Shuttle launches cost what,$20-50M. It shouldn't be that hard to come up with $1M per year.
That's the problem, I'm afraid. Most of NASA's budget is committed to the International Space Station and its supporting Shuttle flights; the rest of NASA is on very tight rations as a result.
Wedged government is good government. If somehow we can keep the Government from starting any new big expensive programs, or screwing around with the tax code in major ways, the current revenue surplus will likely continue, and be automatically applied to reducing the deficit.
Keeping the legeslative and executive branches in different parties won't work 100%, but it will likely work better than having them aligned - complete control by either party scares me, frankly.
For me, Gore in the White House and a Republican Congress is slightly preferrable to Bush and a Democratic Congress, primarly because of the Supreme Court.
The Government is currently taking in more money than it is putting out (Social Security included, both benefits and payroll taxes) and it is required by law to reduce the national debt whenever that condition occurs. Bipartisan rhetorical nonsense aside, the Social Security system has always been just another fungible revenue source.
Do you seriously think that when the Social Security funds "run out" the Government is going to just sit there and turn its pockets inside out? They'll make those payments (or reduced versions thereof) somehow, which will mean borrowing money, which is why it's good that the national debt is being reduced now.
If the legislative and executive branches can't agree on major new programs, or major new wrinkles in the tax code, the current situation has the budget surplus paying down the deficit. It would be nice if we could actually clean up the tax code and delete useless programs, but given government tendencies, this may be the best we can do.
The fact that I am a 44-year-old who will be needing Social Security just as it is expected to run dry (and thus I want the national debt to be as small as possible, so it will be feasible to borrow money to handle the downhill side of the boomers) has nothing to do with my wanting this, of course.
Did a lot of work on basic molecular modeling out of PARC.
Of course, the fact that he left there in 1999 is also relevant to this thread.
*sigh*... I worked for Xerox AI Systems in 1986-8; we were one of Xerox's many attempts to commercialize PARC research in AI and Lisp. Our record for cooperation with PARC was mixed - they loaned us Larry Masinter and Bill van Melle when we were implementing Xerox Common Lisp, and that was an enormous help, but getting stuff out of them without specific direction was often difficult.
... and I mean real standards, as in implementation-independent specifications generated and approved by real standards bodies.
.NET does not qualify here, and won't as long as you have to say "Microsoft.NET" to be completely correct.
Java does not qualify here, and won't until Sun turns at least the core of the language over to a non-vendor-controlled standards body (ISO, IEEE, whatever).
The other obvious cross-platform web choices (PERL, Python, tcl/tk) don't qualify here either. Definition by portable implementation is not sufficient - it can't handle questions like whether JPython and Python are the same language.
A major web project in my company evaluated all of the above as potential platform choices, tried all of them, and chose...
Common Lisp. True, it has no standard socket API, but the rest of the language is controlled by an ISO specification, so it's by-God cross-platform, cross-implementation portable. We have rigidly defined areas of doubt and uncertainty, as opposed to the shifting sands you get with rapidly evolving/mutating systems like Java and.NET.
If that's so, then CORBA sucks even
worse than I thought.
I'll assume you are just badly informed, rather than arrogant.
Perhaps I should have stated that a little differently:
The CORBA model for resource management is appropriate for its domain: big distributed applications. Basing Bonobo on CORBA sounds like a good way to get distributed applications, but if the cost is that you have to accept CORBA's model for resource management even among objects that are all on your machine, that's too much for me.
For someone who clearly knows a bunch about a bunch of different languages, Mr. Arbel doesn't seem to know much about garbage collection in general, and Java/JVM GC in particular.
The JVM is not required to do mark-sweep GC. The JVM spec ifically leaves the implementation of storage management unspecified.
This is good because it means that Java can use modern, higher-performance GC strategies like stop-and-copy or generational GC, both of which have been in use in Lisp and Scheme systems since the 1980's. I strongly suspect that C# will have to use mark-sweep or some other non-relocating GC, since you're allowed to go down below it to assembler, exposing pointers that might need relocation.
Do most JVM implementations really still use mark-sweep GC? Despite James Gosling and Guy Steele both being ur-Lisp hackers?
