Would anyone know if his co-author Robert Morris is the same Robert Morris (or his father) of the infamous Morris internet worm from the late 80's?
Yes, he means the Robert T. Morris of Internet worm fame. In a talk Graham gave at the 1998 Lisp Users and Vendors conference, he even joked about this, saying something like "remember when you do anything at Yahoo Store, your data is being handled by code written by RTFM".
Yes there must be an audit trail, yes machine-generated paper is good, but NO handing out receipts that show how you voted.
The problem is the possibility of vote coercion - your boss/landlord/relative says they'll do something nasty to you unless you bring them a receipt from the polls that shows you voted the way they want you to. It doesn't matter whether the receipt has your name on it or not, as long as it shows how you voted, and it's hard to fake (and if it's easy to fake, it's not a useful receipt).
The real problem is that evolved hardware can specialize itself very precisely for its environment, and the environment includes everything.
Some of Thompson's early evolved designs only worked in a narrow temperature range, because that's what they evolved in (see here for the article that prompted Slashdot to look at this last time).
Company X states that patents A, B,... will be licensed at no-cost to anyone whose code using those patents is covered by any of the free/open-source licenses FooL, BarL,...
Something like this would pull a major thorn out of the community's paw - the threat of legal action for independent rediscovery of patented algorithms.
A company that did this could still use its patents to beat up on other closed-source companies - they would just be saying they don't want to use the same tactics against open-source projects.
... the 50-foot high Teletype machine for a console?
Opposing view links should not be required
on
Republic.Com
·
· Score: 1
... because the presence or absence of links to opposing viewpoints is a useful heuristic as to the sanity of the site.
Sites that have opposing viewpoint links are saying that they are confident enough in their views (and your intelligence) that pointing to the opposition only makes their case stronger. It's also an excellent way to kick sand in the face of the opposition - "we link to them, but they're not brave enough to link to us".
Sites that don't have opposition links are saying the opposite - that they don't trust you to agree with them, that they think the opposition might fool you despite their best efforts, that they're not 100% confident in their arguments.
It will be worth people's time to write real compilers for (Perl/Python/Ruby/whatever) when there are real, non-platform-specific specifications of what the languages mean, so compiler writers have something to aim at that won't move out from under them.
This is the fundamental reason why languages defined by portable implementations have problems growing beyond those implementations. The implementation actually specifies too much of the langauge's behavior - a real spec says explicitly what behavior is implementation-dependent, which tells implementors and users both exactly where the boundaries are.
The Common Lisp and Scheme communities wrote real specs for their languages (ANSI and IEEE, respectively). The user communities routinely beat up implementors and vendors who don't conform to the spec. As a result, they can write portable programs that compile to native code, without getting stuck in the C/C++ quagmire.
Someday there may be a way to answer these questions:
Are JPerl and Perl the same language?
How about JPython and Python?
that doesn't boil down to "ask Larry/Guido". A year or two after that, you might see real compilers for those languages.
1. Installing the patch(es) once is not enough. When Windows pollutes its environment enough so you have to reinstall it, you have to reinstall the patches, too, which means you have to be organized about downloading them, putting them somewhere safe and easy to find, installing them all in order, and cleaning them out as service packs come along. I suspect most users would prefer to live in denial ("I don't really need that").
2. How can we trust Microsoft/whoever to not add extraneous stuff in patches? I would object if they decided to issue a security patch that also upgraded, say, your DNS service to work better with Microsoft servers. I want to be able to choose whether or not to participate in the latest embrace&extend maneuver, but with closed-source patches, there's no way to tell.
... but like the man said, it was possible to hack VM into something almost useable. At Clemson in the mid 70's we had a home-grown edit/batch-submit/output-view system that was about as interactive as you could get with 3270 terminals. The editor could squirt numbered source lines onto your screen, you could edit them there, squirt the screen back, and have your changes accepted in one gulp.
Depressingly like filling out forms in a web browser...
DS1 is running an RS/6000 (IBM's version of the PowerPC architecture) at something like 25 MHz. That was the fastest rad-hardened chip they could get when they started the design.
The operating system is RTWorks, a Unix-like realtime kernel. The system software is written in C and Common Lisp (for the Autonomous Remote Agent stuff).
How about something like a copyleft for patents, in which the holder of a patent states that they will give a no-cost license for its use in any application covered by, say, the GPL? The holder could also grant licenses to non-free software creators in the usual way.
This strikes me as a middle ground between the current situation (patent something and use it to beat everyone senseless) and no patents at all. A patent holder could use it as an advantage over non-free software, while leaving the free/open-source communities alone.
... the day someone successfully patents the first legal argument. When the lawyers realize that their bread and butter has many of the same "patentable" characteristics as algorithms, they just might get a clue about the current situation.
