Journalling file systems usually don't do journalling for data, only for file system structure. Furthermore, the databases themselves do the journalling and recovery for the data.
Using a journalling file system for the partition holding database tables will accomplish pretty much nothing (but it won't do much harm either).
Journalling file system in general don't help much with data integrity either; the kinds of guarantees they usually make are minimal, they don't protect you against many sources of data loss, and they don't even guarantee fast system startups.
I think that's a good deal for someone who has a specific need for Solaris or SPARC. I don't think it will have much effect on Linux, however.
The biggest thing SPARC-based systems have going for them in my mind is that they run a good version of Java on a UNIX platform. The other really good implementation of Java is for NT.
Java on other UNIX platforms (Linux, FreeBSD, etc.) is either not the latest version (only up to 1.1) or still not entirely reliable (1.2 on Linux, etc.).
Solution: Use the MMU to isolate software components from each other and to make their internal structure entirely hidden, leaving only their interfaces visible (an OO approach is implied).
Using the MMU for this purpose is a poor engineering tradeoff.
C and C++ are exceptional among programming languages in terms of the poor fault isolation they provide. Most languages do better. Some, like Java, do very well.
However, while lack of fault isolation is one of the biggest problems in C/C++ software, even with excellent fault isolation, there are still plenty of bugs possible.
The claims are worthless without descriptions of the experimental procedures, peer review, and replication. There are already many ways in which pattern recognition systems and neural networks can greatly outperform humans, even in the presence of noise; that says nothing about whether it is a practical advance or not.
While the press release doesn't say much about neural networks or whether the state of the art in speech recognition has improved, it tells us something about a disregard by USC for standards of scientific conduct: scientific publication by press release is improper.
Aluminum used to be as expensive as gold, and soda wasn't cheap either. Making them the inexpensive mass-market products they are now required some serious innovation and deep thinking.
Likewise, we all want our technical toys to be smaller, so "pioneers" building very expensive wearable computers don't impress me that much--they are doing what is natural. But a company that figures out how to produce them competitively and at a price I can afford, that's the real innovation to me.
First you assume that something must have happened, and then you criticize Sun for your assumption.
FWIW, even if Sun gave them a good discount, that's just fine with me. It's a competitive market, and Microsoft has been "donating" (and probably tax deducting) hardware and software to schools in huge amounts. Until that practice is declared anticompetitive and prohibited by the FTC, Sun has to play along. If they don't already do it, I think they should start special pricing and donations to schools ASAP.
It's nice that Apple is actually trying to go beyond square PC boxes running Windows.
But while an alternative is good, I personally don't like this particular alternative: I don't like the look, I don't like the pointing device or keyboard, and I don't like the software. Imagine Apple would take over the market and all computers would come in semi-transparent rounded lumps with trim in various colors. I'd find it pretty awful.
The notion of basing security on some piece of hardware associated with the CPU or motherboard is intrinsically flawed. I don't want my machine to authenticate as me: the identity of my machine doesn't matter. The machine may get sold or stolen or used by someone else.
Hardware based authentication and security tokens should be based on something portable, and that portable needs to have enough compute power to implement something like zero knowledge proofs. SmartCards fit the bill, and they are cheap. Keyboards should have SmartCard readers, and standard cryptographic methods allow secure transactions to be executed with SmartCards even over untrusted machines.
At best, the computer itself could benefit from hardware encryption that doesn't carry a key, in order to speed up throughput for encrypted data streams. But in the current political climate, putting hardware-based encryption into a PC is futile, since, according to US laws, it cannot be secure anyway.
Of course, e-commerce companies don't like SmartCards because, oh my, the consumer can remove them when they don't want to buy anything and don't want to get tracked. ID chips tied to the CPU or motherboard are great: the kids can order, the software can be used to track people wherever they go, and there is little most people can do about it if they run standard software like Windows.
If IBM wants to drive secure e-commerce, they should be shipping computers with SmartCard enabled keyboards.
It's pretty hard to beat the convenience of paper. It's not only that it's lightweight, high contrast, and durable. Paper is also cheap to produce, cheap to print on, and easy to dispose of. It works without batteries, and it works everywhere. It can't break, and it works in a huge range of temperatures and environmental conditions (much wider than most portable electronics). There is very little cost associated with giving a piece of paper to someone else. And it's easy to understand how to destroy the human readable information on paper (fingerprints, digital watermarks, DNA, etc., of course, is much more difficult to destroy).