... storage management via reference count really, REALLY, REALLY sucks. It's a first-class recipe for memory/resource leaks whose badness has been enshrined in the Hacker's Dictionary for Ghod's sake. Its only advantage is that it spreads the management overhead out so evenly over your whole system that you can't see how much memory and time it really takes up.
Miguel de Icaza seems like an otherwise intelligent guy, so I have to assume that CORBA is forcing the use of reference counting here. If that's so, then CORBA sucks even worse than I thought.
Thinking in metaphors is equivalent to reasoning by analogy, a weak form of induction that makes you feel good about your conclusions but has no real logical validity.
IANAL, but I gather that reasoning by analogy is what passes for logic in the law. This only makes sense when you realize that what's important in court or the legislature is convincing the judge/jury/lawmakers that your argument is the correct one - verisimilitude rather than reality.
... who blame computers for problems better assigned to:
Teachers who don't know how to use computers
Educational sofware writers who can't get past drill-and-kill programs
Administrators and funding agencies who think dumping hardware and software into classrooms is sufficient.
Everyone should go sit down calmly, take a stress pill and read Seymour Papert. Computers are needed in classrooms, but not to teach kids how to surf the web. They are needed to transform education from its current one-way, passive mode to something more active.
The active optics can probably be done with MEMS, and the measuring lights and receivers are probably semiconductor-based already. The whole system could be moved into technologies that are benefiting from Moore's Law, and once it's there the size and cost will drop exponentially.
Hell, this project at the MIT Media Lab wedged a camera, speaker, and enough compute power to recognize and translate 40 words of American Sign Language into a baseball cap. In 1997!
The Rete match algorithm made rule-based systems computationally practical. Discovered by Charles Forgy in 1979 or so, it was probably the most important event in the expert systems boom of the 1980's.
The suggestions of Visual Basic, Logo, Squeak, and JavaScript are all good, because they all have graphical interaction and ways to write a small first program. I would be careful, though, to keep a new programmer from fixating on any of them as the One True Way to communicate with computers. I know a lot of people who are decent hackers in language X but are helpless outside it, and many of them don't want to learn anything different. The people who are going to own the world are the ones who can change their toolsets in Internet time, the ones whose opinions are based on use rather than on prejudice.
The best start you can give your kids is experience in more than one way to compute. Once they have an interest in one language, show them something they already know how to do, but in another environment. If they started out OO, try something that isn't, and vice-versa; if their first language was all drag-and-drop, slip in something with structured text; if they start out in Visual Basic, show them something cross-platform, so they'll know other platforms exist.
Because the vast majority of problems that need computer solutions aren't really that complicated, so high-level languages are overkill for them. Visual Basic is just about right for most applications written it, after all.
Be honest: how much of your time programming do you spend doing things that are really new? Most of the time, the shortest way to a solution is not an expressive language, but the ability to find and reuse code that solves straightforward problems. This is especially true now that we have the Web and searchable on-line code libraries. The size of the community using the langauge can be more important than the utility or safety of the language itself - if language X makes you ten times as productive, but there are a thousand times as many people using language Y and their libraries are broad and deep, you may be better off with Y.
When the problems are unusual, one wizard with the right tools can beat fifty drones; for most programs written today, high-level langauges don't buy you all that much.
As a long-time Lisp hacker, I gnash my teeth and wail as I say this, but it's all still true. Go find a copy of Gabriel's Patterns of Software and read the chapter entitled "The End of History and The Last Programming Language" for further depressing reading.
I work for a NASA contractor. The Government head of our area got a copy of ILOVEYOU from one of our University partners. As we do satellite image processing, it was just conceivable that they might have legitimately sent us a pretty-picture executable, so he tried it.
He was on a Mac, so the program failed; he then forwarded it to another guy running Windows. He got clobbered when he ran it, and I got my copy from him. I read mail on a Unix box with Netscape (and I know better than to ever run an executable attachment), so I was safe.
This shows why e-mail is such an effective way to spread this stuff. The virus was passed from user to user without being executed, through machines where it couldn't execute, and was still dangerous when it arrived.
Seriously. There's a good chance you've already visited a web site running Lisp, and didn't know it (Yahoo Store for one). Lisp's forte is extreme dynamic content that requires persistence or cross-session communications.
