I think it changes with age. When I was quite young, 10 hour a day for days on end wasn't so hard to pull off. Remembering to sleep and shower and brush my teeth were harder. The catch was that a very high percentage of the code I'd write was either pure crap, or could have been done better in less time by writing another program to write that code. As I've gotten older, I've found that it's easier to spend a large number of hours *contemplating* code -- but hard to work other than in smaller bursts actually writing the code. The difference is that when I do write code, the hours are far less wasted.
I've taken this into account and so now my plan is, that when I reach 90, I'll just wake up in the morning and fart. My heavily customized Emacs will analyze the fart and translate it into C. "Oh, boy, I wrote another new OS kernel this morning!"
Well, ok, one of the two above paragraphs is true and not the other.
The engineering authority in charge of the bridge and repairs already gave their answer to this on the morning news (yesterday, I think):
They found the crack. They designed the "band-aid": the saddle, T-bar, rods, etc. They had it fabricated and installed.
In subsequent days, they went back up to look at how it was doing. They found that it was vibrating more than they thought it should: it wasn't as rigid as it was designed to be. They recognized that this would lead to fatigue and failure.
They began designing the improved "band-aid" and planned to install it sometime in coming weeks.
To their surprise, *perhaps* related to unusually high winds, the system failed sooner than they thought it could.
The completed their improved design and are now installing it. (And they are counting their blessings that nobody was killed: they got lucky, that way.)
Cinematic reference to Blade Runner. I was talking about geezer knowledge so it seemed apropos.
The full quote begins with something about how "I've seen things you... humans... will never see..." and continues about lost / underappreciated learning.
I said positive income spin-offs. I don't mean spin-offs that immediately pay for the core R&D activity - only spin-offs that if you regard them as a separate business unit, make sense on their own.
Yes, that does put a kind of pressure on the research. Yes it does distort what kind of research takes place. I think it does both of those in a good an appropriate way for an R&D lab that is a long-term, for-profit investment.
One way that it helps is that it helps the lab develop the personal relationships and the skills needed for technology transfer to those who will operate something closer to customer-facing. That skill set comes in handy when the really big R&D wins come through - so you don't wind up like Xerox PARC in 1980.
Another way it helps is the reverse direction of technology transfer - what we could call "market awareness transfer" from the closer-to-the-front business guys back to the lab. There are gazillions more things that make for interesting research than make for research that is promising for business and, insisting on at least little spin-offs early is a way to help keep the lab folks in touch with business reality.
You misunderstand. By "managing wealth" I very much include not leaving *too* much of a legacy for kids, making sure as little as possible goes towards evil, and getting as much of the surplus doing good works. Buffet is schematically the right idea here, even if I don't agree with all of his particular decisions. My selfish thing is that I wouldn't want to spend 60 hours / week managing various investments. Nor would I want to just hand most of it over to the Gates foundation. $1B today, if you can make a lot of it liquid quickly, is -- I agree -- more than is reasonably needed. It's just a big responsibility and my selfish take is that I'd make a priority of reducing the amount of time I had to personally spend managing that responsibility. There are some causes I'd want a hands-on role in because I think I have intellectual contributions to make but there's a lot of grunt work in responsibly handling that large an amount of money/nominal wealth that I would want to delegate in order to concentrate on what I'm good at.
I've been programming for, like, uh.... about 27 or 28 years. Arguably longer if you wanna go back to really little kid stuff.
If I had that much money - was basically (if I wanted to be) in the leisure class - what I would like to believe about myself is that I would try to secure my family's material conditions really well, try to make as efficient as possible my wealth management program, and, as to hacking.....
There are *so many* really great and valuable potential projects that (a) nobody is investing in; (b) have an investment horizon that is tough because these are projects that will take a good 5 years, let's say, to get to where seeing a return is on the table. A good 10 years before you start to see the possibility of "done".
I would start an R&D lab but a very small one - perhaps 10 people - and while we'd try to have some positive income spin-offs each year from 0 onward, the goal would be to create the kind of environment where we can take off some of the bigger, long-neglected problems.
You kids these days don't know what's possible in a GUI framework. You don't know how to do language design, systems software generally, databases, file systems, or a whole lot of other basics. You've inherited really mediocre crap and you take for granted that that's where things are at. And the industry has ceased production of grey-beards. (Also: get off my lawn!)
"like tears in the rain",
-t
Re:I was "almost" a subject of this experiment
on
My Genome, My Self?
·
· Score: 1
Thank you, "Creepy Crawler", that's right. There's more to it, as well, but that's a big part of it.
-t
Re:I was "almost" a subject of this experiment
on
My Genome, My Self?