Even if you met the convenience and durability of paper with an electronic product, you still have to face the problem that the electronic product is expensive. Would you take a $1000 rollup electronic newspaper with you everywhere?
Paper can be produced more cheaply and more environmentally friendly from renewable resources. One of the primary ones would be (drug free) hemp.
It may be surprising to some, but Sun is actually a business and they have to make a profit. As a business, they have been pushing for some pretty good technologies, they have documented their interfaces well, stuck to standards, and released a lot of code into the market. They have done so because it serves their business interests, not because they are philanthropic.
Compare that with Microsoft. Microsoft resists any careful documentation or standardization of their APIs, Microsoft constantly strives to replace non-proprietary APIs with proprietary APIs, and Microsoft's software (in particular, their APIs) is a mess.
As for StarOffice, if an X11 and UNIX-based office suite catches on widely in corporations, that's good for Linux, all things being equal. The server-based nature of their system is a boon to businesses, something Microsoft can't match right now. If the whole thing runs on Linux, that's nice, but if it doesn't, no harm done.
Sun is no "friend" of Linux beyond what serves their business needs. And that is as it should be. As long as they keep doing what they are doing, Linux and Sun have similar goals and they deserve our support.
The patent office isn't a private company that you could sue. In any case, if you did, you'd be sueing yourself, since their costs largely come out of your pocket as taxes.
Maybe the patent office could be privatized and both the patent office and patent applicants could be found liable and assessed penalties if patents are overturned. The problem right now is that the risk and cost of applying for a frivolous patent is tiny compared to the benefit of getting away with it occasionally.
It might be possible that a lawyer could argue that a company that has obtained an invalid patent has greatly harmed its competitors and should pay fines and penalties. But in the current environment, it's already so hard even to establish that patents with obvious prior art are invalid that I wouldn't hold my breath for that to happen.
Sure, it would be nice if IBM got the modem working (in fact, it would be even nicer if they just got rid of the MWAVE stuff altogether), but I don't think this is a big deal.
So, think of the 600E as a Thinkpad without a modem when you are running Linux. You aren't paying a lot extra for the modem, and the built-in Thinkpad modem is a kludge anyway.
If they do it correctly, a card with a chip embedded can be a lot more secure than either a manual password or even a SecurId card.
Using zero knowledge proofs and bidirectional communications, the remote server can establish to an arbitrary degree of certainty whether the person at the computer is in posession of the card holding the chip. Even if the line is completely unencrypted and open, nobody else can impersonate the holder of the card, no matter how much they listen.
Of course, whether AMEX is doing this right or doing something lame remains to be seen.
One of Lem's books describes a spacecraft dropping a cloud of dust-like smart devices for surveillance. That description almost certainly predates Stephenson.
The idea of using highly redundant systems for telemetry and monitoring doesn't seem to be all that new, and neither is the idea of passive readout. If MEMS can help make the devices smaller, that's a nice evolutionary step, but I won't really hold my breath...
I highly recommend reading Lem; he has a lot of neat ideas, and his stories are often funny or insightful as well.
The media seems to have an attention span of about six months. If something hasn't taken over and obliterated the competition by then, they declare it a failure. That's childish and irrational, of course, but as long as people keep reading the stuff, they keep writing.
Java has been spectacularly successful. Its most important use is probably server side development. Java's run-time safety, reflection, and dynamic loading are crucial features, and its libraries have become more complete and comfortable than those of OpenStep and other applications development platforms. Of course, Java isn't without warts, but it's a lot better than any other widely used platform out there. Java 2 also has made lots of progress on the client side, and apparently will be included with an upcoming browser release by AOL.
I'd expect that Linux, just like Java, will fall from media favor sooner or later. We'll get articles about how Windows 2000 is the best thing since sliced bread and how it will wipe every other operating system from the face of the earth, or about the rebirth of OS/2, or about BeOS or whatever. Just like Java, Linux will continue to be enormously successful.
Now, if we could only get a complete, working Java 2 implementation on Linux; the Blackdown port is so close, yet not quite ready for production use.
As a programmer, I love Python as a tool: it's a no-fuzz, text-based scripting tool that's very useful for many applications.
