Yes, BSD may have used this partition type before Thinkpads (although Thinkpads have had this kind of suspend feature for a long, long time). But messing around with its BIOS and suspend feature would be a major risk for IBM; I can understand that they don't want to do that. In such cases, realistically, open source software is what can and probably should adapt.
Maybe BSD could define an alternative partition type for its standard file system, something that can be used on Thinkpads.
For secure mail, you need a client that contains encryption and that you can trust: you need to be able to trust its encryption, and you need to know that it is free of back doors. The content needs to travel encrypted from your mail client to the recipient's mail client.
A Java applet with well-known source code might begin to give you that kind of trust (if you trust your Java application). A C or Perl program, small enough to be reviewed, might as well.
A web browser with SSL just doesn't do the right thing since the mail arrives in cleartext on the web server, and a closed source client like Outlook simply can't be trusted to be free of backdoors or other problems at all.
Well, given that my deck outside right now measures at 256 times brighter than my computer screen, it's probably not a problem. The human eye is adaptable to a huge range of brightnesses. I believe most retinal damage is caused by actual heating (if you focus on a a really bright, small object) or UV light. With retinal scan technology, I'd mostly worry about the scanning apparatus failing.
I believe it was proven, not just conjectured, in the 70's or 80's that the game of life is Turing universal. The proof, as I recall, was constructive.
How do you construct such a thing? By building up abstractions. You construct "wires" from gliders and mirrors and define basic logic gates. For storage, delay lines (composed of gliders and movable mirrors) are one choice. To put everything together, writing a program that compiles logic into an initial state is probably a good idea.
There are actually two separate issues for identifying hosts: location independence and human readability. Arguably, one problem with the current Internet naming system is that it tries to address both issues with one system. So, people will haggle endlessly over names like "toys.com" because it has to address a unique host and once someone has adopted it, all links to them would go bad if they changed.
Separating the two concerns would give you a system where you have a two step resolution: human readable to location independent numbers, followed by location independent numbers to IP addresses. Such a split gives you a lot more flexibility on mapping the human readable names because you can change the mapping without having all the web pages that point to the affected hosts go bad.
Of course, these location independent numbers should not be phone numbers, since phone numbers do, in fact, change.
Device independent output was a great idea. And if PostScript had remained exclusively a means for sending output to printers, it would have been OK as well. But PostScript has become a document format: people need to manipulate it, reverse it, change fonts, and do all sorts of other things with it. And for that, a procedural programming language is a poor choice; you need something declarative.
PDF is an attempt to adapt PostScript to that purpose, but it isn't a very good one.
It would work itself out: users have no interest in using name servers that don't let them resolve links, and information providers have no interest in registering with servers that aren't widely used.
I think anyone who dreams of an internet populated with contradictory name tables needs to come down to earth. It sounds like a disaster to me. Whenever you told someone a URL, you'd have to explain which DNS table supported it. Let's get real. Most people don't know or care about the issue, and whatever their ISP provides is what they'll use.
Because it's a disaster, it's not a stable state: for any TLD, only one would survive in pretty short order. It's the ISPs that basically make the choice for most people, although you'll probably also start seeing buttons on web pages "click here to add.SNAFU to your universe". Maybe ISPs would even create some form of coordinating body, but with a structure different from ICANN.
Think of it more as a marketplace: rather than having ICANN making some ex cathedra decision about what TLDs are worthy, lots of people would just create TLDs and ISPs and end users would coordinate and pick and choose. Or think of it as the equivalent of the "alt.*" newsgroups.
I'm not sure whether it's really more of a technical issue or whether it's just that the PRC can do this because they have the power to force the vast majority of Chinese speakers to use whatever name server they choose.
Well, it's pretty obviously the latter, since formally, the only domain China has any say in is ".cn".
What is kind of interesting, of course, is how China plans on coordinating with the other countries using Chinese characters; I doubt, for example, that the Japanese are willing to let the Chinese determine all uses of Chinese characters (many of them identical to Japanses Kanji) in the ".com" TLD.
What is clear is that there is no reason for those countries to give VeriSign hundreds of millions of dollars for having been granted a US government monopoly at some point in the past.
Meanwhile, China is threatening to run an alternative DNS
and force VeriSign out of the business of registering Chinese-character domains. Does anyone really
think they'd allow the Chinese-character equivalent of independence.tw?