Check out any of the following:
SIOD - A small, easy-to-extend Scheme interpreter that can be used to do CGI.
AllegroServe - This is a small Common Lisp based web server which can sit behind Apache or do it all itself. They have a very nice mapping between Lisp lists and HTML that makes dynamic content generation a breeze. The source is currently implementation-specific, but it's LGPLed and available at Sourceforge.
CL-HTTP - If you want to do dynamic content and don't want to learn HTML, this is what you want. Query processing and page generation are buried behind deep and powerful abstractions. Long learning curve, but worth it for really complex stuff.
There are others, but I don't have the links handy right now.
Then maybe the next experiment should be done on a satellite platform up in geosynchronous orbit. From the satellite's point of view, nothing will be moving - not the earth, nor any of the other satellites near it. As a bonus, you get hard vacuum and low temperatures for free (cheap, anyway - some venting and thermal blankets).
The problem then is designing an apparatus that will survive launch and still be accurate. Probably the thing to do is take the parts up to the space station, assemble and test it there, then move it slowly and carefully up to GEO - maybe with an ion drive.
A question like this is best asked of people who rarely appear in the media - people who don't normally take the opportunity to speak their minds before a wide audience.
RMS is definitely not like that - he will give you his opinions in any forum at any time. If he has some important question to ask, he has almost certainly already asked it in an appropriate place. This question is not a good way to get something new out of RMS.
Please don't let this question be one of the ones forwarded to RMS unless there are only nine other non-lame questions.
They did, after all, publish this editorial by Charles Ferguson, who supports breaking Microsoft up the right way (OS/apps/etc.). Finally, someone who understands that being a monopoly isn't illegal, but using a monopoly to create other monopolies is.
Then again, they also published this editorial by Robert Samuelson, who trots out all the "Microsoft hasn't hurt anyone" nonsense typically spouted by people who believe that the words "Microsoft", "innovative", and "software" can truthfully be used in the same sentence.
XML is little more than a syntax for writing down tree structures. It is equivalent in expressive power to S-expressions (the core data type in the Lisp family of languages - 42 years old in 2000). I'm far from the only person that has noticed this - go see this PDF by Phillip Wadler at Bell Labs, especially pages 5-8. He also has some other good XML links.
The only reason XML excites people is it looks like something they're familiar with (SGML/HTML). All the old representational issues that the AI community has been grappling with for decades will now be recast into XML terms, and "solved" in half-ass fashion by people who won't know where to look in the literature for existing techniques.
Microsoft loves XML - have you thought about why they like it? It's because they can claim to be "open" and "conforming to standards" by using its syntax, and still have enough control over the underlying semantics to keep developers and users on the upgrade tradmill.
XML - more than enough rope, with a godawful syntax to boot. Long live () and ""!
Slashdot Mirror
Shuttle launches cost what,$20-50M. It shouldn't be that hard to come up with $1M per year.
That's the problem, I'm afraid. Most of NASA's budget is committed to the International Space Station and its supporting Shuttle flights; the rest of NASA is on very tight rations as a result.
Wedged government is good government. If somehow we can keep the Government from starting any new big expensive programs, or screwing around with the tax code in major ways, the current revenue surplus will likely continue, and be automatically applied to reducing the deficit.
Keeping the legeslative and executive branches in different parties won't work 100%, but it will likely work better than having them aligned - complete control by either party scares me, frankly.
For me, Gore in the White House and a Republican Congress is slightly preferrable to Bush and a Democratic Congress, primarly because of the Supreme Court.
I have occasionally fantisized about patenting some useful technique, and then publishing a license that says:
You may use my patent for free if and only if your code conforms to the GPL.
This would give the GPL (at least as it applies to the patented technique) real legal teeth.
The bug is that I can't figure out where the money needed to file suits to enforce the license would come from.
The Government is currently taking in more money than it is putting out (Social Security included, both benefits and payroll taxes) and it is required by law to reduce the national debt whenever that condition occurs. Bipartisan rhetorical nonsense aside, the Social Security system has always been just another fungible revenue source.
Do you seriously think that when the Social Security funds "run out" the Government is going to just sit there and turn its pockets inside out? They'll make those payments (or reduced versions thereof) somehow, which will mean borrowing money, which is why it's good that the national debt is being reduced now.