·
· Score: 1
If you want to put it in cold economic terms, the economic benefit of "insurance for life" (especially in the context of a nation that looks poised to implement national health care well before such a contract would expire) is greatly outweighed by the many exploitative possibilities of having your medical history and your genome exposed. The risk of those things extends to your progeny and your relatives, as well. For example, participation in the PGP experiment - especially as one of the first 10 - makes you a political symbol; an object. People can prove things to third parties by "doing things" to you (or your genetic material-sharing relatives). It's like painting a target on your back.
I also had deeper misgivings because of the scientific sloppiness I witnessed in my work at the Church lab. The only thing in genomics work that is worse than a rush towards the new eugenics is a rush towards the new scientifically false eugenics - which was the rush my supervisor's work was aimed towards when we parted ways. These guys are bad news. On a personal level, I did not find either my supervisor or boss to be people I could sustain even a presumption of respect for. They are, in my experience, sloppy in important ways (me, I try to be sloppy in the unimportant ways) - bad science, bad ethics, money-hungry, self-serving twits.
-t
I was "almost" a subject of this experiment
on
My Genome, My Self?
·
· Score: 2, Interesting
I used to work as a contractor for the George Church lab. My supervisor was a student of Church's. Church was his boss. I was working on bio-informatics (if anyone cares, I can tell you some tricks for regexp-searching of genetic sequences).
My family was under extreme financial duress. In light of that knowledge, my supervisor (tells me, at least) that he took my situation to George and they came up with this: "Sign up to be one of the first 10 PGP subjects. Give us all of your medical records from the past and into the future. Agree to have your sequence published. We think we can get Harvard to agree to pay for your medical insurance for life. Don't you think your family deserves for you to make that trade-off?"
I said, flatly, "no." I pointed out, among other problems, some severe technical problems in the line of sequencing research we working on. Ultimately, we (me and the lab) part ways on less than amicable terms after this.
I think these people are scum.
They were eager to exploit my poverty as leverage to make me a human subject to rather dangerous experimentation based on highly dubious scientific claims - and they punished me for dissenting from this plan, as nearly as I can tell.
Strive to make sure that if any one member of your team were to be, goodness forbid, struck by a bus -- that you could fill there job with at least barely enough efficiency to still have a chance.
Then deal with the business side as if that is the expected outcome.
If you lack the empathic capacity or technical smarts for this, quit your job and go back to programming.
It's not a panopticon.
In a panopticon, the threat of perpetual surveillance is used to enforce discipline.
In the contemporary situation, hidden surveillance is used to create points of control.
-t
...and so it begins. Not on the frontiers of outer space, not launched from Mars during the night...but here, on Slashdot. They have found how to infiltrate our minds and compel us to respond, waste our mod points, and upset the balance of society itself.
Well, yes. "They" [the political, economic, and social situation] have done just that. You have just done it by trivializing the concerns, reinforcing the hegemony that produces the problems Carr is noticing. You have illustrated a tiny sliver of the problem, by example. Good job.
Yes, exploited. It's a professional embarrasment at the start of his career. I'm not sure $25K is a fair price.
It's a basic logic problem. The original challenge problem could be restated: "Prove either that there exists a non-universal machine which emulates the 2-3 machine OR that the 2-3 machine can only be emulated by a universal machine." The proof does neither. Wolfram Inc. reps have come back with "Well, perhaps we should change the definition of universality!" Only, they aren't very concrete in offering an alternative with any rigor and the vague suggestions they are making don't add up (e.g., don't answer Vaughn Pratt's counter-example of paired push-down automata).
What the student proved here may turn out to be an interesting and useful result (not the universality of 2-3 but the universality of this interesting combination of machines). Students should be encouraged to work on such problems. Students should be encouraged to write as well as this student is learning to do -- it's a nicely presented paper (trivial formatting bugs aside). But students shouldn't be encouraged to go in those directions on false premises.
At least two interesting questions come from the students work. To Wolfram Inc.'s credit, they are pointing in the general direction of these new questions (even while not yet acknowledging their mistake). The new questions: Do there exist simple machines whose universality is undecidable (and might 2-3 be an example)? If 2-3's universality is either false or undecidable (and especially in the latter case) can we find any useful structure to what combinations of it with other machines clearly are universal?
I'll leave it as an exercise to figure out how raising those questions relate to the "Priciple of Computational Equivalence" in NKS but, meanwhile, leave the student out of it!
We'll see. With an additional step that "finitizes" the student's construction the proof is rescued and raised to the status of an important lemma -- but if that step isn't very quickly forthcoming, the prize -- in no small part an advertising vehicle -- was administered in a pedagogically misleading way.