How China resolves names is their business, and whether you use their servers is yours.
I think most people outside China would recognize Taiwan as the ultimate authority on "independence.tw" (in Chinese or Western characters), as well as the ultimate authority on the Chinese character equivalent of the ".tw" TLD, and that's how most people would set up their name resolvers.
Why doesn't each country just take control of their own ccTLD, and leave gTLDs to ICANN?
Well, one problem discussed here is that ICANN apparently is trying to exert control over ccTLDs as well. If they stopped trying that, I think people would be less upset with them.
But one might also ask why ICANN should have any significant control over gTLDs or any involvement in the infrastructure. You don't need centralized control over gTLDs to avoid clashes, and you certainly don't need centralized control over any part of the infrastructure to make things work.
There is no reason for having a single organization administering the DNS system. Even with the current system, both individuals and ISPs can perform name resolution using whatever collection of servers they like. ICANN has no business messing with the.uk or.de domains: if people want to resolve names in those domains, they (or their ISPs) just point to the right servers. If your ISP doesn't resolve things the way you like, you can use a third party name server, or you can change ISPs.
There is some coordination needed to help people avoid creating conflicting TLDs. If both the UK and the US create a ".biz" within their servers, it would be bad for both. Serious registrars will cooperate, and if they won't, users just won't point at them. In either case, the function an organization like ICANN would perform would be a minor, administrative one, not justifying their current size, power, or charge structure: maintaining a list of those TLDs.
Even with the current DNS infrastructure, ICANN is technically and administratively superfluous. I hope the ccTLD administrators will leave the current system: sooner or later, it is destined for demise anyway, and it might as well be sooner.
Of course, Windows and Linux share code: networking-related code from BSD. And I wouldn't be surprised if Windows has borrowed whereever legally possible from other open source projects.
You can also bet that Microsoft analyzes open source code. As benchmark-obsessed as that company is, how could they not? If GNU C, Apache, some device driver, or some other piece of software does better on some benchmark, Microsoft will surely analyze what is going on and reimplement whatever they learn in Windows.
Keep in mind that the reverse has happened to some degree for years as well: for example, people have disassembled Windows to figure out how to talk to specific devices.
Given that Microsoft is a large organization with lots of closed source and many fairly inexperienced programmers fresh out of college, I wouldn't be surprised if someone at Microsoft had violated the GPL. But I somehow doubt that it is Microsoft company policy to actually incorporate GPL'ed code--it doesn't make economic sense, and Microsoft is largely about making money.
So, I think altogether, there is a kernel of truth to the assertion that "Microsoft takes code from Linux". But I that isn't the same as widespread, blatant violations of the GPL.
I think contracts that make employees pay for training that they received are open to all sorts of abuse. In fact, there is no reason at all why the value of the training to the employee should be as high as the cost of the training to the company. And the cost of the training could be unreasonably high.
A better approach would be to incentivize employees to stay. Something like: after the training, your salary will be subject to the normal pay increases. But a year after completion of training, you will receive an additional 10% raise, together with a bonus or options valued at 15% of your current salary.
With that kind of approach, the cost and value of the training is properly accounted for by the company and there is no incentive for the company to cheat or force employees to accept unreasonably costly training. The risk of the employee is limited to the value of his new training to the company: if he leaves, he loses exactly as much money as his new training is worth to the company over a year.
Did I think she was ready for the job she took? No, not yet. Certification is nice as a checklist item, and
the training process is useful, but you need real-world experience, too - and she didn't really have
enough for the role yet. I expect she'll ultimately do very well there because she's a fast learner, but
based on her current skills I couldn't have offered her equivalent pay to stay. However, she was being
paid reasonably well, and did leave only a few months after completing MCSE training.
Well, you made a simple economic mistake: cost and pay shouldn't take into account just current utility but also future utility.
The bottom line for us out of this affair was that our company will still pay for education, but we have now
added a 1-year requirement - they have to stay for at least a year after taking their final test for any sort
of certification.
Economically, that is equivalent to lowering people's salaries: training is part of the job, you just shifted your risk onto your employees, risk=money, and you also removed part of your economic incentive to give people a salary increase.
The bargain we now
require is "we'll pay you reasonably and pay for your advancement - but you have to promise us that
you'll let us get some benefit from our investments".