Alternatively, someone might try to defend themselves in court by saying that it is impossible to prove two algorithms are equivalent (because doing so is equivalent to solving the Halting Problem), so you can't really prove infringement. This is of course BS in this context (you can often prove equivalence for a particular pair of algorithms - it's the general case that steps in the Halting Problem), but it might confuse things enough to get you out of court.
... state that they will accept electronic copies of documents only in formats with open, published standards. Damn near every element of the US Government is required by law to submit its documents to the Archive. Currently the Archives accepts only paper (see here), primarly because they can't insure they can save digital stuff readably long enough. Allowing agencies to submit stuff electronically would save everyone a bunch of time and money, and they could still say it's readable forever if they burned the data into CD-ROMS and "printed" it to microfilm.
Doing this would create an enormous market (the Government buys a bunch of software) for open formats, which would hit Microsoft right in its upgrade-treadmill solar plexus.
Check out the liability caps in the FCRA and the FDCPA - $500,000.00 is nothing to these guys.
That makes it harder, but not impossible. Just file multiple suits, each with a group of, say, 100 people asking for $5000 in damages each. Even doing this once per state works out to $25M, plus legal fees and hassle.
Unfortunately, those denied credit are generally those who can't afford to launch a legal attack against a well funded opponent.
When a large entity screws a lot of people over, the normal remedy is a class action lawsuit. The net should help here, as it will help the screwees find each other and get organized. Surely some lawyer somewhere has smelled blood in the water from this problem...
Does anyone know exactly what jobs the people who are suing held at Microsoft? I haven't seen a list in any of the media reports.
I can believe the people who say MS policy is "we hire hackers based on ability - race is irrelevant". This is probably true for the technical positions (or as true as it can get in America).
BUT, MS is a big company, and has lots of jobs that are less technical. I could easily believe racism is alive and well in sales, marketing, packaging design, legal, etc...
but this is a giant program to keep Russian Aerospace engineers employed.
The ISS has always (at least from a US budgetary point of view) been about keeping American aerospace engineers employed. This has been NASA's primary reason for existence for many years now... at least according to Congress.
If NASA was supposed to be about science and aero/astronomical research, maybe we'd have a budget for them instead of for relatively useless stuff like the ISS.
When anyone asks me what it's like in Silicon Valley, that's what I tell them. I lived there from '86 - '88 (working for somewhat-lamented Xerox AI Systems and its descendents). I was fresh out of graduate school at the time, and I thought the cost of living was crazy then.
I currently live in the Washington DC area. Strangely enough, DC and SV do have something in common: if you live there, watch local TV news, and read the {Post, Mercury/News}, you gradually become convinced that {government/high-tech} is the most important thing in the world, and that you know more about it than other deprived people who don't live where you do. When you leave the area, you have to decompress a bit and realize that other things matter, too.
... learn Common Lisp. Seriously. XML is really just S-expressions (in drag, with redundant, overblown syntax - see here; (sorry about the pdf)), so why not use a language designed for munging them?
Perl and XML don't really get along well, primarily because Perl and arbitrarily nested data structures don't really get along well (see here if you want a less biased, but still discouraging opinion).
Strangely enough, Java and XML aren't getting along as well as one would think, if this article on the JDOM project is any indication. Java also has the shifting sands problem - vendor-controlled standards are evil, no matter who the vendor is.
Common Lisp has an ANSI standard that hasn't changed since 1995, open-source multi-threaded web servers that you can add native code to on the fly (see AllegroServe or CL-HTTP) and a lot of other good stuff that I don't have time to list here.
2) If campaign money is speech (Buckley vs Valejo!) then my voice is being drowned out by the roar of corporate cash. Let's investigate public financing so that we know in advance who has bought the candidates - us!
Public financing might reduce the corporate cash problem, but it would increase the two-parties-only problem. The real bugs in our current system (IMHO, of course) are:
Corporations are people under the law, and thus have most of the rights of people by default, including freedom of speech. Corporations became people in the US about a hundred years ago (brain lock, case reference missing). This was a big mistake which is going to be hell to correct.
Political campaigns need big money to buy media time. Broadcast media is expensive, but it is essentially impossible to win any major office without it... today.
One possible way out is the medium you are using right now. The Internet can have arbitrarily wide reach, and is unbelievably cheap for its reach compared to broadcast media. With any luck, a few years from now when everyone is connected, someone will conduct a Net-only campaign with the following two major platform planks:
I communicate with you voters via the Net; I don't need an enormous campaign fund.
My sleazy opponents use expensive, deceptive media ads, and are owned by the people who donate to them.
The only problem with this is that people might start questioning the validity of any expensive media operation designed to influence them. Death of advertising, anyone?
... unless you can afford to check into the Betty Ford Clinic.