I have read and followed all of the related threads on the FOM mailing list. I am also familiar with how restraint in expression works in academia vs. slashdot. Academic restraint of expression in the emails has led to some confusion on slashdot:
Vaughn Pratt's objection is indeed elementary and, if true, unassailable. Vaughn Pratt has pointed out that the proof says, essentially, "the 2 state, 3 color machine in question is universal -- if you add some more states".
Is that objection true? The recent responses from Wolfram Inc. have said, "yes, it is." They say so explicitly. Game over. Except wait:
Things have now turned obscene:
Wolfram Inc. first of all misled this student, badly when they advised him about the paper. Now he has the embarassment of being the victem of a controversy on his record.
Wolfram Inc. exploited this student badly. I have yet to see any public statement from Wolfram Inc. about the matter that does not seem to me to be primarily aimed at promoting the "NKS" book.
Wolfram the man has, in my opinion, embarassed himself badly. The elementary error is not only in the proof, it is in the rules of the contest. It is not written out in the rules of the contest -- the rules cite certain passages of the book, where the error is offered as fact.
The heat-of-the-moment responses from Wolfram Inc. have, clearly to anyone familiar with the dialects and who has read the threads, evaded the issue. In my opinion, they have offered handwaving (and self promotion) of the worst kind.
There is, as indirectly pointed out on the FOM mailing list, some chance that the 2-3 machine may yet turn out not only to not be provably universal, but perhaps even to be a clear refutation of the main thesis of NKS (as nearly as said thesis can be generously rendered as a coherent statement, in my opinion).
So, it's a freaking mess. In the standards of the academic world, this is a 100-year flood kind of boner of a mistake, to put it mildly.
I call upon Wolfram and Wolfram Inc. to own up, make a serious effort to articulate their critic's best arguments in plain language accessible to the slashdot crowd, and then explain how they will proceed justly from there. Otherwise, in my opinion, they will just be furthering their exploitation of others.
-t
p.s.: my best guess is that the lame responses from Wolfram Inc. are there to buy time -- they want 48-96 hours to rack their brains and rescue the proof. And there's a chance! Not a great chance, but a chance. For example, if it turns out that encoding the student's construction in a meta-circular interpreter (writing a program to cause it to simulate itself) results in an initial config that only needs to initialize a finite amount of the tape -- that would complete the proof. (Doesn't seem likely, from what I understand so far, but that's just my guess.)
I even have to go so far as to question his honesty, here.
There are two interesting cases: (1) programs that generate so many writes, so quickly, that no matter what they will have to periodically block waiting for I/O subsystems to catch up; (2) programs that generate bursts of writes mixed with bursts of computing or idling, during which time I/O subsystems have a chance to catch up.
Case (2) is quite extremely well solved by COW simply by using a ring of user-space buffers large enough to hold all the writes in a single burst. The resulting code is portable. On a COW system it makes optimal use of memory and time. You could add lots of linux-specific "vmsplice" (or whatever) code to get the same effect on Linux.
Case (1) is also handled by a ring of buffers if the COW implementation has a particular feature. The feature needed from the COW system is to keep a queue of recent writes. If a page fault occurs on the COW buffer for write N, then don't just copy buffer N, copy buffers N....N+K for some appropriate K. Then all the overhead that Linus is talking about goes away. With that version of COW, applications never lose, sometimes win, and always have portable code. Meanwhile, if you just grow your buffers indefinately waiting for the kernel to tell you it's ok to re-use some, you risk thrashing.
Where I think Linus is being less than frank is that he must know these basic queuing theory observations. I don't want to accuse him of incompetence so it must be an issue of frankness, no?
You only have to give out your private keys if not possessing them is an obstacle to using the software as intended.
You are free to sign code all you like if all you depend on is your reputation. It's when you try to restrict peoples use of the code that you run into problems.
Linus earlier cited Red Hat as a case of a company that will have to adjust to GPLv3. IANAL but I suspect and hope he's right -- they've been cheating for years.
Several commentors are worried that funding from these sources implies inevitable corruption of the effort into a proprietary product owned by The Big Guys. One poster sees a contradiction there and wonders why, at this funding level, one company didn't just fund the whole thing (rather than Google joining Microsoft et al.)
Don't panic. There seem to be a few things going on here:
1) The principle investigators for this project are basically intellectual "hubs". Stunning track records. Long histories of students who go on to "move and shake". Perhaps most importantly: active involvement with people from all over the industry. If you want a group that simultaneously has its fingers on the pulses of both industry and academia and has a far better grasp of both fundamentals and how to systematically move forward in good directions, you could do a whole lot worse. The point: this is, to a degree, a "write your own ticket" group of researchers and they wisely elect to go for independence and diverse funding sources. The Big Companies may be big but this crowd is a bit more immune than most to being bullied. Everyone involved knows and embraces that.
