MS is only one of the firms and, perhaps, not the most involved. The most extensive report I can find comes from the Canadian International Center for Human Rights & Democracy, and names Siemens, Motorola, Cisco Systems, Sun Microsystems, and Nortel Networks. I found two Amnesty reports, here and here.
Cell service is way overpriced, and probably won't work inside most of the campus buildings, anyway. Voice traffic, with predictable levels, predictable average connection length, and limited bandwidth, is scarcely noticeable in a well-designed wireless system.
Turing, von Neumann, and Wiener were influenced by biological concepts; the connections between what used to be called "cybernetics" and biology go back to the very beginnings of modern computer science. Wiener in particular used the phrase "cybernetic organism", later shorted to "cyborg" in, if I remember correctly, the 1930s.
The analysis of biological and ecological systems as information systems is becoming a standard technique.
It is difficult, at this point, to assess the potential for the combination of electronic and biological systems. A great deal depends on two rather abtruse points:
Does quantum computing invalidate the Church-Turing hypothesis?
Are biological systems capable of quantum computing?
If the answer to either of these questions is no, then current computer science will become the core of biological theory. "Yes" to either dramatically changes the relationships of the subjects; the mathematics of quantum computing bears little resemblance to the finite discrete mathematics now studied as the theoretical core of computer science.
Technological factors are social factors.
Who decides what is right or wrong? There is much reason to take care with new technologies in this area; the popular fear of the creation of monstrosities, I think, has some foundation. But what would make such a thing "inherently wrong"?
Y'know, there's nothing to defend such suits, or such cases. That seems to have been what SCO is banking on; if no defense is made, SCO wins by default. Of course, maybe it would have been better if they hadn't sued IBM.:-)
But this is a problem that's going to come back for a sequel, I think.
To begin with--I used to do customer service--most problems aren't reported. On top of which, a firm that talks about such things in the press is going to have problems with MS come contract renewal time. And any hope of a settlement from MS depends on not talking publicly about the problem.
Actually, MS has, I believe, settled many claims for this kind of problem. However, part of such settlement agreement is usually non-disclosure (a common practice in such matters--MS didn't invent it.)
The settlements, however, make no difference to the employees whose work has been compromised by the problems. When one considers that line employees are usually evaluated on their output, and that MS doesn't compensate them for lost time or sheer frustration, management's ability to get settlements from MS makes much less difference.
You can read their conclusions.
Even if their figures are wrong for 1,000 years ago--and they never were very certain--there has still been a dramatic rise in the past century and their modelling results are about as sure as such things can be.
I would like to see the IPCC Third Assessment link included with all future articles on global climate change.
Ask him for his opinion of open source and free software, and about its place in the future. I believe that, yes, he knows the what those phrases mean. I hope he has some interesting things to say.
Hey, editors! If you're listening, maybe you could do questions for Al Gore!
In the National Opinion Research Center count of ballots, there were more than 113,000 ballots cast for two or more presidential candidates. Of those, 75,000 chose Mr. Gore and a minor candidate; 29,000 chose Mr. Bush and a minor candidate.
Had the balloting been properly run there's little doubt that Gore would have had a clear, though slim, victory. (NORC is a social science research unit at the University of Chicago; they are world-class experts in sampling and polling.)
Back when Al Gore was in college, he used to do stand up comedy. As his recent showing on Saturday Night Live (warning: annoying site) shows, he still can. I don't know why he came across so poorly during the election.
He's been pro-technology, pro-space, and has known from computers for a long time.
Since he's on the Apple board, he is apparently an Apple stockholder. Corporate directors don't usually do that much, but the position does provide for the making of and using of connections.
I would have preferred to see testing across a group of users, and perhaps some actual measurements as well. If design is to be more than satisfying one's own prejudices it must rely on user testing. Notably, emacs, Unix, and MacOS classic are all results of design efforts that involved extensive user testing, with MacOS the most formal of the three.
The author did pick up on a MacOS characteristic that I have not seen widely discussed and is likely to influence most user experience: the slowness of immediate feedback. Good on her. On the other hand, I am struck that the author does not recognize the visual precedent of the default XP theme, which appears to be plastic children's toys.
