We're currently using a cylindrical patch antenna on the rocket, and a semi-directional antenna (helical) on the ground. Eventually, we'll build a tracking dish for the ground, and try to phase the rocket antenna to get the antenna pattern to point more down. Right now the rocket antenna pattern is a sort of "donut", which is fairly suboptimal.
Microwave antenna design is hard. If any gurus want to contribute expertise, please drop PSAS a note.
You can set up your own rule weights for SA. See the documentation. I have all the porn features turned way up, and it works much better for my mail stream.
Everyone is smart enough not to use [soundex] for names like "Juan Garza"...
Huh? Soundex was invented for the US Census back in the days of Ellis Island and name changes. What is broken about "Juan Garza", exactly?
Soundex is a very simple algorithm, but performs surprisingly well. A long time ago, some of my undergraduate colleagues implemented a mispelled word suggestion program using it: I recently re-implemented it, and it works well enough to use. That said, anyone using it as their sole key is insane.
Recall the case of Victorias Secret vs. the mom-and-pop Victors Little Secret. The case was trademark related, Victorias Secret claimed VLS was hurting their profits and sales by having a similar name. I dont know what the intermediate rulings in the case were, but each hearing VLS asked evidence be presented of actual harm. The case eventually got before the Supreme Court, who ruled that in trademark cases, the victim (ie, plaintiff) must show evidence of actual harm being done.
This is your evidence that the court system is even-handed? So Victoria's sues VLS, takes them over the course of years through several rounds of litigation, and VLS finally prevails. Hooray for the little guy. Who knows, they may have even been reimbursed for the legal fees and court time that they incurred---years later. If VLS is still in business, they are lucky. Further, there's little to stop Victoria's or someone else from just walking through it again on some slightly different pretext until VLS is dead.
The fact of the matter is that in the US, unless you have deep pockets, you lose de facto at the moment you have to start paying lawyers. Until this situation is remedied (how?), our civil system is primarily a tool for heavily monied interests to muscle out the less-financially-endowed.
The "obvious" answer would have been to use FreeS/WAN or similar to set up an IPSEC tunnel to your wired network and be done with it. Windows supports IPSEC as well, and it seems like it would solve most of your problems.
Am I missing something?
Word on the street is that HB 2892 may be revived via a rarely-used legislative maneuver. An existing unrelated bill, SB 589, scheduled for a floor vote, may be amended by removing the text of the existing SB 589 and replacing it with the text of HB 2892. I don't know any official cite for this information
yet, but I have heard it from "reliable sources". It's not a sure thing, but there's still a slight hope...
As near as I can tell from the popular articles, and the web page referenced in the New Scientist article, "Fast TCP" is not a new protocol, but rather better congestion control for standard TCP. I'm not a network guru by any means, so please take the comments below with a grain of salt.
Currently, TCP implementations use a standard trick to play nice with small router queues. Using precise timing would be better. I hassled Mike Karels over it about 10-15 years ago, but the consensus at the time was that the hardware wasn't up to it. Now it is.
Also, modern routers have gotten clever about queue management, which screws up the trick.
The new proposal is to take advantage of modern HW to measure latencies. Existing TCP could thus be used more efficiently, by allowing larger amounts of data to be outstanding on the network without trashing routers.
It is not widely understood that in 1988 the Internet DOSed itself because of a protocol design issue, and Van Jacobsen got everybody to fix it by a consensus change to the reference implementation of TCP. These articles appear to report (badly) ongoing research into that issue.
A couple of things I might add that haven't been mentioned so far, I think:
Norstrilia and/or the collected short stories of Cordwainer Smith (Henry Linebarger). This is an amazing writer whose stories are literally 50 years ahead of their time in so many ways. His Alpha Ralpha Boulevard has been anthologized about a bajillion times, yet most folks don't seem to have heard of him.
The Illuminatus! Trilogy of Robert Anton Wilson. Very juvenile, silly stuff, but there's nothing quite like it out there, and it's definitely a geek read.