If you're rich, drugs can apparently be just another interesting experience; if they get you into trouble (and you or your friends get you wrung out) they can be used later in life as an example of character building in the face of adversity.
If you're poor, you don't have as far to fall, and once you've hit bottom you don't have as many resources to climb back out.
Slashdot Mirror
Yes there must be an audit trail, yes machine-generated paper is good, but NO handing out receipts that show how you voted.
The problem is the possibility of vote coercion - your boss/landlord/relative says they'll do something nasty to you unless you bring them a receipt from the polls that shows you voted the way they want you to. It doesn't matter whether the receipt has your name on it or not, as long as it shows how you voted, and it's hard to fake (and if it's easy to fake, it's not a useful receipt).
The real problem is that evolved hardware can specialize itself very precisely for its environment, and the environment includes everything.
Some of Thompson's early evolved designs only worked in a narrow temperature range, because that's what they evolved in (see here for the article that prompted Slashdot to look at this last time).
Company X states that patents A, B, ... will be licensed at no-cost to anyone whose code using those patents is covered by any of the free/open-source licenses FooL, BarL, ...
Something like this would pull a major thorn out of the community's paw - the threat of legal action for independent rediscovery of patented algorithms.
A company that did this could still use its patents to beat up on other closed-source companies - they would just be saying they don't want to use the same tactics against open-source projects.
... the 50-foot high Teletype machine for a console?
... because the presence or absence of links to opposing viewpoints is a useful heuristic as to the sanity of the site.
Sites that have opposing viewpoint links are saying that they are confident enough in their views (and your intelligence) that pointing to the opposition only makes their case stronger. It's also an excellent way to kick sand in the face of the opposition - "we link to them, but they're not brave enough to link to us".
Sites that don't have opposition links are saying the opposite - that they don't trust you to agree with them, that they think the opposition might fool you despite their best efforts, that they're not 100% confident in their arguments.
It will be worth people's time to write real compilers for (Perl/Python/Ruby/whatever) when there are real, non-platform-specific specifications of what the languages mean, so compiler writers have something to aim at that won't move out from under them.
This is the fundamental reason why languages defined by portable implementations have problems growing beyond those implementations. The implementation actually specifies too much of the langauge's behavior - a real spec says explicitly what behavior is implementation-dependent, which tells implementors and users both exactly where the boundaries are.
The Common Lisp and Scheme communities wrote real specs for their languages (ANSI and IEEE, respectively). The user communities routinely beat up implementors and vendors who don't conform to the spec. As a result, they can write portable programs that compile to native code, without getting stuck in the C/C++ quagmire.
Someday there may be a way to answer these questions:
Are JPerl and Perl the same language?
How about JPython and Python?
that doesn't boil down to "ask Larry/Guido". A year or two after that, you might see real compilers for those languages.
1. Installing the patch(es) once is not enough. When Windows pollutes its environment enough so you have to reinstall it, you have to reinstall the patches, too, which means you have to be organized about downloading them, putting them somewhere safe and easy to find, installing them all in order, and cleaning them out as service packs come along. I suspect most users would prefer to live in denial ("I don't really need that").
2. How can we trust Microsoft/whoever to not add extraneous stuff in patches? I would object if they decided to issue a security patch that also upgraded, say, your DNS service to work better with Microsoft servers. I want to be able to choose whether or not to participate in the latest embrace&extend maneuver, but with closed-source patches, there's no way to tell.
... but like the man said, it was possible to hack VM into something almost useable. At Clemson in the mid 70's we had a home-grown edit/batch-submit/output-view system that was about as interactive as you could get with 3270 terminals. The editor could squirt numbered source lines onto your screen, you could edit them there, squirt the screen back, and have your changes accepted in one gulp.
Depressingly like filling out forms in a web browser...
DS1 is running an RS/6000 (IBM's version of the PowerPC architecture) at something like 25 MHz. That was the fastest rad-hardened chip they could get when they started the design.
The operating system is RTWorks, a Unix-like realtime kernel. The system software is written in C and Common Lisp (for the Autonomous Remote Agent stuff).
For details, check out http://nmp.jpl.nasa.gov/ds1/
How about something like a copyleft for patents, in which the holder of a patent states that they will give a no-cost license for its use in any application covered by, say, the GPL? The holder could also grant licenses to non-free software creators in the usual way.
This strikes me as a middle ground between the current situation (patent something and use it to beat everyone senseless) and no patents at all. A patent holder could use it as an advantage over non-free software, while leaving the free/open-source communities alone.
... the day someone successfully patents the first legal argument. When the lawyers realize that their bread and butter has many of the same "patentable" characteristics as algorithms, they just might get a clue about the current situation.