2) At the levels of management where funding decisions like this are initiated and made, people are not so out of touch as the average slashdotter is likely to think. Oh sure, they have blind spots. But they are not stupid. They've seen Internet service industry growth increasingly coming from garage projects -- almost to the point that that's the only place it comes from. They do what they can to systemize and potentiate entrepreneurial skunksworking internally but they also know the social and economic limitations of management. Importantly (as can be seen by acquisitions, for example), they know that they need to rely on many, many other people making the up-front R&D investments, most failing, and a few becoming targets for acquisition. One aspect of RAD is that it envisions radically lowering the costs of playing for those external high-risk investors. If today, there are 100 people trying to win the social-network/calendering war, and perhaps 1000 serious novel-network-service efforts overall, and each of these efforts costs many people-months just to get out of "coming soon" state --- an aim of this project is to bump those numbers of people by an order of magnitude or more and shrink the lead-time similarly.
3) This is how corporate investment in academic research is supposed to work (and so it's sad, really, that RAD materials describe this as a "new" model). Corporate investors specifically don't get exclusives and therefore don't invest all that much, individually. What do they get? Partly they get new ideas which, while open to all, the investors hope to be in the best position to use (or the best position to benefit from others using them). Partly (and complementing that) they get less tangible benefits like personal access to PIs and, generally, a leg up on "technology transfer [out of the lab and into the market]".
4) This funding model is an application of a Nash Equilibrium. Let's take Microsoft. They've no shortage of systems researchers that, polite rivalries aside, could be sequestered in a room and could do all of this work just fine -- at far greater expense to Microsoft. What happens if they do that? Google eventually figures out the gist of what problems they're solving and how and obtains the same results, given those hints, far cheaper (and good look trying to repair that with patents -- it just don't work that way). In general, at the bleeding edge like this, the most probable outcome for any of these companies is that they hurt themselves unless they choose a strategy that gives their competitors options other than a direct assault -- RAD is an example of such a winning strategy.
5) "There's something happening here." [Buffalo Springfield]. RAD materials don't talk about it directly and, indeed, they're taking a step-towards rather than looking
Re:I for one welcome our new, err ... old, overlor
on
Ma Bell is Back
·
· Score: 1
> Does this mean that Captain Crunch is coming out of retirement soon?
Well, it's appropriate to whistle a captain on board, right?
Not just pulars. Remember, they also (trying to identify a source annoying noise) discovered cosmic background radiation. They helped find the *really* big map with the "you are here" marker:-)
The Office codebase has a limited future lifecycle. They need a replacement for generic thin clients, for better team-ware support, etc. In the meanwhile, to make any progress on desktops, Unix and GNU/Linux vendors need an Office-like, Office-more-or-less-compatible replacement.
The Unix and GNU/Linux vendors pay for a (conservatively estimated) 100 employees to keep OpenOffice going. So, that's like 8 - 12 million dollars a year, still trying to catch up to a MSFT bread-and-butter product that MSFT itself sees as something they themselves need to leapfrog within the next few years.
Any revenues MSFT might get from making a Linux version available are going to have a hard time helping their position to a greater degree than that 8-12 million hurts their enemies, now and into the future. So the decision not to support Office on Linux is determined by heavy FOSS investment in a dead-from-the-start alternative, mostly by Sun but with plenty of ancillary investment to spread around among other MSFT competitors. It's a no-brainer, really.
Which is why it's delicate and potentially graceful.
Mainland is actively, albeit carefully and incrementally, moving towards greater democracy, economic liberalization, and even greater degrees of province self-management (the last of which there is some already but not obviously of a desirable sort).
Probably the best outcome is that we all just dance around this issue for a few more years and it dissolves, either because "sentiment on the street" changes on one side or the other (less likely) or because there's a clear win-win reconciliation that probably includes *some* form of peaceful reunification.
Meanwhile, companies like Google (as noted in several posts already) ought to mind their own business by picking a least-controversial political authority and following that standard. That may be emotionally rough on some Taiwanese but it's not as though Google doing something different would really advance their cause.
There are folklore points you have to pick up here and there and, sure, those could be better documented.
Mostly, though, it's just a matter of practice. Some of us don't suck as badly as others. Specifically, we care about our tool skills and have some patience in learning them.
That said: the emacs paradigms for user interaction are truly underdeveloped. It could be made vastly more "friendly" to users who won't spend lots of time learning.