The Cyberiad. Stanislaw Lem's greatest work IMHO.
There's plenty of more room, of course. If you haven't read Hofstadter's classic Godel, Escher, Bach: an Eternal Golden Braid, for example, you simply must. But I think we've given enough to get started on:-).
Fastest/.ing ever. What a pain. I'm seriously tired of this.
Has anyone thought seriously about creating a proxy site for/. that automatically caches all the links? I think that if the site and caching was done as a proxy, there would be no copyright issues?
Heck, if anyone wants to put together a demo of this, I'd be willing to host it on an experimental basis. I've got access to a ton of free BW through my academic institution for non-profit stuff. My theory is that if the proxy becomes popular, we can bully VA into hosting it, like they should have done for the past several years...
Alternatively, and maybe even cooler, one could build a proxy that browses through the Google cache, to save bandwidth and storage and further blur the copyright question.
Whatever Sun's fate, Mr Schwartz is probably right that the software industry will not be taken over by free programs, as some geeks would like. The main attraction of open source, as he says, is the fact that it is "great for innovation", not its questionable claim to be free.
"Questionable claim to be free"? Let's leave aside "free/Free" for a moment, as the author seems to indicate the former. (Let's also leave aside the grammatical correctness of the sentence, which looks more like it belongs in a/. article than in The Economist.:-)
Instead, let's ask what this "questionable claim" actually is. Hmm, does open source software have a purchase price? Not really: by definition, it costs $0. How about technical support, is that free? Why, for most open source it is: extensive online help, rapid bugfixes, etc. I know, are any and all costs related to its use zero? Why no, they are not---you still have to pay to field the software and maintain it.
If you told the author of this article you were giving him a free car, with a free warranty for parts and a substantial discount on labor, apparently his response would be "Oh yeah? What about gas?". Sheesh.
Although, the article was pretty well-written otherwise:-).
There also seems to be a great deal of political contention between supporters of 802.11g and 802.11a. As a result, some of the articles posted on the Internet discussing 802.11g and/or 802.11a may have a political slant---indeed, may contain untruths.
Fair enough. Have to say, though, that they don't seem to feel sorry for themselves. My courses continue to be popular, and I have a lot of students working unfunded to learn from me. I guess I believe my approach is working pretty well, or I would change it. But I could be wrong.
The ECG measurement is a key piece of the standard polygraph. One of these can
be combined with a galvanometer (easy), a skin thermometer (easy), and possibly a respiratory rate measurement (harder: standard technique is to wrap the chest with a stress-sensitive band and build a circuit similar to the ECG one) for a lie detector that should be great fun at parties.
I'm an "Absent", but it's a variant known as "make an e-mail appointment". I explicitly tell students up front, "I have no office hours. If you want to talk to me, I'm happy to make an appointment to meet with you almost anytime."
I do encourage long-winded homework questions, although I sometimes don't get around to answering them:-).
Positive consequences: the student can count on my being there; e-mail questions, once stated, often can be resolved without a meeting; I can work around student schedules; I am not trapped in an empty office for no good reason. Negative consequences: some students have a hard time dealing with this approach; there's little economy of scale where several students can be helped at once. Overall, the approach probably works better for the smaller advanced classes I teach than it would for large lecture courses.
Want an "interesting mix of imperative and functional features" in a language any C/C++/Java programmer can start working with immediately? Try Nickle.
If you run a hard line, it is best to use glass instead of copper. Otherwise, the effects of a lightning strike can be pretty dramatic: the lightning tends to seek out shallow runs of copper (although it could care less about glass). If you must run copper, use fancy isolators on both ends to minimize the danger to folks and equipment.
x++ is idiomatic because C is zero-based. Allow me to expand on that statement.
One of the most common uses of the increment operators in C-like languages is in array indexing. Consider for example a simple integer stack:
int stk[3]; int sp = 0;
void push(int s[], int v) { s[sp++] = v; }
int pop(int s[]) { return s[--sp]; }
This is the inspiration for post-increment and pre-decrement: if I automagically post-increment and pre-decrement, the array is usually accessed properly.