The bargain is more simply translated as "our salary/benefits have decreased". If the cost of training is substantial, more sinister translations are possible.
To me, ongoing training is part of any job. You don't expect people to pay you back for the last year of health insurance or the last year of 401k matching when they leave. And some of the most important training is not junk certifications like MSCE, it's on-the-job learning; are you going to make people indebted for that as well?
Forcing employees to take on is business expenses like that looks like just a transparent attempt to tie employees to your company at substandard wages or substandard working conditions. It smells of the kind of financial abuses migrant farmworkers or people who sign into iffy marketing schemes are subject to.
I hope you are honest about your policies with new hires before they sign up. In fact, you could save people more time by just letting us know what your company is so that they don't apply.
One: If they leave within a year, make them pay for the training.
If there is any significant amount of money involved, I would consider this inappropriate. If the company selects the training program, they pay for it; otherwise, there is too little incentive for them to make a good choice, and it opens up employees to all sorts of possibilities for abuse and fraud. If we are only talking a few thousand dollars, it should make no difference to the company anyway.
If the company trains you and adjust your salary and working conditions appropriately, you shouldn't be any more likely to leave after the training than before. If they don't, that should be their problem, not yours.
Consumers in the US also pay additional fees like that (on blank media, including CD-R media, I believe). That was part of a copyright reform act a few years ago, and the deal was the same: consumer pay this extra fee, and in return, big content distributors would stop complaining about consumers taking advantage of fair use provisions (you can find lots of excellent links to the US laws in the discussions the last few times this came up on Slashdot).
But, of course, collecting these fees wasn't enough for the US industry--they keep merrily complaining about piracy and implementing various technical means to deny consumers fair use rights. And the same will probably be happening in Germany.
The International Federation of the Phonographic industry claims "about 500 million CDs are pirated annually by people creating their own CDs from downloaded tunes off the Internet. More than 25 million pirated music files are available online, the group estimates." I find that hard to believe. That would seem to be a large fraction of CD-R/CD-RW sales, but most of them are used for computer backup and other purposes.
Why do newspapers just reprint biased estimates from institutions with an axe to grind? This is where journalists should get busy and do a bit of background research. In particular when the numbers are so out of whack.
This is particularly common with AP stories, where no journalist ever signs responsible for it. The AP seems to have turned into a glorified PR newswire.
we need sound social policy, not GMO
on
Golden Rice
·
· Score: 2
Given that this plant has been developed and can save people's lives and sight, it should probably be made available if it is safe. However, developments like this will not help humanity in the long run.
We have already increased agricultural productivity, yield, and nutritional value many times over over the last few centuries. Initial improvements gave us the free time we needed to build our civilization. But beyond that point, productivity improvements have not stopped hunger, malnutrition, or overpopulation.
At this point, we don't need more productive or more nutritious crops. If we had the political will, we could already feed the current population well with the crops we have. What we really need to work on limiting the size of earth's population through social progress and access to family planning, as well as a more uniform distribution of wealth throughout the world. If we don't accomplish that, huge numbers of people will continue to die from hunger, malnutrition, and war, no matter what genetic gimmicks we invent.
Let me prefix this by saying that I think technically KDE is a good achievement. It's pretty zippy, quite robust, and seems well-written. I think on technical grounds, KDE does not need to fear comparison with Gnome.
But I find the KDE "mission statement" unacceptable:
"To establish KDE as a desktop standard for PCs, workstations, and mobile devices, to promote software development for KDE and to promote the use of KDE by enterprises and individuals."
To me, free software is about diversity, the ability to try things out, and the ability to customize software to everybody's needs. It's the notion of a "standard" that bothers me about Windows, a "standard" that, in the case of Windows is enforced through closed source releases. But KDE can just as effectively "enforce" standards in the open source world--there are plenty of mechanisms. Open source software is meaningless if it results in the same kind of dull standardization that the closed source world has suffered under.
If it is true that KDE is already on 70% of the Linux desktops, as the KDE League claims, I can only conclude that it is time to stop installing it, promoting it, and recommending it, and to work on something else. And if KDE really succeeds at marginalizing other Linux desktops, I'd just switch to Windows. I really don't see a difference between being forced to use a "standard" system from Microsoft or from KDE.