As to achieving a productive and pleasant GUI user experience on Linux... Knowlegeable people who would never in a million years attempt design of an operating system internal without careful thought and study seem to be convinced that they can dream up a GUI without either. If one is convinced there is no commonality in UI experience--that it is all a matter of taste--then why not the designer's taste? In practice, though, there are commonalities in user experience. I believe it is important, here, to pay attention to the ancient distinction between architecture and building; if it's architecture worth living in, it is built with attention to the people who live in it, not just the designer and builders.
My impression is the rationale for the proliferation of standards in this area is that various competing vendors want to own them. Can some knowledgeable person confirm that?
I think there'd be considerable technical and cost advantages in settling on one open standard.
As someone who's gone from developer to user, with my time spent on something very different, I see two problems here:
First, studying software takes time. A lot of it. I've just lost an hour of time to a software config problem that I could have spent on design. Computer professionals enjoy studying software and get paid to do a lot of it; most people don't and don't.
Computers make us all feel stupid--including us sometime pros. Professionals tolerate the feeling better, is all, I think. (And this may be partly because so many of us have spent a lot of childhood feeling really dumb.)
For solutions to these problems, and there has been some good stuff done, I refer all concerned to the Marc Weiser articles on ubiquitous computing, and Tognazzini's Tog on Interface. What satisfied me as a professional does not now satisfy me as a user; the needs of the groups are different.
On the other hand, they do a bunch of grunt-work to keep the code going. Open Source, if not free software, has benefited from these arrangements. People work on Linux partly because it's an interesting thing to work on from the CS viewpoint. That doesn't seem to happen with office packages. So...it looks a bit rough, but maybe it's a better deal than what we've had so far.
Or (amortized over 20 years), perhaps $4/month. Telecomm companies do this all the time; the amortization used to be 40 years.
The prices sound to me fairly good for telecomm equipment; I think you're just surprised because you've never seen them before; when I consider that underground cable cost about $5/foot (fully allocated) in the 1980s, it's cheap.
Geek commune, nothing. Try, poor neighborhood or third-world city. Use IP voice terminals (telephones) as well as data terminals. Data and voice for a fraction of the cost of wiring the place--I think they've got a winner here.
I dunno, though. The UN is a representative organization with no popular vote and many undemocratic states as members. On top of which, the ITU is not an operations organization at all--it runs standards bodies, things like that. I'd be very hesitant to hand it control of such a crucial thing.
Perhaps the USA could operate the central DNS on behalf of the ITU?
As someone who does lighting simulation work, anything that would make my simulations run faster would be welcome. If it could shave even 20% off a Radiance simulation run, it would be worth a fair bit.
More generally, I see numerous applications in research, provided that the compiler and its use on clusters is reasonably-priced.
It seems to me that we are afraid of collective political action. I think that--with a few notable exceptions like Linus Torvalds--we are intolerant of the human uncertainties of political activism. It is not like software design, where one can be a little god; effective democratic political activism demands persuasion of people and embracing their wants and needs.
So I suppose the free software/open source community would have to embrace very different ideals of society if it was going to be effective in the wider democratic politics of our world.
It's a loose translation of a Hindu tenet, "tat tvam asi," Thou art That, apparently from the Chandogya Upannishad, VI, 8, 7. The whole thing is part of an extended discourse on the unity of the spirit; it is roughly equivalent to the christian saying "The kingdom of god is within you" but there is much more material in the on-line copy of the Upannishad I found.
A little known and excellent set of free multimedia and confrerencing tools have been quietly developed for the MBONE. They are intended, naturally, for use in multi-cast environments, but most will work in point-to-point environments as well. Take a look at the University College of London Networked Multimedia Research Group web pages for details and a software archive.
I once took a course with Dr. Alfred Aho, one of the authors of The Dragon Book and a contributor to lex, yacc and awk (he's the 'a' in awk.) Dr. Aho would never call a compiler component an "optimizer;" instead, he called it a "transmogrifier," because the only way a compiler could produce truly optimal code would be if it could write an optimal program--and it can't.
