on a new VCR to an old TV you might have noticed the top half of the picture being totally screwed (hint, it's not bad tracking, it's Macrovision screwing your TV)).
You wont 'loose some quality', you wont get a watchable picture.
So, you ever tried to play commercial video tapes to an older TV? Ever noticed how the upper half of the picture gets distorted? Now, it would be easy to think that's just a case of bad tracking, or a sucky tv, or something. Heck, I did, and got a new TV.
But likely it's due to the last copyprotection that went in this way. Macrovision.
Sure you'll get a signal out of the digital box. Just not a signal that is recordable on a lot of the systems today, because any old device wont work with it, and any new device will honor copy-protection code.
They did get away with it last time, why wouldnt they get away with making everyone buy new stuff again? It's not like it's done over a day, it's done over a device-generation. By the time its implemented all the way and switched on, 80 percent of the available devices will support the new coding, and the rest... well, you had a chance to decide for yourself wether you wanted to get a new TV and video, now you have to if you wish to watch any broadcasts.
This could be worked around by making companies/software developers liable for willful negligence for profit. The problems that really needs fixing in the industry are on one hand the inclusion of completely utterly braindead features that are inherently insecure, but are strategic for corporations like MS, and on the other hand programming and shipping deadlines that result in the inclusion of known suspect code for market reasons.
There is a general problem with software quality, but the real problem is when it _pays_ to ignore security, because then it will become a standard buisness practice. That is the behaviour that any law should be targetted at.
Um, have you ever read your Windows license agreements?
MS is liable for nothing. Your computer could spontaneously blow up and level your house because of the Windows Exploding Computer Feature, and you wouldnt get a dime from them.
Oh, yes, so lets have the government confiscate all buisnesses and replace them with monopolies. This pesky competition thing sure cant be the best thing for consumers and the economy.
Face it, Microsoft is harming the US and world economy constantly, and the harm the company causes will only grow worse.
Different markets tho; Motorola had competition. Cable competition is lacking. And when competition is lacking... well, you want it, you pay.
Just wait 'til the monthly $450 Windows XP charges start rolling in. People sure will 'like the convenience'... but unless the anti-trust cases work out, that'll be the only game in town if you want to use anything electronic.
Um, we just set up yet another Linux install for relatives (we'd figured on making it dual boot, but Windows wasnt compatible with the hardware (cant install windows if the bios cant find the disk if it's too large. No problem with linux tho)).
And guess what? Works perfectly. They click the 'internet' button and it connects. They click the 'mail' button and read their mail... etc etc. And guess something else? They cant blow up the system by mistake, nor can the grandchildren when they play with the computer, because they have separate accounts on the machine.
Face it, Windows is _NOT_ easy to use, only easy to mess up for a new user. The only ones who find it 'easier' than linux are those used to Windows and who've never used anything else.
Those without computer experience wont be installing either Linux or Windows or any other OS soon (so, what would your elderly belgian woman say when the Windows install cd says 'cant find a hard disk'?). Linux has the advantage that when you have installed it for them, and if you tell them not to login with username root, then it wont blow up on them because they or a grandchild clicked somewhere they shouldnt.
For those using a computer for the first time, Linux fits the cheap, easy to use, does their basic stuff, and wont randomly break bill _perfectly_. While Windows is neither cheap, nor any easier, and will randomly break.
Novell doesnt have the desktop clout to ram it down everyone and their grandmothers throat wether they want it or not.
Sure, Novell would have the tech to do it, but hell, this is the IT industry, which isnt about technology. It's about Microsoft taxing you for 30% of your income for the rest of your life. Nothing else.
Nope, my DSL is static IP, and the TOS allow me to run any non-commercial servers I want, for about $35 per month. Sure, it happens it's slow, and sometimes they have outages. But with the usual speed, the usual uptime and the generous terms of service... well, Im not going to complain.
I love my DSL provider. Heck, I used to pay $60 per month for unlimited time modem connection. And before that I paid on average $120 per month for standard charge metered access.
If you want decent DSL service, dont go with the major telecom corps, because they'll do crap like this.