-t
I'd rather overshoot the 2020 deadline by, say, 5
years; forsake the false frugality of reusing
shuttle hardware; and aim for installing a
base. Think of it as the goal of installing
a foreman's shack at a construction site.
Slashdot Mirror
I think it changes with age. When I was quite young, 10 hour a day for days on end wasn't so hard to pull off. Remembering to sleep and shower and brush my teeth were harder. The catch was that a very high percentage of the code I'd write was either pure crap, or could have been done better in less time by writing another program to write that code. As I've gotten older, I've found that it's easier to spend a large number of hours *contemplating* code -- but hard to work other than in smaller bursts actually writing the code. The difference is that when I do write code, the hours are far less wasted.
I've taken this into account and so now my plan is, that when I reach 90, I'll just wake up in the morning and fart. My heavily customized Emacs will analyze the fart and translate it into C. "Oh, boy, I wrote another new OS kernel this morning!"
Well, ok, one of the two above paragraphs is true and not the other.
The engineering authority in charge of the bridge and repairs already gave their answer to this on the morning news (yesterday, I think):
They found the crack. They designed the "band-aid": the saddle, T-bar, rods, etc. They had it fabricated and installed.
In subsequent days, they went back up to look at how it was doing. They found that it was vibrating more than they thought it should: it wasn't as rigid as it was designed to be. They recognized that this would lead to fatigue and failure.
They began designing the improved "band-aid" and planned to install it sometime in coming weeks.
To their surprise, *perhaps* related to unusually high winds, the system failed sooner than they thought it could.
The completed their improved design and are now installing it. (And they are counting their blessings that nobody was killed: they got lucky, that way.)
-t
"time to die?"
Cinematic reference to Blade Runner. I was talking about geezer knowledge so it seemed apropos.
The full quote begins with something about how "I've seen things you ... humans ... will never see..." and continues about lost / underappreciated learning.
-t
I said positive income spin-offs. I don't mean spin-offs that immediately pay for the core R&D activity - only spin-offs that if you regard them as a separate business unit, make sense on their own.
Yes, that does put a kind of pressure on the research. Yes it does distort what kind of research takes place. I think it does both of those in a good an appropriate way for an R&D lab that is a long-term, for-profit investment.
One way that it helps is that it helps the lab develop the personal relationships and the skills needed for technology transfer to those who will operate something closer to customer-facing. That skill set comes in handy when the really big R&D wins come through - so you don't wind up like Xerox PARC in 1980.
Another way it helps is the reverse direction of technology transfer - what we could call "market awareness transfer" from the closer-to-the-front business guys back to the lab. There are gazillions more things that make for interesting research than make for research that is promising for business and, insisting on at least little spin-offs early is a way to help keep the lab folks in touch with business reality.
-t
You misunderstand. By "managing wealth" I very much include not leaving *too* much of a legacy for kids, making sure as little as possible goes towards evil, and getting as much of the surplus doing good works. Buffet is schematically the right idea here, even if I don't agree with all of his particular decisions. My selfish thing is that I wouldn't want to spend 60 hours / week managing various investments. Nor would I want to just hand most of it over to the Gates foundation. $1B today, if you can make a lot of it liquid quickly, is -- I agree -- more than is reasonably needed. It's just a big responsibility and my selfish take is that I'd make a priority of reducing the amount of time I had to personally spend managing that responsibility. There are some causes I'd want a hands-on role in because I think I have intellectual contributions to make but there's a lot of grunt work in responsibly handling that large an amount of money/nominal wealth that I would want to delegate in order to concentrate on what I'm good at.
-t
I've been programming for, like, uh.... about 27 or 28 years. Arguably longer if you wanna go back to really little kid stuff.
If I had that much money - was basically (if I wanted to be) in the leisure class - what I would like to believe about myself is that I would try to secure my family's material conditions really well, try to make as efficient as possible my wealth management program, and, as to hacking.....
There are *so many* really great and valuable potential projects that (a) nobody is investing in; (b) have an investment horizon that is tough because these are projects that will take a good 5 years, let's say, to get to where seeing a return is on the table. A good 10 years before you start to see the possibility of "done".
I would start an R&D lab but a very small one - perhaps 10 people - and while we'd try to have some positive income spin-offs each year from 0 onward, the goal would be to create the kind of environment where we can take off some of the bigger, long-neglected problems.
You kids these days don't know what's possible in a GUI framework. You don't know how to do language design, systems software generally, databases, file systems, or a whole lot of other basics. You've inherited really mediocre crap and you take for granted that that's where things are at. And the industry has ceased production of grey-beards. (Also: get off my lawn!)
