And as of this weekend The Amber Spyglass is now out in paperback (in the UK anyway).
I can't tell you if it's any good -- it's sitting on the table in front of me at the moment, next up to be read -- but the first two were fascinating. Strongly recommended.
The Economist had a good
article back in May about the state of the art (7 qubits), and some of the practical difficilties facing quantum computing.
"Dr Cory says that the program for factorising large numbers [400 digits] will require about 1,000 qubits simply to store the problem. But each of these qubits will require dozens of extra qubits for error-checking. That means that a useful computer will need tens of thousands of qubits to come up with a reliable answer. At seven and counting, that goal is a long way off. "
Strictly speaking, data mining is not statistics. Statistics deals with analysis of time-based data, typically small amount of parameters is sampled many times. Data mining studies non or weakly time-based data, with lots of parameters but just a few samples taken.
Examples: taking measurements of temperature in a region over 50 years and trying to predict climate change is statistical problem, while analysing samples of minerals in an area to try to find oil or gas is data mining. (as, presumably, mineral composition does not change over time so only single sample from each point is taken)
Nope. Disagree.
If you want a unifying theme to pull together most of AI, you could do worse than think of it as a cookbook of techniques for designing, building and then automatically refining statistical models of (some aspect/s of) reality.
Pretty much all of the standard problems of AI can be presented to advantage in this framework (including the identification of new objects/concepts/patterns and associations). It is the evolutionary enhancement that our brain has conferred by automatically doing adaptive and hierarchical statistical modelling (incl. pattern recognition, but also much more) so well that is basically what makes 'intelligence' worth having.
The important thing about such a statistical model is *not* certainty that the model is right -- it never will be (not even for your temperature data). The 'long-run densely-sampled stationary time-series' is a myth -- in reality it just doesn't happen.
The important thing *is* to realise your model is imperfect; to allow (as well as you can) for that uncertainty in your model; to explore different ways of setting up your model; and then to use statistical inference (Bayes theorem etc) to improve your predictive model as the data comes in. Only by allowing for the imperfection can you learn from the data.
The priciples of statistical modelling have a central place in the study of Statistics -- it's the underlying logic that most of the subject is built on. On the other hand the blind application of certain 'standard' statistical tests seems distictly peripheral.
Yeah, that industry has been where he's rediscovered things we had in the 80s just to announce them as "new". I'm amazed that IJCAI didn't boo him off the stage but then again, it's a new Microsoft influenced generation from the AI heyday of the 80s
One big step since the '80s is the advance in Bayesian networks, and MCMC methods for training them.
That, plus the increase in computing power, makes it much more possible to deal realistically with uncertainty and small training sets; it's also now possible (and worthwhile) to embed the systems in end-user applications.
Slightly (OT), but does anyone know if it is still possible to buy a K6-II+ or K6-III anywhere for an ordinary end-user ? (Just one sought, not several thousand).
This looked to be the one upgrade for an old 66 Mhz series 7 motherboard that would have made sense, avoiding the slow main memory with a 400 Mhz on-die L2 cache. But the chips seemed to disappear almost as soon as AMD announced them.
Does anybody know of anywhere where they can still be found ?
(I believe the embedded chips have the same pinouts as the original and the mobile versions, but I could be wrong).
In case anybody else has been wondering about this apparently virulent American plague, that's never been heard of on this side of the Atlantic, Mononucleosis translates into English as Glandular Fever (Epstein-Barr virus).
There was a book about ten years ago, called Brainsex (Amazon UK), by Anne Moir and David Jessel, a couple of quite respectable TV journalists. (There might have been a related TV series ?)
I've never read it, though I think I did see some magazine extracts. The reader's review on Amazon found it interesting, but was dubious about how much weight the experimental evidence deserved.
It's an interesting prospect, if instead of being an effect of the physically rather dubious "dark energy" (vacuum energy density), or the even more dubious "vacuum state tunneling" of inflation, the apparent flatness of space at large scales might be due to plain old magnetic fields.