But why start from scratch when it comes to programming environments? Why not build on the decades of experience with other teaching languages, like Logo, Scheme, and Smalltalk?
And for real-world relevance, the facts of life are that there are hundreds of "little languages" serious computer users need to learn. Rather than pretending that everybody can exist in a little shell, exploring what is out there in the real world might make for a much better introductory CS curriculum.
And, in fact, even the "little languages" casual computer users encounter teach a lot of interesting concepts: event driven programming (dial-up scripts, game character programming), functional programming (database queries), OOP and message passing (Apple's scripting languages), etc.
I think it's nice to see more money go to Python, but the proposal really doesn't make that much sense to me. I think the development of a good introductory computer curriculum should start with no preconceived notions of which language to use.
probably licensing issues; good sign?
on
Is firewire dying?
·
· Score: 2
I suspect many potential FireWire users became wary of FireWire when Apple tried to impose licensing fees of $1/port.
Whenever a single company tries to control an important standard and loses the market, I think that's a good sign. Let's hope other companies notice.
Microsoft seems to admit that there is a backup key and furthermore that a backup key needs to exist to "ensure compliance with the NSA's technical review". It seems to me pretty academic to argue whether they have already shared that "backup key" with anyone.
But I'd ask the more general question: why does this surprise anyone? NT is not an open source product. It would be easy for any developer on the project to slip in a backdoor. Based on experience with other large software systems, I'd expect there to be dozens of backdoors in NT system and applications software. I wouldn't trust NT security further than I can throw a year's worth of MSDN CD's and documentation.
Decentralization in the form of end-user-run PCs doesn't solve any security problems. A single bad line of code in Windows opens up millions of Internet-connected PCs just as surely as a single bad line of code in Hotmail. But in addition to bugs, end users that maintain PCs generally have little experience or understanding of security issues.
Central, server-based applications remove a lot of chores and cares from users. That's no different from other centralized utilities: people used to generate their own power and water, but today, most people rely on utilities. Those utilities generally do pretty well and provide reliable service. Occasionally, they do something dumb, or they just have bad luck, and a lot of people end up having service outages, but from the point of view of each individual, the service is usually still very reliable.
From the point of view of security, a diversity of professionally run computer services both beats a Windows/PC monoculture and a single huge server.
As for Hotmail--what do you expect? It's a free service, so why should they assume any liabilities? If you want a company that stands behind their security, you probably have to pay for the service. And you have to do a little bit of shopping to identify companies and vendors that actually care and know something about security.
There may be lots of myths about FreeBSD, but I doubt that those are keeping people from switching from Linux to FreeBSD.
I was using BSD UNIX for more than a decade, since before it even had TCP/IP networking, and have been using Linux for a few years. Both seem to be good, solid systems, good enough for most server and desktop applications.
Given the momentum behind Linux, however, I see little reason to switch back to BSD. The technical differences seem academic to me (and I'm not convinced BSD always made the better choices). And most of the cutting-edge software that I care about seems to be coming out for Linux first.
I think for FreeBSD to attract more users, it needs to carve out a different niche from the one Linux is successful in. What that niche is, I don't know. What would pull me over to FreeBSD would be if it provided a rock-solid Java 2 implementation for client and server applications (something Linux currently lacks), and perhaps standardized on a desktop built around that.
Export controls are pretty clearly ineffective for keeping strong cryptography out of the hands of terrorists or criminals. That tells us that at the heart of the debate has to be something different.
I think it's pretty clear that the reason why the administration and the three letter agencies are fighting so hard against easing export controls is because they don't want strong cryptography become part of the communications infrastructure.
As soon as export controls are lifted, even just to "friendly" countries, most phone systems and communications standards and systems will incorporate strong cryptography, and routine monitoring of communications (for law enforcement, corporate intelligence, etc.) would become prohibitively expensive. Widespread use of strong cryptography would bring us back to the old days where wiretapping, bugging, etc., required specific targets and physical access.
This issue won't get resolved until the real underlying issues are recognized widely and the subject gets discussed openly.
I don't see a big difference to what has been previously reported. UNISYS isn't giving in at all--they can't sue people who created software with, say, Photoshop, because that's almost certainly prohibited by the license terms they have with Adobe.
In any case, who UNISYS sues or doesn't sue doesn't matter much. The reason why UNISYS is getting such bad press out of it is not because they are pursuing an iffy patent--lots of companies are doing that.