"like tears in the rain", -t
Thank you, "Creepy Crawler", that's right. There's more to it, as well, but that's a big part of it.
-t
I also had deeper misgivings because of the scientific sloppiness I witnessed in my work at the Church lab. The only thing in genomics work that is worse than a rush towards the new eugenics is a rush towards the new scientifically false eugenics - which was the rush my supervisor's work was aimed towards when we parted ways. These guys are bad news. On a personal level, I did not find either my supervisor or boss to be people I could sustain even a presumption of respect for. They are, in my experience, sloppy in important ways (me, I try to be sloppy in the unimportant ways) - bad science, bad ethics, money-hungry, self-serving twits.
-t
I used to work as a contractor for the George Church lab. My supervisor was a student of Church's. Church was his boss. I was working on bio-informatics (if anyone cares, I can tell you some tricks for regexp-searching of genetic sequences).
My family was under extreme financial duress. In light of that knowledge, my supervisor (tells me, at least) that he took my situation to George and they came up with this: "Sign up to be one of the first 10 PGP subjects. Give us all of your medical records from the past and into the future. Agree to have your sequence published. We think we can get Harvard to agree to pay for your medical insurance for life. Don't you think your family deserves for you to make that trade-off?"
I said, flatly, "no." I pointed out, among other problems, some severe technical problems in the line of sequencing research we working on. Ultimately, we (me and the lab) part ways on less than amicable terms after this.
I think these people are scum.
They were eager to exploit my poverty as leverage to make me a human subject to rather dangerous experimentation based on highly dubious scientific claims - and they punished me for dissenting from this plan, as nearly as I can tell.
-t
Strive to make sure that if any one member of your team were to be, goodness forbid, struck by a bus -- that you could fill there job with at least barely enough efficiency to still have a chance.
Then deal with the business side as if that is the expected outcome.
If you lack the empathic capacity or technical smarts for this, quit your job and go back to programming.
-t
It's not a panopticon. In a panopticon, the threat of perpetual surveillance is used to enforce discipline. In the contemporary situation, hidden surveillance is used to create points of control. -t
Well, yes. "They" [the political, economic, and social situation] have done just that. You have just done it by trivializing the concerns, reinforcing the hegemony that produces the problems Carr is noticing. You have illustrated a tiny sliver of the problem, by example. Good job.
-t
Yes, exploited. It's a professional embarrasment at the start of his career. I'm not sure $25K is a fair price.
It's a basic logic problem. The original challenge problem could be restated: "Prove either that there exists a non-universal machine which emulates the 2-3 machine OR that the 2-3 machine can only be emulated by a universal machine." The proof does neither. Wolfram Inc. reps have come back with "Well, perhaps we should change the definition of universality!" Only, they aren't very concrete in offering an alternative with any rigor and the vague suggestions they are making don't add up (e.g., don't answer Vaughn Pratt's counter-example of paired push-down automata).
What the student proved here may turn out to be an interesting and useful result (not the universality of 2-3 but the universality of this interesting combination of machines). Students should be encouraged to work on such problems. Students should be encouraged to write as well as this student is learning to do -- it's a nicely presented paper (trivial formatting bugs aside). But students shouldn't be encouraged to go in those directions on false premises.
At least two interesting questions come from the students work. To Wolfram Inc.'s credit, they are pointing in the general direction of these new questions (even while not yet acknowledging their mistake). The new questions: Do there exist simple machines whose universality is undecidable (and might 2-3 be an example)? If 2-3's universality is either false or undecidable (and especially in the latter case) can we find any useful structure to what combinations of it with other machines clearly are universal?
I'll leave it as an exercise to figure out how raising those questions relate to the "Priciple of Computational Equivalence" in NKS but, meanwhile, leave the student out of it!
We'll see. With an additional step that "finitizes" the student's construction the proof is rescued and raised to the status of an important lemma -- but if that step isn't very quickly forthcoming, the prize -- in no small part an advertising vehicle -- was administered in a pedagogically misleading way.