Somebody had to do this. So, it looks as if it's Ximian who've got their claim staked first.
Well, I don't think they'll starve.
Having a rival to.NET is too important for the rest of the industry to ignore. I imagine Ximian won't find it too tough when they go looking for campaign contributions.
This graph
is one of the scariest things I have seen in a long time.
It's a plot of the temperature variations and CO2 levels over the last 500,000 years measured from ice cores drilled out from Lake Vostok in the Antarctic.
The two series track each other incredibly closely.
As we now have good models for why CO2 should cause temperature change, but not the other way round, it is something to take very seriously.
The figure was taken from The Economist magazine, a paper not usually associated with extreme anti-business views.
Two recent articles gave good summaries of our present state of knowledge about global warming, and how both the data and the models have improved over the last ten years:
The Reality of Global Warming (14 June 2001)
Article on the report for the Bush administration from the National Academy of Sciences.
(Titles given are those used in the magazine's
index of its environmental stories online.)
One worrying new possibility is that there may be an abrupt change (bifurcation) in the ecosystem response as the temperature rises. At the moment about 50% of the manmade CO2 emissions are being absorbed by the Amazon rain forest. But the latest Hadley Centre models predict that if the temperature continues to rise, this greatly increases the frequency of much drier weather in this region, causing the forest to dry out, ultimately leading to uncontrollable forest fires. This would release vast amounts of more CO2 into the atmosphere if the whole lot went up -- perhaps ten times as much as human activities.
(And that is not the ultimate nightmare positive-feedback scenario, which is the enormous amounts of methane hydrate locked up at the bottom of the ocean in the arctic permafrost. The only thing that keeps it stable is the high pressure and low temperature. There is thought to have been a runaway destabilisation 55 million years ago, which raised the temperature 15 degrees C in less than 20 years).
I suppose somebody might come up with a techno-fix solution. But the complacency of gambling on that is like playing Russian roulette with five of the six chambers loaded.
A short paper by Pam Samuelson of Berkeley (1999) does an excellent job of underlining the significance of the Statute of Anne, as affirmed by the Founding Fathers, and contrasting it with the various worrying current developments in copyright law.
It's along the same lines as the MSNBC piece, but the history and the analysis are much sharper.
This is what she has to say specifically about the 1710 Statute
The development that ushered in the modern era of copyright was the
English Parliaments passage of the Statute of Anne in 1710 [18]. On
its face, this statute was not only a repudiation of several principal
tenets of the stationers copyright system; it was also a redirection
of copyright's purpose away from censorship and toward freedom of
expression principles and an effort to promote real competition among
printers and booksellersthat is, to break the stranglehold that major
players in the Stationers' Company had over the book trade. Insofar
as that monopoly continued in revised form, the statute provided
recourse for those injured by excessive prices of books.
The key aspects of the Statute of Anne for achieving these goals were
these: First, the act granted rights to authors, not to publishers.
Second, it did so for the utilitarian purpose of inducing learned
men to write and publish books. Third, the act established a larger
societal purpose for copyright, namely, to promote learning. Fourth,
it granted rights only in newly authored books. Thereafter, ancient
books were in the public domain and could be printed by anyone.
Fifth, it limited the duration of copyright to fourteen year terms
(renewable for another fourteen years if the author was living at the
end of that term), thus abolishing perpetual copyrights [19]. Sixth,
the statute conferred rights of a limited character (not to control
all uses, but to control the printing and reprinting of protected
works). Seventh, it imposed a responsibility on publishers to deposit
copies of their works with designated libraries. Eighth, it provided
a system for redressing grievances about overpriced books.