While the compiler has the opportunity to rewrite the code every time it processes it, it's not very good at rewriting code, so its abilities are limited.
A well-designed compiler on a well-designed CPU can produce very good assembler code--I doubt I could outsmart a compiler in a RISC assembler optimization problem--but I believe that code tweaking is usually of value only in inner loops and for certain kinds of data storage.
All of which adds up to the reality that, pending the arrival of artifical intelligence, human programmers are still needed.
Let me suggest that some of the confusion about the nature of "mainframes" can be cleared up by thinking of mainframes as giant servers designed for i/o intensive operations like, for instance, printing a whole month's telephone bills for several states.
Linux in a VM/390 environment is very promising, and would be even more promising if someone did a variant of XML designed to control 3270-family terminals which, if one squints a little, are not all that different from character-only browsers. As a native OS on the S/390 family, I've got my doubts; Linux doesn't currently support multi-processors well, and I don't know how it cope with the extremely fast parallel network that connects the 390 mass storage system. Still, IBM has people who've been working with those technologies for decades, so perhaps they know how to make it work.
Slashdot Mirror
"I'm no supporter of Bush but Kerry is the greater of the two evils."
A Kerry administration will not include Ashcroft, Rumsfeld, or Cheney. (If you are an environmentalist, add Gale Norton.)
I'll take it.
Why the hostility to Amnesty International?
Cell service is way overpriced, and probably won't work inside most of the campus buildings, anyway. Voice traffic, with predictable levels, predictable average connection length, and limited bandwidth, is scarcely noticeable in a well-designed wireless system.
- Does quantum computing invalidate the Church-Turing hypothesis?
- Are biological systems capable of quantum computing?
If the answer to either of these questions is no, then current computer science will become the core of biological theory. "Yes" to either dramatically changes the relationships of the subjects; the mathematics of quantum computing bears little resemblance to the finite discrete mathematics now studied as the theoretical core of computer science.Y'know, there's nothing to defend such suits, or such cases. That seems to have been what SCO is banking on; if no defense is made, SCO wins by default. Of course, maybe it would have been better if they hadn't sued IBM. :-)
But this is a problem that's going to come back for a sequel, I think.
To begin with--I used to do customer service--most problems aren't reported. On top of which, a firm that talks about such things in the press is going to have problems with MS come contract renewal time. And any hope of a settlement from MS depends on not talking publicly about the problem.
The settlements, however, make no difference to the employees whose work has been compromised by the problems. When one considers that line employees are usually evaluated on their output, and that MS doesn't compensate them for lost time or sheer frustration, management's ability to get settlements from MS makes much less difference.
I would like to see the IPCC Third Assessment link included with all future articles on global climate change.
US Center for Disease Control
World Health Organization (United Nations)
Hmmm, guess governments do something useful after all.
Hey, editors! If you're listening, maybe you could do questions for Al Gore!
The author did pick up on a MacOS characteristic that I have not seen widely discussed and is likely to influence most user experience: the slowness of immediate feedback. Good on her. On the other hand, I am struck that the author does not recognize the visual precedent of the default XP theme, which appears to be plastic children's toys.
As to achieving a productive and pleasant GUI user experience on Linux... Knowlegeable people who would never in a million years attempt design of an operating system internal without careful thought and study seem to be convinced that they can dream up a GUI without either. If one is convinced there is no commonality in UI experience--that it is all a matter of taste--then why not the designer's taste? In practice, though, there are commonalities in user experience. I believe it is important, here, to pay attention to the ancient distinction between architecture and building; if it's architecture worth living in, it is built with attention to the people who live in it, not just the designer and builders.
My impression is the rationale for the proliferation of standards in this area is that various competing vendors want to own them. Can some knowledgeable person confirm that?
I think there'd be considerable technical and cost advantages in settling on one open standard.
-
First, studying software takes time. A lot of it. I've just lost an hour of time to a software config problem that I could have spent on design. Computer professionals enjoy studying software and get paid to do a lot of it; most people don't and don't.
-
Computers make us all feel stupid--including us sometime pros. Professionals tolerate the feeling better, is all, I think. (And this may be partly because so many of us have spent a lot of childhood feeling really dumb.)