Slashdot Mirror
Um, if you're working on a commercial project, be careful if you look at worldforge beyond architecture documentation, without at least talking to your legal department (or lawyer) first. WF is GPL and you are no more safe against copyright infringement suits if you copy GPL code than you would be if you got your hands on the code to EverQuest, Anarchy Online, or any other MMORPG.
Looking at code is tricky buisness. You can do it in some ways, and you're probably safe as long as you're not creating a direct competitor with a commercial product, or a work-alike of something. If you do a clean-room implementation of something there cannot be even a chance that you've seen the competitors code. If you're doing your own thing, and just glance at the code to get a basic feel for how they've done things, then write your own, different code, you can probably consider it fair use. Maybe. But it's tricky territory.
Stable systems run without a crash for longer than Windows 2000 has even been released.
To anyone upgrading from a dll-rotted Win 95 I'm sure Windows 2000 or XP seem really stable. To anyone who hasnt had their expectations quite so lowered, it still isnt near good enough.
Not quite. Microsoft isnt giving anyone anything, they're submitting a standard to a standards body. Ximian and others want to implement the good parts of that standard, to increase development speed and portability (and probably slow down executable speed (there's a reason all those 'lets make desktop apps in java' initiatives went boom)).
.NET thing might actually only be a plot to trick the rest of the world to waste resources trying to implement something that wont work in practice, and when everyone is all worn out and tired, along they come and reveal what they've *really* been working on meanwhile... MS BOB 2.0.
The point of it is that even if (well, rather when, in my opinion) Microsoft starts corrupting the standard it doesnt really matter because the point isnt really interoperability with MS, it is getting another useful development platform that isnt as tied to languages as the ones we have now.
Think of it this way; the entire world is coding in assembler. Along comes the evil empire and says 'hey, we've made up this C thing'. We can build a C compiler too, and use it for what it's worth. Sure enough, the Evil Empire went along and built incompatible API's outside their C reference which made the portability idea useless, but the language and the associated standards would still be a good thing.
Of course, I dont think Miguel is even near critical enough of MS, and heck, the whole
Yep, and in sane patenting laws you cannot patent algorithms any more than you can patent any other form of basic mathematics, or discoveries. Patents should apply to inventions as originally intended.
True. Or, well, large datacenters hopefully (ha) have the sense to consolidate their DB's onto DB servers with 20-500 databases each.
Still, there are piles of machines with loads that could easily be served with a 5 yearold low to midrange machine... which will be the death of the large machines, it's getting hard to find applications of any but the most estoteric use to corporations that actually need 'serious' processing power anymore. Which, of course, is the reason that all the iron vendors are adding partitioning capabilities to their systems like mad so they can sell on the consolidation capability.
Uhm, QC is usually done for the separate components, and those are pretty much the same wether you build the computer yourself or a company builds it.
I have yet to see any computer assembly QC that exceeds 'it boots. cool.'. And I can do that very well myself, thankyou. Me (or rather an assembly line worker), my (well, his) screwdriver, and a poweron is what you get, despite your requirements.
You dont get reliable and predictable from anyone. Apple does not burn in the computers for several weeks to months, nor does anyone else, and that is the time you'd need to ensure that no components are gonna go poof due to miniscule and by QC undetectable chip flaws. Even then you can get faults from thermal stress, shipping, handling, etc, and eventually random faults and after that physical wearing out of things like disks, fans, etc.
And let me tell you, if you have a scientific cluster or any other cluster doing critical calculations and your software isnt doing savepoints you need drag the programmer out into the parking lot and flog him in public (or visit the supplier with a squad of goons with big sticks), because you _are_ going to have outages (and man do you have a software quality problem).
Well, hey, I can write a massively parallel distributed computing manual in _one slashdot comment_!
Requirements: Any machine supported by the seti-at-home client and some form of internet connection.
Installation: Download and install the client.
Select a parallel application: Easy! Already done for you!
Select nodes: Not your problem!
Congratulations, your computer is now part of a massively paralell computing network.
What? You want to do something else? Well... there's a 560 page manual... you know, parallel computing isnt that easy.
There is a difference between a quickstart guide and reference manuals. Comments such as that are just silly.