While it took about fifty additional years for pre-modern system to
die out [20], the modern law of copyright emerged from the Statute
of Anne's precepts. Censorship held no place of honor in this new
copyright system which, in the main, embraced Enlightenment values
that also influenced the framers of the U.S. Constitution. The clause
of this constitution that empowers Congress to promote the progress
of science and the useful arts by securing to authors and inventors
an exclusive right in their respective writings and discoveries for
limited times should be viewed in historical context as an American
endorsement of England's repudiation of the speech-suppressing,
anti-competitive and otherwise repressive pre-modern copyright system
that the English Parliament meant to reshape by the Statute of Anne.
Core elements of the Statute of Anne are reflected in Article I,
sec. 8, cl. 8's purpose ("to promote Science"), in the persons to whom
rights were to be granted ("authors"), and in the duration of rights
("for limited times").
... The constitutional copyright
clause, properly construed, embodies first amendment and anti-monopoly
principles.... There is a "dormant copyright clause" waiting to be reawakened in the
case law -- and hopefully in Congress -- after a long sleep in which
the clause has become a meaningless cliche [22].
The rest of the paper, which analyses the contrast between this history and current recent developments, is strongly recommended.
The first 'modern' copyright law was the Statute of Queen Anne (1710).
This created the legal status quo in the Colonies which continued largely unchanged by the US Founding Fathers.
From Britannica:
The Statute of Anne, passed in England in 1710, was a milestone in the history of copyright law. It recognized that authors should be the primary beneficiaries of copyright law; it also established the idea that such laws should have only limited duration, after which works could pass into public domain. The designated period came to be set at 28 years. Similar laws were enacted in Denmark (1741), the United States (1790), and France (1793).
The BSD version is deliberately crippleware, and intended to stay that way.
It is missing many of the key libraries that would most applications would use; and a lot of the internals are intentionally low performance.
It appears to be a bare minimum for (a).NET to be teachable on university CS courses; and (b) MS to be able to point to 'two independent implementations' for standards approval -- which it needs to reassure CIO's that the platform won't get hijacked from underneath them.
We haven't seen the licence yet, of course; but I would imagine that any patches would belong to MS; independent distribution of patched code would be forbidden; and that MS would have the right to suppress any patches which altered the functionality of the product.
This is one product I think MS have no desire to see working any better.
The idea of.NET is to create an environment where it is as easy to click-and-paste together server apps as it is to create standalone scripts at the moment in VB -- easier, if anything.
It isn't meant as just an API for invoking applications -- it is meant as a complete API for creating them. With its support libraries, we are talking about something as all encompassing as win16 and win32 MFC were in their time, but of course now much much wider.
(Do platform capabilities have to grow with Moore's Law too ?)
The idea is that -- for example -- you could click together a standalone app to run on your workstation; but transfer it to a server, and all your local forms and GUI are automatically translated to web equivalents.
You can write objects in whichever language you like - VB, C#, perl, python - and not have to worry about bindings or compatibility; you can automatically inherit and extend any of the objects written in any of the languages, either at runtime or at compile time, without even any header files to worry about -- even if they are bought-in objects pre-compiled to bytecode.
The aim is to make it easy to do things at a much much higher level. It is on a completely different scale of developer-friendliness than having to worry about the minutiae of interfacing with Corba, or drawing a page box by box with GTK.
Just plug in, and off you go.
I am not saying that the underlying technology is necessarily particularly advanced -- but to create all of the components to make the whole platform is a huge undertaking.
The big questions are whether MS can deliver it, and (especially) what the performance will be like. But if they do get it right, their aim is to de-skill the whole business of writing everyday garden-variety server applications.
Try looking it up on a mailing list, and all you get is endless postings from aa@bb.cc.net . Or try Google and it ignores it altogether.
This is what annoys me the most about.NET
Otherwise I'm with Xandis (who appears to have got modded down off the face of the earth for daring to suggest MS might have done something right for once, so here's a copy):
It's not that bad. It is easy to remember and pronounce. It can also fit nicely on mugs, t-shirts, flags, etc. Having a DOT along with a NET captures both the business angle (DOT-com) along with the generic angle (interNET). It is so non-specific it meets their need for being the name of a framework that will be the core of everything net-related in the universe. I think Red Hat would be wise to follow suit and produce a competing framework titled DOT HAT.
