They're inherent in the medium, they're not extra copies that the Canadian government grants you as a right in exchange for a medium fee.
That's correct, but you've contradicted your first point. Here you acknowledge that the consumer has paid only for the download, not for any additional copying. The industry agrees with you. The proposal at hand is, like the media tax, intended to cover the case of additional copying not inherent in downloading. I'm not defending it, just explaining it.
It helps to remember that just because two situations are superficially similar that [sic] doesn't mean the same principles apply to both.
It doesn't mean the converse either. Really, you need to work on your logic a bit.
Referring to my original comment, the principle is to compensate the industry for the legal right of users in Canada to copy material for their own use.
That compensation can be extracted at various points. I don't know that in fact it's not also hidden in the retail cost of a recorded CD. If it's the same amount per unit as for blank CDs, you'd never notice, the price of recorded CDs being so high already. Or perhaps recorded CDs were overlooked in the original decision.
It helps to remember that all regulatory decisions are based on an application of principle. That's usually the more important focus for concern than a particular detail of implementation. And if, in your view, a decision seems unjustified or nonsensical, you probably aren't looking at the same set of principles that were weighed while making the decision.
The existing media tax in Canada compensates the industry for the legal right of users in Canada to copy material for their own use.
We occasionally hear expressions of doubt from the music industry about this arrangement, despite having proposed it originally. If the industry now asks for the principle to be extended to downloads, then clearly it is underscoring its endorsement of the arrangement.
It's good to have this question cleared up, because if a legislative reversal were to take place, I'd want my money back. I buy CDs and DVDs exclusively for system backup. That's my data and yet I have to pay a copy tax on the media. Not that I'm complaining, just pointing out that the industry can't have it both ways.
This calls for proper application of cryptography.
I'm not sure how it would look, but let's play with the idea a bit. What you'd like would be for each transmission to be signed in a particular way:
For collecting ordinary traffic data, transmissions need to be authenticated and encrypted (in order to prevent obvious abuses of the system) but remain completely anonymous.
In case of theft, there needs to be a way to add information to the system so that a particular signature is no longer anonymous but instead is identified with a particular vehicle.
The information which permits the above identity association must have effect only for a specified window of log data.
There has to be an efficient way to search the transmission logs for messages with a given identity.
It's obvious that, underlying all of this, the vehicle owner has to be given a digital identity as well, in order to provide authentication for all the other functions. I think that would address your concern about the unit being compromised, though I'm not sure how those would arise.
Let's see, what else? There is a trust issue here that can't be entirely avoided. No matter what the manufacturer may claim about the data remaining anonymous, there is no design which prevents transmission of identifying data as part of the payload. About the only way to mitigate this risk is to have the system periodically inspected by a trusted third party. That would also address your concern about network identifiers lower in the stack finding their way into the transmission logs where they could potentially be exploited by the manufacturer. Any other party with access to the identifiers, such as a telecommunications carrier, would make not be able to correlate them with transmitted data since that is encrypted.
My understanding is that a contract requires an offer, an acceptance, and the exchange of consideration. This definition is essentially unchanged from ancient times.
Your purchase of a book is a contract, because it meets this test. A gift is not a contract. The "cease and desist" letter is a gift.
Copyright is something else entirely, related to ownership of title in a creative work, under the terms of the Berne Convention or local law.
Whether it's a purchase or a gift, the author of a "cease and desist" letter has title to it. However, not every such work is subject to copyright. For example, advertizing and other promotional materials are excluded. But the strongest point is that copyright law explicitly permits fair use of a work "for purposes such as criticism." (15 USCS Section 107).
The specific capability that is being exercised here is to capture traffic at the subscriber edge of the domestic network.
Carnivore and Echelon are placed on the core networks, not the edge. They're only good for intercepting traffic that passes through the core, for example, international traffic.
CALEA, by contrast, mandates the capture of domestic traffic. Users of the domestic network are, by definition, individuals located inside the US.
So eventually (at least in theory) the prices will hit some sort of equilibrium point.
Or they might be so dynamically unstable that the system would thrash itself a couple of times and then end up jammed in the opposite state from that intended.
The only control input is the price per copy, which as formulated has a destabilizing effect on market share. Charge more when there are more copies being purchased? That's not your usual supply and demand economics, certainly not with nonrival goods where there is no scarcity of supply.