Similarly, consider the usual
backward-counting while loop:
int i = 3; while(--i >= 0) { stk[i] = 0; }
(Of course, the forward-counting case cannot be written naturally with pre- or post-increment in this style; hey, no idiom is perfect:-), but if you use an increment operator inside the while body, it should be a post-increment of the last reference to i.)
One should do what one wants, but following the post-increment pre-decrement idiom may help improve code quality.
(Excuse the code formatting. Why can't I indent in <ecode>?)
Nobody got outraged when that new-fangled mechanical auto-mobile contraption started to outpace the world's fastest human runners.
Allow me to recommend to you the legend of John Henry. About the time period you mention, too. I always mention this story in the Intro AI class I teach.
Heh. Back in the day, I spent about 8 hours playing multiplayer air combat with acm on an SGI Onyx system (a predecessor of the
Onyx 3000), with its incredibly-realistic-for-the-time 3D rendering and physics.
On the drive home, I found myself needing to cross 4 lanes of traffic to make a light. Without thinking, I spotted a small opening, stepped on the gas, and floored it, squeezing through quite nicely. Then I realized what I had done.
As a happy Debian user, it seems to me like Petreley is mostly solving a non-problem. Debian has its problems, but "software is hard to install correctly" is rarely one of them. I think that Petreley's suggestions are actually way more germane to simplifying Windows software installation. "DLL hell" got its name there for a reason.
Petreley seems to not understand the model that has evolved for shared resources (libraries, data, etc.) in Linux. He is correct that nobody cares about the disk space anymore: what they care about is that they can now upgrade functionality without reinstalling packages, as long as they preserve interfaces. I can install a new libc, xine back-end, Mozilla icon, etc., and if I'm careful every app on my box will see the new version without further action. That's a big feature, as long as I'm careful to re-install all the apps if the interface to the newly-installed thing changes. Fortunately, apt gets this case mostly right. I wish they had a better scheme than putting major version numbers in package names, but I guess I can live with it.
I used to run Red Hat everywhere. I now think Debian-based distros such as Knoppix will win in the long run. They're getting to be no harder to install, and you only have to do it once per box: when you want to upgrade to current or install new bits you just push a button and it happens. What could be easier?
As someone who has done actual research on Oregon HB 2892 (reading the bill and talking to its sponsor), I'd like to try to clear up some misconceptions with a short FAQ:
Q: What did HB 2892 do?
A: Two things: Require state agencies to (1) consider open source in procurement, and (2) procure only software that supports open formats for data storage and interchange.
Q: Why the past tense in the previous question?
A: Because it appears that the language of the bill has been compromised to increase its chance of passage. I haven't yet looked for the revised wording.
Q: Why is (1) necessary---can't state agencies consider open source anyhow?
A: According to the bill's sponsor, the nature of state procurement rules makes open source procurement difficult. Because there is no sponsoring organization that will bid contracts for typical open source alternatives, agencies may be bound by law or regulation to ignore them. (1) changes that.
Q: Doesn't the language of (1) force open software on state agencies?
A: No, it forces them to consider it. In a plain reading of the rules, a state agency should be prepared to explain why it selected a particular package over open source alternatives. HB 2892 has no detailed description of the criteria or methods of consideration.
Q: Do state agencies use a lot of open source anyhow?
A: Yes. Agencies that already use open source software generally support the bill: see above.
Q: Is (1) the most important part of the bill?
A: No, both provisions (1) and (2) are important. Perhaps the chief concern of the bill's sponsor, Rep. Barnhart, is legacy systems and lock-in. (1) addresses this issue by encouraging continuously-maintainable systems. (2) addresses the issue by allowing seamless replacement of systems.
So, if I'm Mr. Burns and I find these things crawling around on the land around my powerplant, what's stopping me from sending Smithers out to pick them up and throw them in the lake?
Heh. When my doctoral advisor was at Stanford, Shakey the Robot was there, and always had a crowd of graduate students following it around. My advisor got tired of this, and told the folks running Shakey: "If Shakey wanders into my office, it's not coming out."