But the most important thing about.NET (from a marketing point of view) is the scope of what it lays claim to.
Whether it is actually true or not, the umbrella name suggests that MS has a complete integrated package to offer -- everything you need to click-and-paste together apps on your standalone system, which can then be seamlessly transferred to run as server apps without changing a line of code.
It may still be vapourware, but IBM, Sun, Oracle etc don't even have a name yet for such an all encompassing dream.
Another thing to remember about.NET is that whether or not there is an Open Source implementation, the majority of the traffic will be going through Microsoft's servers... At least, for the HailStorm portion of things.
.NET is an API and development environment for writing server-based apps.
Hailstorm is a tiny part of the picture.
The 'traffic' will be going to the vast number of third-party apps, which.NET is designed to make it much, much easier to write.
It isn't the Hailstorm servers you need to worry about -- it's all the servers running those third party apps.
Do you want to see Open Source locked out of what is likely to become the dominant segment of the server market ?
Copying the base platform is the easy part, and isn't going to get you very far... copying the library is probably much trickier
There was a good article recently in Linux Magazine on what the CLR would need
GCC.NET
-- Mark Mitchell, release manager for gcc 3.0 and CTO of CodeSourcery
What is Required for GCC to Support Microsoft's.NET? (April 2001)
A couple of other articles (less focussed) discussed why adopting some of.NET might be a good idea:
Independent State
-- Interview with Dick Hardt of ActiveState.
Pages 3-4 discuss.NET : What it is, what it is like to code for, why linux needs an implementation, what needs to be done. (April 2001)
Embrace and Extend
-- Jon Udell from Byte
What linux can learn from.NET's component architecture. (February 2001)
For all the throwaway remarks Bruce Perens makes, I don't think it is going to be an easy job -- and cloning the libraries will be a massive undertaking. (For comparison, just look how much of the standard Java libraries are still to be implemented by GCJ).
But even incomplete first steps could be very worth while, especially
Mechanisms to inherit and efficiently extend objects from pre-built libraries, using a common cross-language ABI.
C#, which does have a certain amount going for it.
I guess its a step towards weaning Microsofties away from Microsoft servers, but it seems like a lot of effort
I think you are altogether too sanguine.
The aim of the.NET API is to dramatically lower the bar for writing server-based apps, leading to a huge expansion in such platforms. MS's services like Hailstorm are only a tiny part of the picture. Unless the unix community can effectively market an alternative API for creating server applications which is both as friendly for end-users and as easy to develop for, we risk lock-out from what is likely to become the dominant sector of the server market.
I was talking to a friend who works for a major UK retailer (a UK top 50 company).
They are about to massively upgrade the software they use for sharing live sales and warehouse stock-level data with their supply chain. The company reckoned that.NET was exactly the sort of foundation they needed, and considered it at some length last week, before (reluctantly) agreeing that there was no way they could commit to something before it had had a widely evaluated and reviewed 1.0 release.
To amplify what cthugha was writing above: unless linux can implement this platform, or offer an alternative which is both as easy to use and as easy to develop for, MS will lock in a lot of the server marketplace.
APIs can be reproduced, so copyright is essentially useless against reimplementation.
It is true that APIs are not protected against cloning under copyright law; but if they implement a novel and efficient solution to a technical problem, they may indeed have protected it under patent law.
For example, Microsoft has a patent
(US 5,297,284) on the layout of the vtables of pointers to functions used in COM objects with multiple inheritance.
It is therefore legally forbidden to add a compatibility option in gcc to clone this. (Although according to this post on the gcc list the WINE people do have a workaround).
If you think that MS Legal haven't done their level best to protect.NET against independent third party clones, then you are naive. The recent pre-announcement of the crippleware.NET SDK for BSD explicitly mentions that it includes the licensing of relevant patents, as does MS's development agreement with Corel. Independent implementers are unlikely to be so favoured.