Conversely, it allows a real momentum to develop around the network effects of illegal copying. As you point out, that's not only driven directly by popularity (not a control input) but also amplified by the pricing model itself, which premises that there is a scarcity of supply. The problem is that illegal copies disprove the premise by providing a very competitive supply of product. And once the effect gets going to the point where illegal copies are commonplace, they become an extremely difficult brand to dislodge. Moreover, there's really no control input to stop the process. What, lower prices? Bit late for that, mate.
It's just not a good business strategy to drive your customers into the arms of the competition. So I don't get it. I hate the notion of DRM, but this isn't a viable alternative. You have to make it cheap and obvious in order for the masses to go there and sustain a network effect.
Which is a shame for me, personally, because my musical tastes are rarefied and I'm going to be paying extra to get at that super cool material that nobody has heard of. But the higher prices make it worthwhile for distribution to take place, or a little bit of piracy for that matter. Maybe that would take away some of the pain.
I'm not able to reproduce either the Java or the JavaScript popup in the following environment:
Mozilla 1.7.13
Sun JDK 1.5.0_02
Red Hat Linux release 9 (Shrike)
The Java popup works as advertized in Firefox 2.0.0.6, same environment in other respects.
Here's an example of why it's useful to support older versions of critical software in your environment. I know, it's all very conservative, but sometimes it's really nice to be able to roll back.
Maybe if ODF is so good we should consider switching our default formats now. Maybe that should be the first step in our migration.
It would be more practical and less risky to migrate to Open Office as the first step.
Get through the process of deploying it and supporting it. Let people become familiar with using it. Trial it initially to be sure that you can comprehensively edit your existing corpus of MS documents. You're likely to turn up a few which need to be cleaned up, and you should budget for this modest effort when considering how to promote the migration.
Once you've got the necessary software integrated and accepted in the workplace, you've overcome the technical and administrative obstacles. Conversion to ODF then essentially becomes a policy decision. Sweet!
I'm not an admin, I don't get to choose or install what I want.
I was with you up to there. Yes, in fact you do get to install what software you want. Put it in your home directory, duh. Open Office has no problem with that, neither does most other Unix software.
It's outrageous for a person in his position to misstate such an elementary security principle. Fine, if he doesn't know about security then he can just keep modestly quiet. The creepy thing is that he pretends to know. We could use a lot less of that.
Digitally signing the ballot eliminates the anonymity of it
Not so, fortunately. Think about it. You can verify a signed object against a public key without knowing who owns the corresponding private key. There is nothing in the key pair itself which carries identity.
And if you make use of a certificate infrastructure, you can verify that the public half of the key pair was signed by an authority whose identity you do know.
Certificates can be used to carry many sorts of identity, including anonymized identities. It's not the case that they must do so. This property makes them very useful for voting systems.
I know it! It's perverse, but it happens as a corollary of the very same hubris. While executives may regard themselves as versatile and adaptable to any situation, it's obvious that someone still has to deliver the necessary expertise. Oh look, there are some technical people. Since the generalist role is already taken, we'll create roles for a few specialists and get them to do all that boring "expert" stuff that those people like to do.
But just as the generalist role has been underconstrained, now the specialist role is overconstrained as well as subordinate. And because there's very little common ground between the two roles, they don't end up communicating well. You can see this in business and government both.
In my opinion, it all began with a false dichotomy in the first place. You need someone who is respected in both communities. We used to be really good at advancing such people to positions of high office, but not so much lately. Some organizations still have it, thankfully, though as a professional you have to really shop around to find it.
In government? Well, cynicism is easy. It's necessary to cultivate some optimism about that or you'll slit your wrists. But yes, it's possible to find excellent and able people in high places. And I think we just have to keep on pushing for that and not sit back cynically and watch it all run down.
That's a valid point. If part of what's broken is due to mindset then the solution needs a new mindset.
But that makes the challenge doubly hard. It still does no good to bring some unqualified individual in from out of nowhere. That just serves to shuffle the oligarchy around a bit without adding any merit. The ideal individual has to be expert in how the present system works, and also somehow untainted by it.
Now if you could parachute in some respected executive from some other patent jurisdiction, that might do it. There would still be significant friction and mistrust at first, but the move would also generate optimism and a flow of new ideas.
She has a generally strong political background but appears to be qualified in no particular whatsoever. That's a far cry from being "totally" qualified.