If the legislative and executive branches can't agree on major new programs, or major new wrinkles in the tax code, the current situation has the budget surplus paying down the deficit. It would be nice if we could actually clean up the tax code and delete useless programs, but given government tendencies, this may be the best we can do.
The fact that I am a 44-year-old who will be needing Social Security just as it is expected to run dry (and thus I want the national debt to be as small as possible, so it will be feasible to borrow money to handle the downhill side of the boomers) has nothing to do with my wanting this, of course.
Did a lot of work on basic molecular modeling out of PARC.
Of course, the fact that he left there in 1999 is also relevant to this thread.
*sigh*... I worked for Xerox AI Systems in 1986-8; we were one of Xerox's many attempts to commercialize PARC research in AI and Lisp. Our record for cooperation with PARC was mixed - they loaned us Larry Masinter and Bill van Melle when we were implementing Xerox Common Lisp, and that was an enormous help, but getting stuff out of them without specific direction was often difficult.
... when they show up in a click-through EULA, along with the various things they feel it necessary to bind your future behavior.
... and I mean real standards, as in implementation-independent specifications generated and approved by real standards bodies.
.NET does not qualify here, and won't as long as you have to say "Microsoft .NET" to be completely correct.
.NET.
Java does not qualify here, and won't until Sun turns at least the core of the language over to a non-vendor-controlled standards body (ISO, IEEE, whatever).
The other obvious cross-platform web choices (PERL, Python, tcl/tk) don't qualify here either. Definition by portable implementation is not sufficient - it can't handle questions like whether JPython and Python are the same language.
A major web project in my company evaluated all of the above as potential platform choices, tried all of them, and chose...
Common Lisp. True, it has no standard socket API, but the rest of the language is controlled by an ISO specification, so it's by-God cross-platform, cross-implementation portable. We have rigidly defined areas of doubt and uncertainty, as opposed to the shifting sands you get with rapidly evolving/mutating systems like Java and
... by the Technical Advisory Committee to Develop a Federal Information Processing Standard for the Federal Key Management Infrastructure which chooses to refer to itself as "Bob" rather than "TACDFIPSFKMI". See here if you don't believe me. As they both have something to do with Federal crypto standards, it would be too confusing to have them both named "Bob".
Perhaps I should have stated that a little differently:
The CORBA model for resource management is appropriate for its domain: big distributed applications. Basing Bonobo on CORBA sounds like a good way to get distributed applications, but if the cost is that you have to accept CORBA's model for resource management even among objects that are all on your machine, that's too much for me.
The JVM is not required to do mark-sweep GC. The JVM spec ifically leaves the implementation of storage management unspecified.
This is good because it means that Java can use modern, higher-performance GC strategies like stop-and-copy or generational GC, both of which have been in use in Lisp and Scheme systems since the 1980's. I strongly suspect that C# will have to use mark-sweep or some other non-relocating GC, since you're allowed to go down below it to assembler, exposing pointers that might need relocation.
Do most JVM implementations really still use mark-sweep GC? Despite James Gosling and Guy Steele both being ur-Lisp hackers?
Miguel de Icaza seems like an otherwise intelligent guy, so I have to assume that CORBA is forcing the use of reference counting here. If that's so, then CORBA sucks even worse than I thought.
Thinking in metaphors is equivalent to reasoning by analogy, a weak form of induction that makes you feel good about your conclusions but has no real logical validity.
IANAL, but I gather that reasoning by analogy is what passes for logic in the law. This only makes sense when you realize that what's important in court or the legislature is convincing the judge/jury/lawmakers that your argument is the correct one - verisimilitude rather than reality.
Teachers who don't know how to use computers
Educational sofware writers who can't get past drill-and-kill programs
Administrators and funding agencies who think dumping hardware and software into classrooms is sufficient.
Everyone should go sit down calmly, take a stress pill and read Seymour Papert. Computers are needed in classrooms, but not to teach kids how to surf the web. They are needed to transform education from its current one-way, passive mode to something more active.
Hell, this project at the MIT Media Lab wedged a camera, speaker, and enough compute power to recognize and translate 40 words of American Sign Language into a baseball cap. In 1997!
I give it five years, max.