Estimate of Lambda -- more info
on
Universe is Flat
·
· Score: 2
By itself Boomerang says very little about the value of the cosmological constant Lambda; but as this diagram shows, with its accompanying caption here (second box from the bottom),
when combined with data from supernovas it suggests that the vacuum energy density ("dark energy")
accounts for about two-thirds of the total mass-energy of the universe.
Such "dark energy" acts as a source for gravitational attraction and lensing, but it also exerts an outward pressure similar to the Caisimir effect, which accelerates the expansion.
Slashdot Mirror
I can't tell you if it's any good -- it's sitting on the table in front of me at the moment, next up to be read -- but the first two were fascinating. Strongly recommended.
It's arguable that a strong faction of the founding fathers of the US left Europe specifically to be able to indulge in religious based persecution.
If you look at the history of (for example) C17 Massachusetts, the early authorities there didn't go a whole bundle on religious tolerance.
"Dr Cory says that the program for factorising large numbers [400 digits] will require about 1,000 qubits simply to store the problem. But each of these qubits will require dozens of extra qubits for error-checking. That means that a useful computer will need tens of thousands of qubits to come up with a reliable answer. At seven and counting, that goal is a long way off. "
unless you know how to change this pref ?
Examples: taking measurements of temperature in a region over 50 years and trying to predict climate change is statistical problem, while analysing samples of minerals in an area to try to find oil or gas is data mining. (as, presumably, mineral composition does not change over time so only single sample from each point is taken)
Nope. Disagree.
If you want a unifying theme to pull together most of AI, you could do worse than think of it as a cookbook of techniques for designing, building and then automatically refining statistical models of (some aspect/s of) reality.
Pretty much all of the standard problems of AI can be presented to advantage in this framework (including the identification of new objects/concepts/patterns and associations). It is the evolutionary enhancement that our brain has conferred by automatically doing adaptive and hierarchical statistical modelling (incl. pattern recognition, but also much more) so well that is basically what makes 'intelligence' worth having.
The important thing about such a statistical model is *not* certainty that the model is right -- it never will be (not even for your temperature data). The 'long-run densely-sampled stationary time-series' is a myth -- in reality it just doesn't happen. The important thing *is* to realise your model is imperfect; to allow (as well as you can) for that uncertainty in your model; to explore different ways of setting up your model; and then to use statistical inference (Bayes theorem etc) to improve your predictive model as the data comes in. Only by allowing for the imperfection can you learn from the data.
The priciples of statistical modelling have a central place in the study of Statistics -- it's the underlying logic that most of the subject is built on. On the other hand the blind application of certain 'standard' statistical tests seems distictly peripheral.
One big step since the '80s is the advance in Bayesian networks, and MCMC methods for training them.
That, plus the increase in computing power, makes it much more possible to deal realistically with uncertainty and small training sets; it's also now possible (and worthwhile) to embed the systems in end-user applications.
This looked to be the one upgrade for an old 66 Mhz series 7 motherboard that would have made sense, avoiding the slow main memory with a 400 Mhz on-die L2 cache. But the chips seemed to disappear almost as soon as AMD announced them.
Does anybody know of anywhere where they can still be found ?
(I believe the embedded chips have the same pinouts as the original and the mobile versions, but I could be wrong).
In case anybody else has been wondering about this apparently virulent American plague, that's never been heard of on this side of the Atlantic, Mononucleosis translates into English as Glandular Fever (Epstein-Barr virus).
I've never read it, though I think I did see some magazine extracts. The reader's review on Amazon found it interesting, but was dubious about how much weight the experimental evidence deserved.
Can anybody else give an opinion on the book ?
It's an interesting prospect, if instead of being an effect of the physically rather dubious "dark energy" (vacuum energy density), or the even more dubious "vacuum state tunneling" of inflation, the apparent flatness of space at large scales might be due to plain old magnetic fields.