What irks people is that UNISYS seems to have allowed GIF to become a de-facto standard on the web without ever asserting their intellectual property rights. If they had made their intentions clear from the beginning, nobody would have used their technology. LZW isn't intrinsically valuable (there are lots of reasonable compression algorithms), it's only valuable because people have widely chosen to adopt it.
In fact, in other areas of property rights, UNISYS would have forfeited their rights through non-enforcement (c.f., real-estate easements, trademark law, etc.); it's an oddity of patent law that they are getting away with this. I think it also reflects poorly on their ethical standards; as a business partner, I'd be left wondering what kinds of rabbits the UNISYS legal department might be pulling out of their hat after I had signed a contract with them.
I think getting Linux software to run more easily on Windows matters a lot.
Many people use both Linux and Windows and would like to have the same tools available on both, including tools and software they develop themselves.
And making free Linux/UNIX software like the Gimp available on Windows makes Windows users more familiar with Linux software. That makes it easier for Windows users to transition onto non-Windows platforms, and it makes it easier for institutions to deploy a single set of tools across all platforms.
As a developer, I look carefully at where I invest time to learn new tools. A tool that is free on only some platforms and costs lots of money on other platforms is not very attractive to me.
I believe that Qt could probably succeed as well as it does if both its UNIX and Windows versions were released under GPL (not LGPL), with separate commercial licenses for commercial developers. I still find the current licensing situation of Qt to be a major obstacle to using it.
and the designers of the Java AWT probably should've followed the Netscape IFC team's lead and just reimplemented the OpenStep ApplicationKit.
That is exactly what Sun did with JFC/Swing. With Java 1.2, Sun also incorporated the OpenStep/DisplayPostscript imaging model, much of the WebObjects application server functionality, as well as extensive reflection capabilities.
Sun has hired lots of OpenStep/ObjC folks, and they also have a number of Smalltalk, Self, and CommonLisp experts, so Java carries on those traditions.
Slashdot Mirror
Using a journalling file system for the partition holding database tables will accomplish pretty much nothing (but it won't do much harm either).
Journalling file system in general don't help much with data integrity either; the kinds of guarantees they usually make are minimal, they don't protect you against many sources of data loss, and they don't even guarantee fast system startups.
The biggest thing SPARC-based systems have going for them in my mind is that they run a good version of Java on a UNIX platform. The other really good implementation of Java is for NT.
Java on other UNIX platforms (Linux, FreeBSD, etc.) is either not the latest version (only up to 1.1) or still not entirely reliable (1.2 on Linux, etc.).
Using the MMU for this purpose is a poor engineering tradeoff.
C and C++ are exceptional among programming languages in terms of the poor fault isolation they provide. Most languages do better. Some, like Java, do very well.
However, while lack of fault isolation is one of the biggest problems in C/C++ software, even with excellent fault isolation, there are still plenty of bugs possible.
While the press release doesn't say much about neural networks or whether the state of the art in speech recognition has improved, it tells us something about a disregard by USC for standards of scientific conduct: scientific publication by press release is improper.
Likewise, we all want our technical toys to be smaller, so "pioneers" building very expensive wearable computers don't impress me that much--they are doing what is natural. But a company that figures out how to produce them competitively and at a price I can afford, that's the real innovation to me.
FWIW, even if Sun gave them a good discount, that's just fine with me. It's a competitive market, and Microsoft has been "donating" (and probably tax deducting) hardware and software to schools in huge amounts. Until that practice is declared anticompetitive and prohibited by the FTC, Sun has to play along. If they don't already do it, I think they should start special pricing and donations to schools ASAP.
But while an alternative is good, I personally don't like this particular alternative: I don't like the look, I don't like the pointing device or keyboard, and I don't like the software. Imagine Apple would take over the market and all computers would come in semi-transparent rounded lumps with trim in various colors. I'd find it pretty awful.
Hardware based authentication and security tokens should be based on something portable, and that portable needs to have enough compute power to implement something like zero knowledge proofs. SmartCards fit the bill, and they are cheap. Keyboards should have SmartCard readers, and standard cryptographic methods allow secure transactions to be executed with SmartCards even over untrusted machines.
At best, the computer itself could benefit from hardware encryption that doesn't carry a key, in order to speed up throughput for encrypted data streams. But in the current political climate, putting hardware-based encryption into a PC is futile, since, according to US laws, it cannot be secure anyway.