-t
I have read and followed all of the related threads on the FOM mailing list. I am also familiar with how restraint in expression works in academia vs. slashdot. Academic restraint of expression in the emails has led to some confusion on slashdot: Vaughn Pratt's objection is indeed elementary and, if true, unassailable. Vaughn Pratt has pointed out that the proof says, essentially, "the 2 state, 3 color machine in question is universal -- if you add some more states". Is that objection true? The recent responses from Wolfram Inc. have said, "yes, it is." They say so explicitly. Game over. Except wait: Things have now turned obscene: Wolfram Inc. first of all misled this student, badly when they advised him about the paper. Now he has the embarassment of being the victem of a controversy on his record. Wolfram Inc. exploited this student badly. I have yet to see any public statement from Wolfram Inc. about the matter that does not seem to me to be primarily aimed at promoting the "NKS" book. Wolfram the man has, in my opinion, embarassed himself badly. The elementary error is not only in the proof, it is in the rules of the contest. It is not written out in the rules of the contest -- the rules cite certain passages of the book, where the error is offered as fact. The heat-of-the-moment responses from Wolfram Inc. have, clearly to anyone familiar with the dialects and who has read the threads, evaded the issue. In my opinion, they have offered handwaving (and self promotion) of the worst kind. There is, as indirectly pointed out on the FOM mailing list, some chance that the 2-3 machine may yet turn out not only to not be provably universal, but perhaps even to be a clear refutation of the main thesis of NKS (as nearly as said thesis can be generously rendered as a coherent statement, in my opinion). So, it's a freaking mess. In the standards of the academic world, this is a 100-year flood kind of boner of a mistake, to put it mildly. I call upon Wolfram and Wolfram Inc. to own up, make a serious effort to articulate their critic's best arguments in plain language accessible to the slashdot crowd, and then explain how they will proceed justly from there. Otherwise, in my opinion, they will just be furthering their exploitation of others. -t p.s.: my best guess is that the lame responses from Wolfram Inc. are there to buy time -- they want 48-96 hours to rack their brains and rescue the proof. And there's a chance! Not a great chance, but a chance. For example, if it turns out that encoding the student's construction in a meta-circular interpreter (writing a program to cause it to simulate itself) results in an initial config that only needs to initialize a finite amount of the tape -- that would complete the proof. (Doesn't seem likely, from what I understand so far, but that's just my guess.)
I even have to go so far as to question his honesty, here.
There are two interesting cases: (1) programs that generate so many writes, so quickly, that no matter what they will have to periodically block waiting for I/O subsystems to catch up; (2) programs that generate bursts of writes mixed with bursts of computing or idling, during which time I/O subsystems have a chance to catch up.
Case (2) is quite extremely well solved by COW simply by using a ring of user-space buffers large enough to hold all the writes in a single burst. The resulting code is portable. On a COW system it makes optimal use of memory and time. You could add lots of linux-specific "vmsplice" (or whatever) code to get the same effect on Linux.
Case (1) is also handled by a ring of buffers if the COW implementation has a particular feature. The feature needed from the COW system is to keep a queue of recent writes. If a page fault occurs on the COW buffer for write N, then don't just copy buffer N, copy buffers N....N+K for some appropriate K. Then all the overhead that Linus is talking about goes away. With that version of COW, applications never lose, sometimes win, and always have portable code. Meanwhile, if you just grow your buffers indefinately waiting for the kernel to tell you it's ok to re-use some, you risk thrashing.
Where I think Linus is being less than frank is that he must know these basic queuing theory observations. I don't want to accuse him of incompetence so it must be an issue of frankness, no?
-t
You only have to give out your private keys
if not possessing them is an obstacle to using
the software as intended.
You are free to sign code all you like if all
you depend on is your reputation. It's when
you try to restrict peoples use of the code
that you run into problems.
Linus earlier cited Red Hat as a case of a company
that will have to adjust to GPLv3. IANAL but
I suspect and hope he's right -- they've been
cheating for years.
-t
will be reached when ham-black-hat-hackers start taking over
space probes. Who will be the first to hack the rovers?
C'mon, go for it. Prove for us the bulk of humanity sucks.
-t
Several commentors are worried that funding from these sources implies inevitable corruption of the effort into a proprietary product owned by The Big Guys. One poster sees a contradiction there and wonders why, at this funding level, one company didn't just fund the whole thing (rather than Google joining Microsoft et al.)
Don't panic. There seem to be a few things going on here:
1) The principle investigators for this project are basically intellectual "hubs". Stunning track records. Long histories of students who go on to "move and shake". Perhaps most importantly: active involvement with people from all over the industry. If you want a group that simultaneously has its fingers on the pulses of both industry and academia and has a far better grasp of both fundamentals and how to systematically move forward in good directions, you could do a whole lot worse. The point: this is, to a degree, a "write your own ticket" group of researchers and they wisely elect to go for independence and diverse funding sources. The Big Companies may be big but this crowd is a bit more immune than most to being bullied. Everyone involved knows and embraces that.