Well, I don't think they'll starve.
Having a rival to .NET is too important for the rest of the industry to ignore. I imagine Ximian won't find it too tough when they go looking for campaign contributions.
This graph is one of the scariest things I have seen in a long time. It's a plot of the temperature variations and CO2 levels over the last 500,000 years measured from ice cores drilled out from Lake Vostok in the Antarctic. The two series track each other incredibly closely.
As we now have good models for why CO2 should cause temperature change, but not the other way round, it is something to take very seriously.
The figure was taken from The Economist magazine, a paper not usually associated with extreme anti-business views. Two recent articles gave good summaries of our present state of knowledge about global warming, and how both the data and the models have improved over the last ten years:
- The Reality of Global Warming (14 June 2001)
- The science and politics of global warming (16 Nov 2000)
(Titles given are those used in the magazine's index of its environmental stories online.)Article on the report for the Bush administration from the National Academy of Sciences.
Background piece before the last 'Kyoto' meeting in the Hague
One worrying new possibility is that there may be an abrupt change (bifurcation) in the ecosystem response as the temperature rises. At the moment about 50% of the manmade CO2 emissions are being absorbed by the Amazon rain forest. But the latest Hadley Centre models predict that if the temperature continues to rise, this greatly increases the frequency of much drier weather in this region, causing the forest to dry out, ultimately leading to uncontrollable forest fires. This would release vast amounts of more CO2 into the atmosphere if the whole lot went up -- perhaps ten times as much as human activities.
(And that is not the ultimate nightmare positive-feedback scenario, which is the enormous amounts of methane hydrate locked up at the bottom of the ocean in the arctic permafrost. The only thing that keeps it stable is the high pressure and low temperature. There is thought to have been a runaway destabilisation 55 million years ago, which raised the temperature 15 degrees C in less than 20 years).
I suppose somebody might come up with a techno-fix solution. But the complacency of gambling on that is like playing Russian roulette with five of the six chambers loaded.
It's along the same lines as the MSNBC piece, but the history and the analysis are much sharper.
This is what she has to say specifically about the 1710 Statute
The rest of the paper, which analyses the contrast between this history and current recent developments, is strongly recommended.
This created the legal status quo in the Colonies which continued largely unchanged by the US Founding Fathers.
From Britannica:
For further information see this discussion by Brian Forte.The full text of the Statute is online here
Did you even read the original post ?
It is missing many of the key libraries that would most applications would use; and a lot of the internals are intentionally low performance.
It appears to be a bare minimum for (a) .NET to be teachable on university CS courses; and (b) MS to be able to point to 'two independent implementations' for standards approval -- which it needs to reassure CIO's that the platform won't get hijacked from underneath them.
We haven't seen the licence yet, of course; but I would imagine that any patches would belong to MS; independent distribution of patched code would be forbidden; and that MS would have the right to suppress any patches which altered the functionality of the product.
This is one product I think MS have no desire to see working any better.
It isn't meant as just an API for invoking applications -- it is meant as a complete API for creating them. With its support libraries, we are talking about something as all encompassing as win16 and win32 MFC were in their time, but of course now much much wider. (Do platform capabilities have to grow with Moore's Law too ?)
The idea is that -- for example -- you could click together a standalone app to run on your workstation; but transfer it to a server, and all your local forms and GUI are automatically translated to web equivalents.
You can write objects in whichever language you like - VB, C#, perl, python - and not have to worry about bindings or compatibility; you can automatically inherit and extend any of the objects written in any of the languages, either at runtime or at compile time, without even any header files to worry about -- even if they are bought-in objects pre-compiled to bytecode.
The aim is to make it easy to do things at a much much higher level. It is on a completely different scale of developer-friendliness than having to worry about the minutiae of interfacing with Corba, or drawing a page box by box with GTK. Just plug in, and off you go.