Fortunately, the UNIX and X11 architecture still makes that possibility fairly remote on the desktop. But on palmtops, Troll Tech really does want to take over with Qt/Embedded, which would not seem to allow non-Qt applications to co-exist on the same screen.
If KDE succeeds at its mission, it will have failed.
What is the future of Gnome and KDE on non-Linux desktops?
Because of the cost to commercial developers of the underlying toolkit, Qt, I think the future of KDE on non-Linux desktops is not too bright. OS vendors have little interest in putting their own customers at the mercy of a third party vendor.
The only way I see that change is if either someone buys Troll Tech outright and re-releases the toolkit for free, or if individual vendors make deals similar to Motif licenses, where the Motif stuff was simply included with the OS, free even for commercial development. But do we really want to go back to those days?
Too hard ? If everyone was like you there would be no Linux.
Well, you are making my point: if something isn't easy enough to try, people won't bother. Linux provided enough functionality and advantages over alternatives for early adopters to get to the point where easy-to-install distributions got created and it became useful to many more people. Darwin seems to still have to cross that threshold.
I am saddened if you can't even install a *nix OS without a wonderful graphical install program.
That would be sad, indeed. Fortunately, I can assure you that that's not a problem: I have manually entered boot loaders in machine code and installed 4.1BSD from tape. But, these days, IMO, a reasonably mature system should come with some kind of installer, just like it comes with a lot of other things that systems didn't use to come with.
That's interesting, but I still don't see the difference. I thought both Darwin and the Hurd retained the Mach core and the Mach architecture of interacting server processes. Both provide UNIX compatibility through Mach "servers". I also thought that some of Darwin's non-UNIX facilities (traditional MacOS stuff) was implemented taking advantage of the Mach architecture rather than being built on top of its UNIX personality.
Are you saying Apple somehow removed the ability to implement additional Mach servers? Does the UNIX personality of Darwin violate the modularity of the Mach server architecture? If that were true, then that would indeed be a strong argument in favor of the Hurd. Or are you saying that the Hurd happens to come with a wider variety of servers right now? If that's the case, probably many of the Hurd servers could be ported over quickly.
Maybe it would be good for someone to write a more detailed technical comparison of the Hurd and Darwin. So far, the bird's eye view is that both system are based on the Mach microkernel, both have a variety of servers for them, and both have a UNIX personality. So, to me, the question remains: why pick one over the other? (Of course, there may still also be licensing issues.)
The rational design for that kind of database is to put the image data into the file system and use the relational database for indexing and lookups. Most of the open source databases are perfectly up to the task of providing indexing for that kind of data. In fact, the amount of metadata in such applications is small compared to the kinds of data encountered in many commercial applications, so this is actually not even a particularly interesting benchmark for high-end database systems.
Putting multimedia data into the file system is the implementation strategy many commercial databases (including some versions of DB2) take behind the scenes for storing multimedia objects, even if they hide it behind a database API. They can still provide all the database facilities (transactions, indexing, access control, etc.) on top of such an implementation.
With that kind of architecture, you don't need a very powerful machine or high performance database to be able to serve image data at disk bandwidth or network bandwidth.
I would very much like to give Darwin a try, but it seems like it's still too hard.
What made me pick up Linux in 1994 was the fact that it came as a complete distribution, pretty much ready to install and run. Those distributions had a number of limitations when it came to drivers and tools, but they were usable and could be used to solve specific problems.
I haven't been able to find a complete distribution for the PC based on the Darwin kernel. Such a distribution would require the kernel, the command line utilities, development tools, X11, and at least one desktop (Gnome, KDE, GNUStep,...). Such a distribution would be useful even if the set of available drivers is pretty limited (IDE, maybe a SCSI card, a couple of common Ethernet cards).
Darwin is based on the same technology as the Hurd, and it has seen more maintenance over the years. Perhaps it would also offer better compatibility with a commercial OS (MacOS X). So, why not use the Darwin kernel instead of the Hurd?
I think this is mostly done for Gray and Microsoft to get publicity for their database. This is the continuation of TerraServer and other projects like that. Microsoft is trying to demonstrate "scalability" of their database and servers and to get it into the hardcore scientific server area.
As usual, Microsoft is late to the party and comes with their own agenda. Microsoft products are oriented towards small business and desktop applications. That's what their evolution is driven by and that's what they are designed for. Whether this kind of data should be in a relational database is questionable to begin with. And it certainly doesn't need to be on an expensive, proprietary operating system and in a proprietary format.