Alternatively, someone might try to defend themselves in court by saying that it is impossible to prove two algorithms are equivalent (because doing so is equivalent to solving the Halting Problem), so you can't really prove infringement. This is of course BS in this context (you can often prove equivalence for a particular pair of algorithms - it's the general case that steps in the Halting Problem), but it might confuse things enough to get you out of court.
it's an accurate description of the conditions and politics that create impossible project schedules.
The reason you should read it despite the source is that most of the good stuff in it is text Yourdon got from other people via email surveys.
Doing this would create an enormous market (the Government buys a bunch of software) for open formats, which would hit Microsoft right in its upgrade-treadmill solar plexus.
Check out the liability caps in the FCRA and the FDCPA - $500,000.00 is nothing to these guys. That makes it harder, but not impossible. Just file multiple suits, each with a group of, say, 100 people asking for $5000 in damages each. Even doing this once per state works out to $25M, plus legal fees and hassle.
Unfortunately, those denied credit are generally those who can't afford to launch a legal attack against a well funded opponent. When a large entity screws a lot of people over, the normal remedy is a class action lawsuit. The net should help here, as it will help the screwees find each other and get organized. Surely some lawyer somewhere has smelled blood in the water from this problem...
Does anyone know exactly what jobs the people who are suing held at Microsoft? I haven't seen a list in any of the media reports.
I can believe the people who say MS policy is "we hire hackers based on ability - race is irrelevant". This is probably true for the technical positions (or as true as it can get in America).
BUT, MS is a big company, and has lots of jobs that are less technical. I could easily believe racism is alive and well in sales, marketing, packaging design, legal, etc...
Could't Intel argue they will be using 50% less power in 18 months or so?
Only in the sense that they'll be producing twice as many transistors per kilowatt-hour.
1. The people who play games
2. The people who write games
Yeah, yeah #2 is pretty much a subset of #1, but you get the idea.
but this is a giant program to keep Russian Aerospace engineers employed.
The ISS has always (at least from a US budgetary point of view) been about keeping American aerospace engineers employed. This has been NASA's primary reason for existence for many years now... at least according to Congress.
If NASA was supposed to be about science and aero/astronomical research, maybe we'd have a budget for them instead of for relatively useless stuff like the ISS.
When anyone asks me what it's like in Silicon Valley, that's what I tell them. I lived there from '86 - '88 (working for somewhat-lamented Xerox AI Systems and its descendents). I was fresh out of graduate school at the time, and I thought the cost of living was crazy then.
I currently live in the Washington DC area. Strangely enough, DC and SV do have something in common: if you live there, watch local TV news, and read the {Post, Mercury/News}, you gradually become convinced that {government/high-tech} is the most important thing in the world, and that you know more about it than other deprived people who don't live where you do. When you leave the area, you have to decompress a bit and realize that other things matter, too.
... learn Common Lisp. Seriously. XML is really just S-expressions (in drag, with redundant, overblown syntax - see here; (sorry about the pdf)), so why not use a language designed for munging them?
Perl and XML don't really get along well, primarily because Perl and arbitrarily nested data structures don't really get along well (see here if you want a less biased, but still discouraging opinion).
Strangely enough, Java and XML aren't getting along as well as one would think, if this article on the JDOM project is any indication. Java also has the shifting sands problem - vendor-controlled standards are evil, no matter who the vendor is.
Common Lisp has an ANSI standard that hasn't changed since 1995, open-source multi-threaded web servers that you can add native code to on the fly (see AllegroServe or CL-HTTP) and a lot of other good stuff that I don't have time to list here.
According to Google:
http://www.tourolaw.edu/patch/CaseSummary.html
- Corporations are people under the law, and thus have most of the rights of people by default, including freedom of speech. Corporations became people in the US about a hundred years ago (brain lock, case reference missing). This was a big mistake which is going to be hell to correct.
- Political campaigns need big money to buy media time. Broadcast media is expensive, but it is essentially impossible to win any major office without it
... today.
One possible way out is the medium you are using right now. The Internet can have arbitrarily wide reach, and is unbelievably cheap for its reach compared to broadcast media. With any luck, a few years from now when everyone is connected, someone will conduct a Net-only campaign with the following two major platform planks:- I communicate with you voters via the Net; I don't need an enormous campaign fund.
- My sleazy opponents use expensive, deceptive media ads, and are owned by the people who donate to them.
The only problem with this is that people might start questioning the validity of any expensive media operation designed to influence them. Death of advertising, anyone?... unless you can afford to check into the Betty Ford Clinic.
If you're rich, drugs can apparently be just another interesting experience; if they get you into trouble (and you or your friends get you wrung out) they can be used later in life as an example of character building in the face of adversity.
If you're poor, you don't have as far to fall, and once you've hit bottom you don't have as many resources to climb back out.