I am not saying that the underlying technology is necessarily particularly advanced -- but to create all of the components to make the whole platform is a huge undertaking.
The big questions are whether MS can deliver it, and (especially) what the performance will be like. But if they do get it right, their aim is to de-skill the whole business of writing everyday garden-variety server applications.
Try looking it up on a mailing list, and all you get is endless postings from aa@bb.cc.net . Or try Google and it ignores it altogether.
This is what annoys me the most about .NET
Otherwise I'm with Xandis (who appears to have got modded down off the face of the earth for daring to suggest MS might have done something right for once, so here's a copy):
But the most important thing about .NET (from a marketing point of view) is the scope of what it lays claim to.
Whether it is actually true or not, the umbrella name suggests that MS has a complete integrated package to offer -- everything you need to click-and-paste together apps on your standalone system, which can then be seamlessly transferred to run as server apps without changing a line of code.
It may still be vapourware, but IBM, Sun, Oracle etc don't even have a name yet for such an all encompassing dream.
-no text-
Hailstorm is a tiny part of the picture.
The 'traffic' will be going to the vast number of third-party apps, which .NET is designed to make it much, much easier to write.
It isn't the Hailstorm servers you need to worry about -- it's all the servers running those third party apps.
Do you want to see Open Source locked out of what is likely to become the dominant segment of the server market ?
There was a good article recently in Linux Magazine on what the CLR would need
- GCC.NET
.NET? (April 2001)
A couple of other articles (less focussed) discussed why adopting some of-- Mark Mitchell, release manager for gcc 3.0 and CTO of CodeSourcery
What is Required for GCC to Support Microsoft's
-- Interview with Dick Hardt of ActiveState.
Pages 3-4 discuss
-- Jon Udell from Byte
What linux can learn from
For all the throwaway remarks Bruce Perens makes, I don't think it is going to be an easy job -- and cloning the libraries will be a massive undertaking. (For comparison, just look how much of the standard Java libraries are still to be implemented by GCJ).
But even incomplete first steps could be very worth while, especially
I guess its a step towards weaning Microsofties away from Microsoft servers, but it seems like a lot of effort
I think you are altogether too sanguine.
The aim of the .NET API is to dramatically lower the bar for writing server-based apps, leading to a huge expansion in such platforms. MS's services like Hailstorm are only a tiny part of the picture. Unless the unix community can effectively market an alternative API for creating server applications which is both as friendly for end-users and as easy to develop for, we risk lock-out from what is likely to become the dominant sector of the server market.
They are about to massively upgrade the software they use for sharing live sales and warehouse stock-level data with their supply chain. The company reckoned that .NET was exactly the sort of foundation they needed, and considered it at some length last week, before (reluctantly) agreeing that there was no way they could commit to something before it had had a widely evaluated and reviewed 1.0 release.
To amplify what cthugha was writing above: unless linux can implement this platform, or offer an alternative which is both as easy to use and as easy to develop for, MS will lock in a lot of the server marketplace.
(no text)
APIs can be reproduced, so copyright is essentially useless against reimplementation.
It is true that APIs are not protected against cloning under copyright law; but if they implement a novel and efficient solution to a technical problem, they may indeed have protected it under patent law.
For example, Microsoft has a patent (US 5,297,284) on the layout of the vtables of pointers to functions used in COM objects with multiple inheritance. It is therefore legally forbidden to add a compatibility option in gcc to clone this. (Although according to this post on the gcc list the WINE people do have a workaround).
If you think that MS Legal haven't done their level best to protect .NET against independent third party clones, then you are naive. The recent pre-announcement of the crippleware .NET SDK for BSD explicitly mentions that it includes the licensing of relevant patents, as does MS's development agreement with Corel. Independent implementers are unlikely to be so favoured.
Such "dark energy" acts as a source for gravitational attraction and lensing, but it also exerts an outward pressure similar to the Caisimir effect, which accelerates the expansion.