Scientists already have excellent open-source tools to build long-term, stable, large-scale data collections. They would be foolish to tie research projects that can span decades to the fortunes of a company in the middle of a battle for the US business computing market, merely to gain some trinkets and give that company a publicity boost.
Slashdot Mirror
Maybe BSD could define an alternative partition type for its standard file system, something that can be used on Thinkpads.
A Java applet with well-known source code might begin to give you that kind of trust (if you trust your Java application). A C or Perl program, small enough to be reviewed, might as well.
A web browser with SSL just doesn't do the right thing since the mail arrives in cleartext on the web server, and a closed source client like Outlook simply can't be trusted to be free of backdoors or other problems at all.
Well, given that my deck outside right now measures at 256 times brighter than my computer screen, it's probably not a problem. The human eye is adaptable to a huge range of brightnesses. I believe most retinal damage is caused by actual heating (if you focus on a a really bright, small object) or UV light. With retinal scan technology, I'd mostly worry about the scanning apparatus failing.
How do you construct such a thing? By building up abstractions. You construct "wires" from gliders and mirrors and define basic logic gates. For storage, delay lines (composed of gliders and movable mirrors) are one choice. To put everything together, writing a program that compiles logic into an initial state is probably a good idea.
Separating the two concerns would give you a system where you have a two step resolution: human readable to location independent numbers, followed by location independent numbers to IP addresses. Such a split gives you a lot more flexibility on mapping the human readable names because you can change the mapping without having all the web pages that point to the affected hosts go bad.
Of course, these location independent numbers should not be phone numbers, since phone numbers do, in fact, change.
PDF is an attempt to adapt PostScript to that purpose, but it isn't a very good one.
It would work itself out: users have no interest in using name servers that don't let them resolve links, and information providers have no interest in registering with servers that aren't widely used.
I think anyone who dreams of an internet populated with contradictory name tables needs to come down to earth. It sounds like a disaster to me. Whenever you told someone a URL, you'd have to explain which DNS table supported it. Let's get real. Most people don't know or care about the issue, and whatever their ISP provides is what they'll use.
Because it's a disaster, it's not a stable state: for any TLD, only one would survive in pretty short order. It's the ISPs that basically make the choice for most people, although you'll probably also start seeing buttons on web pages "click here to add .SNAFU to your universe". Maybe ISPs would even create some form of coordinating body, but with a structure different from ICANN.
Think of it more as a marketplace: rather than having ICANN making some ex cathedra decision about what TLDs are worthy, lots of people would just create TLDs and ISPs and end users would coordinate and pick and choose. Or think of it as the equivalent of the "alt.*" newsgroups.
I'm not sure whether it's really more of a technical issue or whether it's just that the PRC can do this because they have the power to force the vast majority of Chinese speakers to use whatever name server they choose.
Well, it's pretty obviously the latter, since formally, the only domain China has any say in is ".cn".
What is kind of interesting, of course, is how China plans on coordinating with the other countries using Chinese characters; I doubt, for example, that the Japanese are willing to let the Chinese determine all uses of Chinese characters (many of them identical to Japanses Kanji) in the ".com" TLD.
What is clear is that there is no reason for those countries to give VeriSign hundreds of millions of dollars for having been granted a US government monopoly at some point in the past.
How China resolves names is their business, and whether you use their servers is yours.
I think most people outside China would recognize Taiwan as the ultimate authority on "independence.tw" (in Chinese or Western characters), as well as the ultimate authority on the Chinese character equivalent of the ".tw" TLD, and that's how most people would set up their name resolvers.
Why doesn't each country just take control of their own ccTLD, and leave gTLDs to ICANN?
Well, one problem discussed here is that ICANN apparently is trying to exert control over ccTLDs as well. If they stopped trying that, I think people would be less upset with them.
But one might also ask why ICANN should have any significant control over gTLDs or any involvement in the infrastructure. You don't need centralized control over gTLDs to avoid clashes, and you certainly don't need centralized control over any part of the infrastructure to make things work.
There is some coordination needed to help people avoid creating conflicting TLDs. If both the UK and the US create a ".biz" within their servers, it would be bad for both. Serious registrars will cooperate, and if they won't, users just won't point at them. In either case, the function an organization like ICANN would perform would be a minor, administrative one, not justifying their current size, power, or charge structure: maintaining a list of those TLDs.