Of course, e-commerce companies don't like SmartCards because, oh my, the consumer can remove them when they don't want to buy anything and don't want to get tracked. ID chips tied to the CPU or motherboard are great: the kids can order, the software can be used to track people wherever they go, and there is little most people can do about it if they run standard software like Windows.
If IBM wants to drive secure e-commerce, they should be shipping computers with SmartCard enabled keyboards.
Even if you met the convenience and durability of paper with an electronic product, you still have to face the problem that the electronic product is expensive. Would you take a $1000 rollup electronic newspaper with you everywhere?
Paper can be produced more cheaply and more environmentally friendly from renewable resources. One of the primary ones would be (drug free) hemp.
Compare that with Microsoft. Microsoft resists any careful documentation or standardization of their APIs, Microsoft constantly strives to replace non-proprietary APIs with proprietary APIs, and Microsoft's software (in particular, their APIs) is a mess.
As for StarOffice, if an X11 and UNIX-based office suite catches on widely in corporations, that's good for Linux, all things being equal. The server-based nature of their system is a boon to businesses, something Microsoft can't match right now. If the whole thing runs on Linux, that's nice, but if it doesn't, no harm done.
Sun is no "friend" of Linux beyond what serves their business needs. And that is as it should be. As long as they keep doing what they are doing, Linux and Sun have similar goals and they deserve our support.
Maybe the patent office could be privatized and both the patent office and patent applicants could be found liable and assessed penalties if patents are overturned. The problem right now is that the risk and cost of applying for a frivolous patent is tiny compared to the benefit of getting away with it occasionally.
It might be possible that a lawyer could argue that a company that has obtained an invalid patent has greatly harmed its competitors and should pay fines and penalties. But in the current environment, it's already so hard even to establish that patents with obvious prior art are invalid that I wouldn't hold my breath for that to happen.
So, think of the 600E as a Thinkpad without a modem when you are running Linux. You aren't paying a lot extra for the modem, and the built-in Thinkpad modem is a kludge anyway.
Using zero knowledge proofs and bidirectional communications, the remote server can establish to an arbitrary degree of certainty whether the person at the computer is in posession of the card holding the chip. Even if the line is completely unencrypted and open, nobody else can impersonate the holder of the card, no matter how much they listen.
Of course, whether AMEX is doing this right or doing something lame remains to be seen.
The idea of using highly redundant systems for telemetry and monitoring doesn't seem to be all that new, and neither is the idea of passive readout. If MEMS can help make the devices smaller, that's a nice evolutionary step, but I won't really hold my breath...
I highly recommend reading Lem; he has a lot of neat ideas, and his stories are often funny or insightful as well.
Java has been spectacularly successful. Its most important use is probably server side development. Java's run-time safety, reflection, and dynamic loading are crucial features, and its libraries have become more complete and comfortable than those of OpenStep and other applications development platforms. Of course, Java isn't without warts, but it's a lot better than any other widely used platform out there. Java 2 also has made lots of progress on the client side, and apparently will be included with an upcoming browser release by AOL.
I'd expect that Linux, just like Java, will fall from media favor sooner or later. We'll get articles about how Windows 2000 is the best thing since sliced bread and how it will wipe every other operating system from the face of the earth, or about the rebirth of OS/2, or about BeOS or whatever. Just like Java, Linux will continue to be enormously successful.
Now, if we could only get a complete, working Java 2 implementation on Linux; the Blackdown port is so close, yet not quite ready for production use.
But why start from scratch when it comes to programming environments? Why not build on the decades of experience with other teaching languages, like Logo, Scheme, and Smalltalk?
And for real-world relevance, the facts of life are that there are hundreds of "little languages" serious computer users need to learn. Rather than pretending that everybody can exist in a little shell, exploring what is out there in the real world might make for a much better introductory CS curriculum.
And, in fact, even the "little languages" casual computer users encounter teach a lot of interesting concepts: event driven programming (dial-up scripts, game character programming), functional programming (database queries), OOP and message passing (Apple's scripting languages), etc.
I think it's nice to see more money go to Python, but the proposal really doesn't make that much sense to me. I think the development of a good introductory computer curriculum should start with no preconceived notions of which language to use.
Whenever a single company tries to control an important standard and loses the market, I think that's a good sign. Let's hope other companies notice.