Totally qualified would mean qualified in all respects, not just maybe a vague hand wave because she knows how to work the political machine.
That appears to be the essential problem. People somehow have developed the hubris that any sufficiently talented generalist can master a new field in short order and lead it to success. Not so. A senior position in any field requires someone who knows the field. And at our present level of complexity, that degree of mastery takes a lifetime.
You can settle for less in the name of political expediency, but don't fool yourself that it's just as good an appointment as someone who actually knew the field well. A totally qualified candidate would know the field and be able to get along with the political establishment. You're getting half a slice of pie.
On that note, I've encountered a surprisingly strong relationship between cleaning up code and understanding it.
Anecdotal though this observation may be, I've been following the practice for about thirty years now and it's never failed me yet. Best of all, it minimizes mental fatigue. In fact, for much of the time it's almost therapeutic. It progresses gradually from rote activity to insightful redesign.
Setting aside the various possible refinements such as versioning, group walkthroughs, and regression testing, it goes basically like this:
Reformat the code for consistent indentation, layout, use of whitespace and so on, with the intention of clearly expressing lexical structure. Make no other changes.
Change symbol names for consistent capitalization, name length, symmetries in naming style. Add definitions for constants. Remove unused declarations. The intention is primarily to improve readability, but you'll probably notice that you've improved meaning as a side effect. At this point, isolated parts of the code are starting to look competently written, and you have at least moderate confidence that you understand how those parts work.
Move things around for consistent grouping. What exactly this means will depend a lot on the size and complexity of the application, as well as how it fits into the larger corpus of software to be maintained. This is essentially an exercise in librarianship, of making things easy to find.
Make what we might call syntactic structural changes. Replace weird tests and loop constructs with obvious ones, make more sensible variable assignments. You'll probably find that all sorts of little things get changed at this stage, and some disappear outright. This is where you have to stay sharp and do the most aggressive testing and debugging, because it's easy to introduce obscure errors or to remove something that looks weird but is responsible for correct behavior. In short, this is the dirtiest part of the exercise right here.
Begin semantic restructuring without changing the architecture. The work should be getting easier again, because you're only looking to replace messy semantics with clean semantics. For example, you might recognize that several functions that do nearly the same thing could be merged, that similar methods could be moved into a superclass, that the code should not have raw library calls scattered about but should be passing them through a slightly more abstract layer.
Refactor the architecture where it has become obvious that refactoring is called for.
It's a somewhat slow and contemplative process, rather like spring cleaning. It takes time, but it's not stressful, and except for wbat I've called syntactic restructuring, it doesn't require a great deal of alertness. On the contrary, you're transforming the code so that it will appear obvious even to a not very alert person. And, amazingly, it usually is obvious at the end of the process.
static type checking... the strange choice of "float" over "rational"... lack of the "symbol" type as a builtin
You're talking about the Lisp Machine, my man, speaking of MIT. Dynamic type checking, atom as a base type, a great platform for symbolic math and lazy evaluation.
That project drew some outstanding architects and programmers too, speaking of what does it take to make a great computer scientist. But in commercialization it couldn't keep up with the mainstream, fell behind in performance and economies of scale. It embodied good ideas, though evidently not sufficiently compelling to alter the prevailing fashion.
My hope is that, if those ideas are truly sound, they'll make a reappearance eventually. And take heart that in the larger scheme of things the Lisp Machine is still very much a conservative Von Neumann architecture. More radical alternatives may make an appearance in their own time.
I take your point that simply because something lies within the current fashion is not much of a claim. But people seem fond of making such claims anyway. [sigh]
Creative time such as project work, planning, research, checking in with colleagues
In terms of responsible time management, that's good enough. A good technical manager encourages his staff to do the same, and applies the second item with restraint.
You've built an eloquent case around a set of points which are a pleasure to read. Much of what you say has merit.
However, I think it needs to be said that science is not the exclusive club you make it out to be. The universals of human behavior and social elitism notwithstanding, anyone can be a scientist. All they have to do is play by the rules. Yes, if you wanted to make the case that there is a pecking order for access to the Hubble space telescope or the Large Hadron Collider, you'd get no argument from me. But nowadays the tools of a computer scientist are available to anyone.