The operators put foil tape in front of my advisor's office door, and a photosensor in the base of Shakey, to guard against this calamity.
Slashdot Mirror
Yes. See the documentation.
We're currently using a cylindrical patch antenna on the rocket, and a semi-directional antenna (helical) on the ground. Eventually, we'll build a tracking dish for the ground, and try to phase the rocket antenna to get the antenna pattern to point more down. Right now the rocket antenna pattern is a sort of "donut", which is fairly suboptimal.
Microwave antenna design is hard. If any gurus want to contribute expertise, please drop PSAS a note.
You can set up your own rule weights for SA. See the documentation. I have all the porn features turned way up, and it works much better for my mail stream.
Everyone is smart enough not to use [soundex] for names like "Juan Garza"...
Huh? Soundex was invented for the US Census back in the days of Ellis Island and name changes. What is broken about "Juan Garza", exactly?
Soundex is a very simple algorithm, but performs surprisingly well. A long time ago, some of my undergraduate colleagues implemented a mispelled word suggestion program using it: I recently re-implemented it, and it works well enough to use. That said, anyone using it as their sole key is insane.
Recall the case of Victorias Secret vs. the mom-and-pop Victors Little Secret. The case was trademark related, Victorias Secret claimed VLS was hurting their profits and sales by having a similar name. I dont know what the intermediate rulings in the case were, but each hearing VLS asked evidence be presented of actual harm. The case eventually got before the Supreme Court, who ruled that in trademark cases, the victim (ie, plaintiff) must show evidence of actual harm being done.
This is your evidence that the court system is even-handed? So Victoria's sues VLS, takes them over the course of years through several rounds of litigation, and VLS finally prevails. Hooray for the little guy. Who knows, they may have even been reimbursed for the legal fees and court time that they incurred---years later. If VLS is still in business, they are lucky. Further, there's little to stop Victoria's or someone else from just walking through it again on some slightly different pretext until VLS is dead.
The fact of the matter is that in the US, unless you have deep pockets, you lose de facto at the moment you have to start paying lawyers. Until this situation is remedied (how?), our civil system is primarily a tool for heavily monied interests to muscle out the less-financially-endowed.
The "obvious" answer would have been to use FreeS/WAN or similar to set up an IPSEC tunnel to your wired network and be done with it. Windows supports IPSEC as well, and it seems like it would solve most of your problems. Am I missing something?
Word on the street is that HB 2892 may be revived via a rarely-used legislative maneuver. An existing unrelated bill, SB 589, scheduled for a floor vote, may be amended by removing the text of the existing SB 589 and replacing it with the text of HB 2892. I don't know any official cite for this information yet, but I have heard it from "reliable sources". It's not a sure thing, but there's still a slight hope...
As near as I can tell from the popular articles, and the web page referenced in the New Scientist article, "Fast TCP" is not a new protocol, but rather better congestion control for standard TCP. I'm not a network guru by any means, so please take the comments below with a grain of salt.
Currently, TCP implementations use a standard trick to play nice with small router queues. Using precise timing would be better. I hassled Mike Karels over it about 10-15 years ago, but the consensus at the time was that the hardware wasn't up to it. Now it is. Also, modern routers have gotten clever about queue management, which screws up the trick.
The new proposal is to take advantage of modern HW to measure latencies. Existing TCP could thus be used more efficiently, by allowing larger amounts of data to be outstanding on the network without trashing routers.
It is not widely understood that in 1988 the Internet DOSed itself because of a protocol design issue, and Van Jacobsen got everybody to fix it by a consensus change to the reference implementation of TCP. These articles appear to report (badly) ongoing research into that issue.
Oregon's HB 2892 has sections covering both open source and open formats. It has recently been revived, and there's still a chance of passage.
A couple of things I might add that haven't been mentioned so far, I think:
- Norstrilia and/or the collected short stories of Cordwainer Smith (Henry Linebarger). This is an amazing writer whose stories are literally 50 years ahead of their time in so many ways. His Alpha Ralpha Boulevard has been anthologized about a bajillion times, yet most folks don't seem to have heard of him.