But I'd ask the more general question: why does this surprise anyone? NT is not an open source product. It would be easy for any developer on the project to slip in a backdoor. Based on experience with other large software systems, I'd expect there to be dozens of backdoors in NT system and applications software. I wouldn't trust NT security further than I can throw a year's worth of MSDN CD's and documentation.
Central, server-based applications remove a lot of chores and cares from users. That's no different from other centralized utilities: people used to generate their own power and water, but today, most people rely on utilities. Those utilities generally do pretty well and provide reliable service. Occasionally, they do something dumb, or they just have bad luck, and a lot of people end up having service outages, but from the point of view of each individual, the service is usually still very reliable.
From the point of view of security, a diversity of professionally run computer services both beats a Windows/PC monoculture and a single huge server.
As for Hotmail--what do you expect? It's a free service, so why should they assume any liabilities? If you want a company that stands behind their security, you probably have to pay for the service. And you have to do a little bit of shopping to identify companies and vendors that actually care and know something about security.
I was using BSD UNIX for more than a decade, since before it even had TCP/IP networking, and have been using Linux for a few years. Both seem to be good, solid systems, good enough for most server and desktop applications.
Given the momentum behind Linux, however, I see little reason to switch back to BSD. The technical differences seem academic to me (and I'm not convinced BSD always made the better choices). And most of the cutting-edge software that I care about seems to be coming out for Linux first.
I think for FreeBSD to attract more users, it needs to carve out a different niche from the one Linux is successful in. What that niche is, I don't know. What would pull me over to FreeBSD would be if it provided a rock-solid Java 2 implementation for client and server applications (something Linux currently lacks), and perhaps standardized on a desktop built around that.
I think it's pretty clear that the reason why the administration and the three letter agencies are fighting so hard against easing export controls is because they don't want strong cryptography become part of the communications infrastructure.
As soon as export controls are lifted, even just to "friendly" countries, most phone systems and communications standards and systems will incorporate strong cryptography, and routine monitoring of communications (for law enforcement, corporate intelligence, etc.) would become prohibitively expensive. Widespread use of strong cryptography would bring us back to the old days where wiretapping, bugging, etc., required specific targets and physical access.
This issue won't get resolved until the real underlying issues are recognized widely and the subject gets discussed openly.
In any case, who UNISYS sues or doesn't sue doesn't matter much. The reason why UNISYS is getting such bad press out of it is not because they are pursuing an iffy patent--lots of companies are doing that.
What irks people is that UNISYS seems to have allowed GIF to become a de-facto standard on the web without ever asserting their intellectual property rights. If they had made their intentions clear from the beginning, nobody would have used their technology. LZW isn't intrinsically valuable (there are lots of reasonable compression algorithms), it's only valuable because people have widely chosen to adopt it.
In fact, in other areas of property rights, UNISYS would have forfeited their rights through non-enforcement (c.f., real-estate easements, trademark law, etc.); it's an oddity of patent law that they are getting away with this. I think it also reflects poorly on their ethical standards; as a business partner, I'd be left wondering what kinds of rabbits the UNISYS legal department might be pulling out of their hat after I had signed a contract with them.
Many people use both Linux and Windows and would like to have the same tools available on both, including tools and software they develop themselves.
And making free Linux/UNIX software like the Gimp available on Windows makes Windows users more familiar with Linux software. That makes it easier for Windows users to transition onto non-Windows platforms, and it makes it easier for institutions to deploy a single set of tools across all platforms.
As a developer, I look carefully at where I invest time to learn new tools. A tool that is free on only some platforms and costs lots of money on other platforms is not very attractive to me.
I believe that Qt could probably succeed as well as it does if both its UNIX and Windows versions were released under GPL (not LGPL), with separate commercial licenses for commercial developers. I still find the current licensing situation of Qt to be a major obstacle to using it.
GTK, however, is being ported to Windows, and it's good enough already to use Gimp on Windows with it.
That is exactly what Sun did with JFC/Swing. With Java 1.2, Sun also incorporated the OpenStep/DisplayPostscript imaging model, much of the WebObjects application server functionality, as well as extensive reflection capabilities.
Sun has hired lots of OpenStep/ObjC folks, and they also have a number of Smalltalk, Self, and CommonLisp experts, so Java carries on those traditions.