For solutions to these problems, and there has been some good stuff done, I refer all concerned to the Marc Weiser articles on ubiquitous computing, and Tognazzini's Tog on Interface. What satisfied me as a professional does not now satisfy me as a user; the needs of the groups are different.On the other hand, they do a bunch of grunt-work to keep the code going. Open Source, if not free software, has benefited from these arrangements. People work on Linux partly because it's an interesting thing to work on from the CS viewpoint. That doesn't seem to happen with office packages. So...it looks a bit rough, but maybe it's a better deal than what we've had so far.
Geek commune, nothing. Try, poor neighborhood or third-world city. Use IP voice terminals (telephones) as well as data terminals. Data and voice for a fraction of the cost of wiring the place--I think they've got a winner here.
I dunno, though. The UN is a representative organization with no popular vote and many undemocratic states as members. On top of which, the ITU is not an operations organization at all--it runs standards bodies, things like that. I'd be very hesitant to hand it control of such a crucial thing. Perhaps the USA could operate the central DNS on behalf of the ITU?
There is extensive coverage in Computerworld, here.
As someone who does lighting simulation work, anything that would make my simulations run faster would be welcome. If it could shave even 20% off a Radiance simulation run, it would be worth a fair bit.
More generally, I see numerous applications in research, provided that the compiler and its use on clusters is reasonably-priced.
It seems to me that we are afraid of collective political action. I think that--with a few notable exceptions like Linus Torvalds--we are intolerant of the human uncertainties of political activism. It is not like software design, where one can be a little god; effective democratic political activism demands persuasion of people and embracing their wants and needs. So I suppose the free software/open source community would have to embrace very different ideals of society if it was going to be effective in the wider democratic politics of our world.
It's a loose translation of a Hindu tenet, "tat tvam asi," Thou art That, apparently from the Chandogya Upannishad, VI, 8, 7. The whole thing is part of an extended discourse on the unity of the spirit; it is roughly equivalent to the christian saying "The kingdom of god is within you" but there is much more material in the on-line copy of the Upannishad I found.
A little known and excellent set of free multimedia and confrerencing tools have been quietly developed for the MBONE. They are intended, naturally, for use in multi-cast environments, but most will work in point-to-point environments as well. Take a look at the University College of London Networked Multimedia Research Group web pages for details and a software archive.
I once took a course with Dr. Alfred Aho, one of the authors of The Dragon Book and a contributor to lex, yacc and awk (he's the 'a' in awk.) Dr. Aho would never call a compiler component an "optimizer;" instead, he called it a "transmogrifier," because the only way a compiler could produce truly optimal code would be if it could write an optimal program--and it can't.
While the compiler has the opportunity to rewrite the code every time it processes it, it's not very good at rewriting code, so its abilities are limited.
A well-designed compiler on a well-designed CPU can produce very good assembler code--I doubt I could outsmart a compiler in a RISC assembler optimization problem--but I believe that code tweaking is usually of value only in inner loops and for certain kinds of data storage.
All of which adds up to the reality that, pending the arrival of artifical intelligence, human programmers are still needed.
"I do think [employers] should help pay for ergonomic furniture and an appropriate working environment."--Slashdot
"An employer is responsible for ensuring that its employees have a safe and healthful workplace, not a safe and healthful home,"--OSHA
Looks like OSHA's done something Slashdot agrees with. Oh the humanity!
An LA Times article giving views both pro- and con- can be read here.
Let me suggest that some of the confusion about the nature of "mainframes" can be cleared up by thinking of mainframes as giant servers designed for i/o intensive operations like, for instance, printing a whole month's telephone bills for several states.
Linux in a VM/390 environment is very promising, and would be even more promising if someone did a variant of XML designed to control 3270-family terminals which, if one squints a little, are not all that different from character-only browsers. As a native OS on the S/390 family, I've got my doubts; Linux doesn't currently support multi-processors well, and I don't know how it cope with the extremely fast parallel network that connects the 390 mass storage system. Still, IBM has people who've been working with those technologies for decades, so perhaps they know how to make it work.