- The Illuminatus! Trilogy of Robert Anton Wilson. Very juvenile, silly stuff, but there's nothing quite like it out there, and it's definitely a geek read.
- The Cyberiad. Stanislaw Lem's greatest work IMHO.
There's plenty of more room, of course. If you haven't read Hofstadter's classic Godel, Escher, Bach: an Eternal Golden Braid, for example, you simply must. But I think we've given enough to get started onFastest /.ing ever. What a pain. I'm seriously tired of this.
Has anyone thought seriously about creating a proxy site for /. that automatically caches all the links? I think that if the site and caching was done as a proxy, there would be no copyright issues?
Heck, if anyone wants to put together a demo of this, I'd be willing to host it on an experimental basis. I've got access to a ton of free BW through my academic institution for non-profit stuff. My theory is that if the proxy becomes popular, we can bully VA into hosting it, like they should have done for the past several years...
Alternatively, and maybe even cooler, one could build a proxy that browses through the Google cache, to save bandwidth and storage and further blur the copyright question.
"Questionable claim to be free"? Let's leave aside "free/Free" for a moment, as the author seems to indicate the former. (Let's also leave aside the grammatical correctness of the sentence, which looks more like it belongs in a /. article than in The Economist. :-)
Instead, let's ask what this "questionable claim" actually is. Hmm, does open source software have a purchase price? Not really: by definition, it costs $0. How about technical support, is that free? Why, for most open source it is: extensive online help, rapid bugfixes, etc. I know, are any and all costs related to its use zero? Why no, they are not---you still have to pay to field the software and maintain it.If you told the author of this article you were giving him a free car, with a free warranty for parts and a substantial discount on labor, apparently his response would be "Oh yeah? What about gas?". Sheesh.
Although, the article was pretty well-written otherwise :-).
I'll try:
Fair enough. Have to say, though, that they don't seem to feel sorry for themselves. My courses continue to be popular, and I have a lot of students working unfunded to learn from me. I guess I believe my approach is working pretty well, or I would change it. But I could be wrong.
The ECG measurement is a key piece of the standard polygraph. One of these can be combined with a galvanometer (easy), a skin thermometer (easy), and possibly a respiratory rate measurement (harder: standard technique is to wrap the chest with a stress-sensitive band and build a circuit similar to the ECG one) for a lie detector that should be great fun at parties.
I'm an "Absent", but it's a variant known as "make an e-mail appointment". I explicitly tell students up front, "I have no office hours. If you want to talk to me, I'm happy to make an appointment to meet with you almost anytime." I do encourage long-winded homework questions, although I sometimes don't get around to answering them :-).
Positive consequences: the student can count on my being there; e-mail questions, once stated, often can be resolved without a meeting; I can work around student schedules; I am not trapped in an empty office for no good reason. Negative consequences: some students have a hard time dealing with this approach; there's little economy of scale where several students can be helped at once. Overall, the approach probably works better for the smaller advanced classes I teach than it would for large lecture courses.
Want an "interesting mix of imperative and functional features" in a language any C/C++/Java programmer can start working with immediately? Try Nickle.
From his comments, Chromatic would like Nickle. But I'm the co-designer---I have to say that :-).
If you run a hard line, it is best to use glass instead of copper. Otherwise, the effects of a lightning strike can be pretty dramatic: the lightning tends to seek out shallow runs of copper (although it could care less about glass). If you must run copper, use fancy isolators on both ends to minimize the danger to folks and equipment.
x++ is idiomatic because C is zero-based. Allow me to expand on that statement.
One of the most common uses of the increment operators in C-like languages is in array indexing. Consider for example a simple integer stack:
This is the inspiration for post-increment and pre-decrement: if I automagically post-increment and pre-decrement, the array is usually accessed properly.Similarly, consider the usual backward-counting while loop:
(Of course, the forward-counting case cannot be written naturally with pre- or post-increment in this style; hey, no idiom is perfectOne should do what one wants, but following the post-increment pre-decrement idiom may help improve code quality.