Even with the current DNS infrastructure, ICANN is technically and administratively superfluous. I hope the ccTLD administrators will leave the current system: sooner or later, it is destined for demise anyway, and it might as well be sooner.
You can also bet that Microsoft analyzes open source code. As benchmark-obsessed as that company is, how could they not? If GNU C, Apache, some device driver, or some other piece of software does better on some benchmark, Microsoft will surely analyze what is going on and reimplement whatever they learn in Windows.
Keep in mind that the reverse has happened to some degree for years as well: for example, people have disassembled Windows to figure out how to talk to specific devices.
Given that Microsoft is a large organization with lots of closed source and many fairly inexperienced programmers fresh out of college, I wouldn't be surprised if someone at Microsoft had violated the GPL. But I somehow doubt that it is Microsoft company policy to actually incorporate GPL'ed code--it doesn't make economic sense, and Microsoft is largely about making money.
So, I think altogether, there is a kernel of truth to the assertion that "Microsoft takes code from Linux". But I that isn't the same as widespread, blatant violations of the GPL.
A better approach would be to incentivize employees to stay. Something like: after the training, your salary will be subject to the normal pay increases. But a year after completion of training, you will receive an additional 10% raise, together with a bonus or options valued at 15% of your current salary.
With that kind of approach, the cost and value of the training is properly accounted for by the company and there is no incentive for the company to cheat or force employees to accept unreasonably costly training. The risk of the employee is limited to the value of his new training to the company: if he leaves, he loses exactly as much money as his new training is worth to the company over a year.
Well, you made a simple economic mistake: cost and pay shouldn't take into account just current utility but also future utility.
The bottom line for us out of this affair was that our company will still pay for education, but we have now added a 1-year requirement - they have to stay for at least a year after taking their final test for any sort of certification.
Economically, that is equivalent to lowering people's salaries: training is part of the job, you just shifted your risk onto your employees, risk=money, and you also removed part of your economic incentive to give people a salary increase.
The bargain we now require is "we'll pay you reasonably and pay for your advancement - but you have to promise us that you'll let us get some benefit from our investments".
The bargain is more simply translated as "our salary/benefits have decreased". If the cost of training is substantial, more sinister translations are possible.
Forcing employees to take on is business expenses like that looks like just a transparent attempt to tie employees to your company at substandard wages or substandard working conditions. It smells of the kind of financial abuses migrant farmworkers or people who sign into iffy marketing schemes are subject to.
I hope you are honest about your policies with new hires before they sign up. In fact, you could save people more time by just letting us know what your company is so that they don't apply.
If there is any significant amount of money involved, I would consider this inappropriate. If the company selects the training program, they pay for it; otherwise, there is too little incentive for them to make a good choice, and it opens up employees to all sorts of possibilities for abuse and fraud. If we are only talking a few thousand dollars, it should make no difference to the company anyway.
If the company trains you and adjust your salary and working conditions appropriately, you shouldn't be any more likely to leave after the training than before. If they don't, that should be their problem, not yours.
But, of course, collecting these fees wasn't enough for the US industry--they keep merrily complaining about piracy and implementing various technical means to deny consumers fair use rights. And the same will probably be happening in Germany.
Why do newspapers just reprint biased estimates from institutions with an axe to grind? This is where journalists should get busy and do a bit of background research. In particular when the numbers are so out of whack.
This is particularly common with AP stories, where no journalist ever signs responsible for it. The AP seems to have turned into a glorified PR newswire.
We have already increased agricultural productivity, yield, and nutritional value many times over over the last few centuries. Initial improvements gave us the free time we needed to build our civilization. But beyond that point, productivity improvements have not stopped hunger, malnutrition, or overpopulation.
At this point, we don't need more productive or more nutritious crops. If we had the political will, we could already feed the current population well with the crops we have. What we really need to work on limiting the size of earth's population through social progress and access to family planning, as well as a more uniform distribution of wealth throughout the world. If we don't accomplish that, huge numbers of people will continue to die from hunger, malnutrition, and war, no matter what genetic gimmicks we invent.