Of course, the article goes on to describe how they switched from applescript to tcp/ip... followed by the amazing 'they can transfer bigger chunks of data between nodes but their latency is less' which 'balances it out'. I find the conclusion that it is no harder to write for multiple processors (clusters) than it is to write for two processors (SMP) rather interesting too. Apparently he's got his hands on some damn hot networking technology, because last I looked the memory bus in an computer was a bit faster than your average network, which makes the problem quite a bit different.
Oh, well, someone has apparently been overdosing heavily on those one page quickstart guides.
Um, how about 'Microsoft'? Or maybe you missed the upgrade terms for XP? Either you upgrade now or you get to pay full price to upgrade later. Those are the breaks for a lot of companies.
As long as the code is GPL you can go the other way. Fork immediately at license change and convince as many developers as possible to either contribute to both branches or exclusively to the still-GPL branch. You need exclusive copyright to change licenses, but you dont need that to fork the code.
Actually, there are going to be less problems for anyone starting now. When Loki started there was no SDL; they bootstrapped that project. Neither were the distribution tools there.
Above all, when Loki started, Linux was still rather difficult to get working well with X. Some things still arent 'consumer' level easy, like getting 3d acceleration working. But the XFree 4 architecture improvements solves a lot of the problems that were putting a serious cramp on the ease of installation.
The API issue isnt really as big as it seems. The solution is called static linking. It's a disgusting solution but that's how it's usually done. How many games on Windows ship depending on specific versions of general system libraries? Trusting dynamic libraries is death-by-support for any binary-only release program.
The situation is better today. Largely thanks to Loki.
Yep, and USS was, while UNIX '95 compliant, not what your average Unix admin would call 'Unix' when they used it. From PITA EBCDIC problems to lack of things like ftp (did come later tho).
Your average unix sysadmin is willing to live with logfiles in a slightly non-standard place, and can accept that some OSs' have severe problems getting the erase char right, and has probably given up on getting lvms standardized between OSs' but he's not willing to live with a whole new world of imaginative new ways of being entirely different.
USS felt like it'd been forked off from mainstream unix in the early to mid 80's, spent 15 years in a closet somewhere and then had a programmer with a unix 95 spec thrown in with it a year before release. Almost, but not quite, entirely unlike unix.
Not quite right. The diodes _on_ the XP work just fine, there just arent many (any?) motherboards that support them.
The thermal diodes _outside_ the CPU, on the motherboards, are the ones that are slow. It's a case of the motherboard vendors not updating design to use the new thermal sensor in the XP.
Should you find a motherboard that does support the XP thermal diode it wont burn.
Funny, last time I checked, neither authors, nor musicians were making much money from their creativity. Yet they do keep producing it despite that they arent the ones making money off it anymore, dont they...
Your fantasy about copyright is long gone; it's what we _should_ have, but it isnt what we _do_ have. That is exactly why we do need to tear the current legislation up, because neither the common good nor the creative people gain anything from the current situation where the distributors are the only ones making any money.
Entirely true. It is, however, 'illegal' to do so in the US. And it's beyond most people who dont even understand that they cant record their TV shows like they used because it has been intentionally screwed up by the equipment providers, not because they cant 'get their video to work correctly with their digital decoder'.
What makes you think that someone doesnt have the right to do whatever they want with anyones artistic material?
Copyright? Copyright is granted as a means to ensure the maximum availability and production of new material for the common good, by offering a limited exclusive right to copy the material.
Of course, this idea is becoming more and more corrupted, far beyond what is for 'the public good'. IMO, it's time to tear up all copyright (and patent) laws, and start over with the public (rather than the corporate interest) good.
The same thing that happens to anyone who tries to build a non-macrovision compliant device today; they'll get sued by whoever is in charge (JVC, I believe, for VHS).
Of course, a few slip through, but that wont really change anything. Oh, and anyone trying to market and sell such a device in the US would get DMCA'd up their ass.
Actually, the reality is, just because you can see it doesnt mean your recording device does.