The Rete match algorithm made rule-based systems computationally practical. Discovered by Charles Forgy in 1979 or so, it was probably the most important event in the expert systems boom of the 1980's.
The best start you can give your kids is experience in more than one way to compute. Once they have an interest in one language, show them something they already know how to do, but in another environment. If they started out OO, try something that isn't, and vice-versa; if their first language was all drag-and-drop, slip in something with structured text; if they start out in Visual Basic, show them something cross-platform, so they'll know other platforms exist.
Be honest: how much of your time programming do you spend doing things that are really new? Most of the time, the shortest way to a solution is not an expressive language, but the ability to find and reuse code that solves straightforward problems. This is especially true now that we have the Web and searchable on-line code libraries. The size of the community using the langauge can be more important than the utility or safety of the language itself - if language X makes you ten times as productive, but there are a thousand times as many people using language Y and their libraries are broad and deep, you may be better off with Y.
When the problems are unusual, one wizard with the right tools can beat fifty drones; for most programs written today, high-level langauges don't buy you all that much.
As a long-time Lisp hacker, I gnash my teeth and wail as I say this, but it's all still true. Go find a copy of Gabriel's Patterns of Software and read the chapter entitled "The End of History and The Last Programming Language" for further depressing reading.
I work for a NASA contractor. The Government head of our area got a copy of ILOVEYOU from one of our University partners. As we do satellite image processing, it was just conceivable that they might have legitimately sent us a pretty-picture executable, so he tried it.
He was on a Mac, so the program failed; he then forwarded it to another guy running Windows. He got clobbered when he ran it, and I got my copy from him. I read mail on a Unix box with Netscape (and I know better than to ever run an executable attachment), so I was safe.
This shows why e-mail is such an effective way to spread this stuff. The virus was passed from user to user without being executed, through machines where it couldn't execute, and was still dangerous when it arrived.
Check out any of the following:
SIOD - A small, easy-to-extend Scheme interpreter that can be used to do CGI.
AllegroServe - This is a small Common Lisp based web server which can sit behind Apache or do it all itself. They have a very nice mapping between Lisp lists and HTML that makes dynamic content generation a breeze. The source is currently implementation-specific, but it's LGPLed and available at Sourceforge.
CL-HTTP - If you want to do dynamic content and don't want to learn HTML, this is what you want. Query processing and page generation are buried behind deep and powerful abstractions. Long learning curve, but worth it for really complex stuff.
There are others, but I don't have the links handy right now.
Exactly correct. What it means is that now nobody but Microsoft may use the obvious mechanism for maintaining package information on Windows systems.
Why they would want to make all non-Microsoft software packages harder to install and maintain on their OS remains a mystery...
Then maybe the next experiment should be done on a satellite platform up in geosynchronous orbit. From the satellite's point of view, nothing will be moving - not the earth, nor any of the other satellites near it. As a bonus, you get hard vacuum and low temperatures for free (cheap, anyway - some venting and thermal blankets).
The problem then is designing an apparatus that will survive launch and still be accurate. Probably the thing to do is take the parts up to the space station, assemble and test it there, then move it slowly and carefully up to GEO - maybe with an ion drive.
A question like this is best asked of people who rarely appear in the media - people who don't normally take the opportunity to speak their minds before a wide audience.
RMS is definitely not like that - he will give you his opinions in any forum at any time. If he has some important question to ask, he has almost certainly already asked it in an appropriate place. This question is not a good way to get something new out of RMS.
Please don't let this question be one of the ones forwarded to RMS unless there are only nine other non-lame questions.
Then again, they also published this editorial by Robert Samuelson, who trots out all the "Microsoft hasn't hurt anyone" nonsense typically spouted by people who believe that the words "Microsoft", "innovative", and "software" can truthfully be used in the same sentence.
The only reason XML excites people is it looks like something they're familiar with (SGML/HTML). All the old representational issues that the AI community has been grappling with for decades will now be recast into XML terms, and "solved" in half-ass fashion by people who won't know where to look in the literature for existing techniques.
Microsoft loves XML - have you thought about why they like it? It's because they can claim to be "open" and "conforming to standards" by using its syntax, and still have enough control over the underlying semantics to keep developers and users on the upgrade tradmill.
XML - more than enough rope, with a godawful syntax to boot. Long live () and ""!