2) At the levels of management where funding decisions like this are initiated and made, people are not so out of touch as the average slashdotter is likely to think. Oh sure, they have blind spots. But they are not stupid. They've seen Internet service industry growth increasingly coming from garage projects -- almost to the point that that's the only place it comes from. They do what they can to systemize and potentiate entrepreneurial skunksworking internally but they also know the social and economic limitations of management. Importantly (as can be seen by acquisitions, for example), they know that they need to rely on many, many other people making the up-front R&D investments, most failing, and a few becoming targets for acquisition. One aspect of RAD is that it envisions radically lowering the costs of playing for those external high-risk investors. If today, there are 100 people trying to win the social-network/calendering war, and perhaps 1000 serious novel-network-service efforts overall, and each of these efforts costs many people-months just to get out of "coming soon" state --- an aim of this project is to bump those numbers of people by an order of magnitude or more and shrink the lead-time similarly.
3) This is how corporate investment in academic research is supposed to work (and so it's sad, really, that RAD materials describe this as a "new" model). Corporate investors specifically don't get exclusives and therefore don't invest all that much, individually. What do they get? Partly they get new ideas which, while open to all, the investors hope to be in the best position to use (or the best position to benefit from others using them). Partly (and complementing that) they get less tangible benefits like personal access to PIs and, generally, a leg up on "technology transfer [out of the lab and into the market]".
4) This funding model is an application of a Nash Equilibrium. Let's take Microsoft. They've no shortage of systems researchers that, polite rivalries aside, could be sequestered in a room and could do all of this work just fine -- at far greater expense to Microsoft. What happens if they do that? Google eventually figures out the gist of what problems they're solving and how and obtains the same results, given those hints, far cheaper (and good look trying to repair that with patents -- it just don't work that way). In general, at the bleeding edge like this, the most probable outcome for any of these companies is that they hurt themselves unless they choose a strategy that gives their competitors options other than a direct assault -- RAD is an example of such a winning strategy.
5) "There's something happening here." [Buffalo Springfield]. RAD materials don't talk about it directly and, indeed, they're taking a step-towards rather than looking
> Does this mean that Captain Crunch is coming out of retirement soon? Well, it's appropriate to whistle a captain on board, right?
Not just pulars. Remember, they also (trying to identify a source annoying noise) discovered cosmic background radiation. They helped find the *really* big map with the "you are here" marker :-)
Bell Labs' Cosmology page.
-t
The Office codebase has a limited future lifecycle. They need a replacement for generic thin clients, for better team-ware support, etc. In the meanwhile, to make any progress on desktops, Unix and GNU/Linux vendors need an Office-like, Office-more-or-less-compatible replacement.
The Unix and GNU/Linux vendors pay for a (conservatively estimated) 100 employees to keep OpenOffice going. So, that's like 8 - 12 million dollars a year, still trying to catch up to a MSFT bread-and-butter product that MSFT itself sees as something they themselves need to leapfrog within the next few years.
Any revenues MSFT might get from making a Linux version available are going to have a hard time helping their position to a greater degree than that 8-12 million hurts their enemies, now and into the future. So the decision not to support Office on Linux is determined by heavy FOSS investment in a dead-from-the-start alternative, mostly by Sun but with plenty of ancillary investment to spread around among other MSFT competitors. It's a no-brainer, really.
-t
Which is why it's delicate and potentially graceful.
Mainland is actively, albeit carefully and incrementally, moving
towards greater democracy, economic liberalization, and even
greater degrees of province self-management (the last of which
there is some already but not obviously of a desirable sort).
Probably the best outcome is that we all just dance around this
issue for a few more years and it dissolves, either because
"sentiment on the street" changes on one side or the other
(less likely) or because there's a clear win-win reconciliation
that probably includes *some* form of peaceful reunification.
Meanwhile, companies like Google (as noted in several posts already)
ought to mind their own business by picking a least-controversial
political authority and following that standard. That may be
emotionally rough on some Taiwanese but it's not as though Google
doing something different would really advance their cause.
-t
There are folklore points you have to pick up here and there and, sure, those could be better documented. Mostly, though, it's just a matter of practice. Some of us don't suck as badly as others. Specifically, we care about our tool skills and have some patience in learning them. That said: the emacs paradigms for user interaction are truly underdeveloped. It could be made vastly more "friendly" to users who won't spend lots of time learning. -t
What's the best way [to learn].
Allocate a couple of hours. Run emacs. Type C-h t ("control-h t").
That runs a venerable tutorial which gives you enough foundation that you can explore from there as the mood suits you.
-t
First reaction: this is all wrong.
4 guys? 7 days?
I'd rather overshoot the 2020 deadline by, say, 5 years; forsake the false frugality of reusing shuttle hardware; and aim for installing a base. Think of it as the goal of installing a foreman's shack at a construction site.
Plus, it should be an international effort.
-t