Try Macrovision. Screws up most VCR's available today. Screws up a lot of old TV's (if you've tried to play commercial videotapes on a new VCR to an old TV you might have noticed the top half of the picture being totally screwed (hint, it's not bad tracking, it's Macrovision screwing your TV)).
You wont 'loose some quality', you wont get a watchable picture.
So, you ever tried to play commercial video tapes to an older TV? Ever noticed how the upper half of the picture gets distorted? Now, it would be easy to think that's just a case of bad tracking, or a sucky tv, or something. Heck, I did, and got a new TV.
But likely it's due to the last copyprotection that went in this way. Macrovision.
Sure you'll get a signal out of the digital box. Just not a signal that is recordable on a lot of the systems today, because any old device wont work with it, and any new device will honor copy-protection code.
They did get away with it last time, why wouldnt they get away with making everyone buy new stuff again? It's not like it's done over a day, it's done over a device-generation. By the time its implemented all the way and switched on, 80 percent of the available devices will support the new coding, and the rest... well, you had a chance to decide for yourself wether you wanted to get a new TV and video, now you have to if you wish to watch any broadcasts.
This could be worked around by making companies/software developers liable for willful negligence for profit. The problems that really needs fixing in the industry are on one hand the inclusion of completely utterly braindead features that are inherently insecure, but are strategic for corporations like MS, and on the other hand programming and shipping deadlines that result in the inclusion of known suspect code for market reasons.
There is a general problem with software quality, but the real problem is when it _pays_ to ignore security, because then it will become a standard buisness practice. That is the behaviour that any law should be targetted at.
Um, have you ever read your Windows license agreements?
MS is liable for nothing. Your computer could spontaneously blow up and level your house because of the Windows Exploding Computer Feature, and you wouldnt get a dime from them.
Oh, yes, so lets have the government confiscate all buisnesses and replace them with monopolies. This pesky competition thing sure cant be the best thing for consumers and the economy.
Face it, Microsoft is harming the US and world economy constantly, and the harm the company causes will only grow worse.
Different markets tho; Motorola had competition. Cable competition is lacking. And when competition is lacking... well, you want it, you pay.
Just wait 'til the monthly $450 Windows XP charges start rolling in. People sure will 'like the convenience'... but unless the anti-trust cases work out, that'll be the only game in town if you want to use anything electronic.
Um, we just set up yet another Linux install for relatives (we'd figured on making it dual boot, but Windows wasnt compatible with the hardware (cant install windows if the bios cant find the disk if it's too large. No problem with linux tho)).
And guess what? Works perfectly. They click the 'internet' button and it connects. They click the 'mail' button and read their mail... etc etc. And guess something else? They cant blow up the system by mistake, nor can the grandchildren when they play with the computer, because they have separate accounts on the machine.
Face it, Windows is _NOT_ easy to use, only easy to mess up for a new user. The only ones who find it 'easier' than linux are those used to Windows and who've never used anything else.
Those without computer experience wont be installing either Linux or Windows or any other OS soon (so, what would your elderly belgian woman say when the Windows install cd says 'cant find a hard disk'?). Linux has the advantage that when you have installed it for them, and if you tell them not to login with username root, then it wont blow up on them because they or a grandchild clicked somewhere they shouldnt.
For those using a computer for the first time, Linux fits the cheap, easy to use, does their basic stuff, and wont randomly break bill _perfectly_. While Windows is neither cheap, nor any easier, and will randomly break.
Novell doesnt have the desktop clout to ram it down everyone and their grandmothers throat wether they want it or not.
Sure, Novell would have the tech to do it, but hell, this is the IT industry, which isnt about technology. It's about Microsoft taxing you for 30% of your income for the rest of your life. Nothing else.
Nope, my DSL is static IP, and the TOS allow me to run any non-commercial servers I want, for about $35 per month. Sure, it happens it's slow, and sometimes they have outages. But with the usual speed, the usual uptime and the generous terms of service... well, Im not going to complain.
I love my DSL provider. Heck, I used to pay $60 per month for unlimited time modem connection. And before that I paid on average $120 per month for standard charge metered access.
If you want decent DSL service, dont go with the major telecom corps, because they'll do crap like this.