(Excuse the code formatting. Why can't I indent in <ecode>?)
Nobody got outraged when that new-fangled mechanical auto-mobile contraption started to outpace the world's fastest human runners.
Allow me to recommend to you the legend of John Henry. About the time period you mention, too. I always mention this story in the Intro AI class I teach.
Heh. Back in the day, I spent about 8 hours playing multiplayer air combat with acm on an SGI Onyx system (a predecessor of the Onyx 3000), with its incredibly-realistic-for-the-time 3D rendering and physics.
On the drive home, I found myself needing to cross 4 lanes of traffic to make a light. Without thinking, I spotted a small opening, stepped on the gas, and floored it, squeezing through quite nicely. Then I realized what I had done.
Resolved: remember that I don't have bonus lives.
As a happy Debian user, it seems to me like Petreley is mostly solving a non-problem. Debian has its problems, but "software is hard to install correctly" is rarely one of them. I think that Petreley's suggestions are actually way more germane to simplifying Windows software installation. "DLL hell" got its name there for a reason.
Petreley seems to not understand the model that has evolved for shared resources (libraries, data, etc.) in Linux. He is correct that nobody cares about the disk space anymore: what they care about is that they can now upgrade functionality without reinstalling packages, as long as they preserve interfaces. I can install a new libc, xine back-end, Mozilla icon, etc., and if I'm careful every app on my box will see the new version without further action. That's a big feature, as long as I'm careful to re-install all the apps if the interface to the newly-installed thing changes. Fortunately, apt gets this case mostly right. I wish they had a better scheme than putting major version numbers in package names, but I guess I can live with it.
I used to run Red Hat everywhere. I now think Debian-based distros such as Knoppix will win in the long run. They're getting to be no harder to install, and you only have to do it once per box: when you want to upgrade to current or install new bits you just push a button and it happens. What could be easier?
As someone who has done actual research on Oregon HB 2892 (reading the bill and talking to its sponsor), I'd like to try to clear up some misconceptions with a short FAQ:
Q: What did HB 2892 do?
A: Two things: Require state agencies to (1) consider open source in procurement, and (2) procure only software that supports open formats for data storage and interchange.
Q: Why the past tense in the previous question?
A: Because it appears that the language of the bill has been compromised to increase its chance of passage. I haven't yet looked for the revised wording.
Q: Why is (1) necessary---can't state agencies consider open source anyhow?
A: According to the bill's sponsor, the nature of state procurement rules makes open source procurement difficult. Because there is no sponsoring organization that will bid contracts for typical open source alternatives, agencies may be bound by law or regulation to ignore them. (1) changes that.
Q: Doesn't the language of (1) force open software on state agencies?
A: No, it forces them to consider it. In a plain reading of the rules, a state agency should be prepared to explain why it selected a particular package over open source alternatives. HB 2892 has no detailed description of the criteria or methods of consideration.
Q: Do state agencies use a lot of open source anyhow?
A: Yes. Agencies that already use open source software generally support the bill: see above.
Q: Is (1) the most important part of the bill?
Hope this helps.A: No, both provisions (1) and (2) are important. Perhaps the chief concern of the bill's sponsor, Rep. Barnhart, is legacy systems and lock-in. (1) addresses this issue by encouraging continuously-maintainable systems. (2) addresses the issue by allowing seamless replacement of systems.
So, if I'm Mr. Burns and I find these things crawling around on the land around my powerplant, what's stopping me from sending Smithers out to pick them up and throw them in the lake?
Heh. When my doctoral advisor was at Stanford, Shakey the Robot was there, and always had a crowd of graduate students following it around. My advisor got tired of this, and told the folks running Shakey: "If Shakey wanders into my office, it's not coming out."
The operators put foil tape in front of my advisor's office door, and a photosensor in the base of Shakey, to guard against this calamity.
Or so I've been told...