But I find the KDE "mission statement" unacceptable:
To me, free software is about diversity, the ability to try things out, and the ability to customize software to everybody's needs. It's the notion of a "standard" that bothers me about Windows, a "standard" that, in the case of Windows is enforced through closed source releases. But KDE can just as effectively "enforce" standards in the open source world--there are plenty of mechanisms. Open source software is meaningless if it results in the same kind of dull standardization that the closed source world has suffered under.
If it is true that KDE is already on 70% of the Linux desktops, as the KDE League claims, I can only conclude that it is time to stop installing it, promoting it, and recommending it, and to work on something else. And if KDE really succeeds at marginalizing other Linux desktops, I'd just switch to Windows. I really don't see a difference between being forced to use a "standard" system from Microsoft or from KDE.
Fortunately, the UNIX and X11 architecture still makes that possibility fairly remote on the desktop. But on palmtops, Troll Tech really does want to take over with Qt/Embedded, which would not seem to allow non-Qt applications to co-exist on the same screen.
If KDE succeeds at its mission, it will have failed.
Because of the cost to commercial developers of the underlying toolkit, Qt, I think the future of KDE on non-Linux desktops is not too bright. OS vendors have little interest in putting their own customers at the mercy of a third party vendor.
The only way I see that change is if either someone buys Troll Tech outright and re-releases the toolkit for free, or if individual vendors make deals similar to Motif licenses, where the Motif stuff was simply included with the OS, free even for commercial development. But do we really want to go back to those days?
Well, you are making my point: if something isn't easy enough to try, people won't bother. Linux provided enough functionality and advantages over alternatives for early adopters to get to the point where easy-to-install distributions got created and it became useful to many more people. Darwin seems to still have to cross that threshold.
I am saddened if you can't even install a *nix OS without a wonderful graphical install program.
That would be sad, indeed. Fortunately, I can assure you that that's not a problem: I have manually entered boot loaders in machine code and installed 4.1BSD from tape. But, these days, IMO, a reasonably mature system should come with some kind of installer, just like it comes with a lot of other things that systems didn't use to come with.
Are you saying Apple somehow removed the ability to implement additional Mach servers? Does the UNIX personality of Darwin violate the modularity of the Mach server architecture? If that were true, then that would indeed be a strong argument in favor of the Hurd. Or are you saying that the Hurd happens to come with a wider variety of servers right now? If that's the case, probably many of the Hurd servers could be ported over quickly.
Maybe it would be good for someone to write a more detailed technical comparison of the Hurd and Darwin. So far, the bird's eye view is that both system are based on the Mach microkernel, both have a variety of servers for them, and both have a UNIX personality. So, to me, the question remains: why pick one over the other? (Of course, there may still also be licensing issues.)
Putting multimedia data into the file system is the implementation strategy many commercial databases (including some versions of DB2) take behind the scenes for storing multimedia objects, even if they hide it behind a database API. They can still provide all the database facilities (transactions, indexing, access control, etc.) on top of such an implementation.
With that kind of architecture, you don't need a very powerful machine or high performance database to be able to serve image data at disk bandwidth or network bandwidth.
What made me pick up Linux in 1994 was the fact that it came as a complete distribution, pretty much ready to install and run. Those distributions had a number of limitations when it came to drivers and tools, but they were usable and could be used to solve specific problems.
I haven't been able to find a complete distribution for the PC based on the Darwin kernel. Such a distribution would require the kernel, the command line utilities, development tools, X11, and at least one desktop (Gnome, KDE, GNUStep, ...). Such a distribution would be useful even if the set of available drivers is pretty limited (IDE, maybe a SCSI card, a couple of common Ethernet cards).
Darwin is based on the same technology as the Hurd, and it has seen more maintenance over the years. Perhaps it would also offer better compatibility with a commercial OS (MacOS X). So, why not use the Darwin kernel instead of the Hurd?
As usual, Microsoft is late to the party and comes with their own agenda. Microsoft products are oriented towards small business and desktop applications. That's what their evolution is driven by and that's what they are designed for. Whether this kind of data should be in a relational database is questionable to begin with. And it certainly doesn't need to be on an expensive, proprietary operating system and in a proprietary format.
Scientists already have excellent open-source tools to build long-term, stable, large-scale data collections. They would be foolish to tie research projects that can span decades to the fortunes of a company in the middle of a battle for the US business computing market, merely to gain some trinkets and give that company a publicity boost.