Computer science is not a pure meritocracy either, but I'd argue that it contains fewer inherent barriers to entry than most scientific disciplines, and being a young discipline, perhaps even fewer social barriers. But that doesn't mean that any twit with a loud voice is going to get equal consideration. I disagree with your claim that we dissuade people from trying, at least no more than any other humans do in comparable settings in science, the humanities, the arts, even at a barbeque competition. If someone wants to go off and do their own thing, fine, I don't see anyone stopping them. But I also think it's reasonable to require some demonstration of merit for an idea to gain acceptance within a community of practice.
However we may resolve the foregoing question, conditions don't seem to be stopping people from doing all kinds of crazy shit with computers. I don't know if encouraging more of this is required in order for it to become a socially or scientifically fruitful activity. Its inherent randomness is hard to organize purposefully. Yes, in principle more monkeys with typewriters could produce more works of Shakespeare, and likewise we could set up some kind of random funding institute. I had a similar idea when I was a teenager of wrapping many millions of joints in hundred-dollar bills and then shovelling them out of light aircraft over populated areas. Surely that would have produced some positive, though highly unpredictable, creative outcomes. But I could never get support from the government despite the liberal funding climate of the day.
I'd also like to respond to the suggestion that the Linux community is perhaps being hypocritical or splitting hairs when rejecting alternative approaches to the application of computer science. I think that, to be credible as a constructive suggestion, it has to be made with respect to particular examples and not as a sweeping generalization. In other words, your dissing of a hypothesized dissing falls to its own argument. And of course it's a strawman argument in any case. The Linux community does not stand in place of the computer science community at large. What could be said is that they're both meritocracies involving the use of computers. I see that as a common success, not as a common failing.
No doubt things are missing from the corpus of identified things that could engage computer science. But which in particular are important? Hypothetical lack is a fine notion, but on an equally sound philosophical footing, I think it is reasonable to assume that things are okay as they are, in the absence of evidence to the contrary. I don't share your view that there is a gap which needs to be closed, but again, it would help if the discussion could focus on particulars.
In particular terms, I know no more dedicated group of professionals than the dozens -- or perhaps by now hundreds -- of computer scientists I've seen acting on thesis committees and with granting agencies, sitting on conference panels, editing research journals and reviewing submissions, supervising grad students, teaching undergraduates, developing curricula, and all of the other duties expected of someone working in science. These are people who encounter and debate new and sometimes outrageously alternative ideas every day of the week, and who I've often often seen make the point of asking, in these settings, whether compute
Slashdot Mirror
That's correct, but you've contradicted your first point. Here you acknowledge that the consumer has paid only for the download, not for any additional copying. The industry agrees with you. The proposal at hand is, like the media tax, intended to cover the case of additional copying not inherent in downloading. I'm not defending it, just explaining it.
It helps to remember that just because two situations are superficially similar that [sic] doesn't mean the same principles apply to both.
It doesn't mean the converse either. Really, you need to work on your logic a bit.
That compensation can be extracted at various points. I don't know that in fact it's not also hidden in the retail cost of a recorded CD. If it's the same amount per unit as for blank CDs, you'd never notice, the price of recorded CDs being so high already. Or perhaps recorded CDs were overlooked in the original decision.
It helps to remember that all regulatory decisions are based on an application of principle. That's usually the more important focus for concern than a particular detail of implementation. And if, in your view, a decision seems unjustified or nonsensical, you probably aren't looking at the same set of principles that were weighed while making the decision.
But they don't regret it. They're endorsing the principle by asking for it to be applied more broadly.
We occasionally hear expressions of doubt from the music industry about this arrangement, despite having proposed it originally. If the industry now asks for the principle to be extended to downloads, then clearly it is underscoring its endorsement of the arrangement.
It's good to have this question cleared up, because if a legislative reversal were to take place, I'd want my money back. I buy CDs and DVDs exclusively for system backup. That's my data and yet I have to pay a copy tax on the media. Not that I'm complaining, just pointing out that the industry can't have it both ways.
I'm not sure how it would look, but let's play with the idea a bit. What you'd like would be for each transmission to be signed in a particular way:
-
For collecting ordinary traffic data, transmissions need to be authenticated and encrypted (in order to prevent obvious abuses of the system) but remain completely anonymous.
-
In case of theft, there needs to be a way to add information to the system so that a particular signature is no longer anonymous but instead is identified with a particular vehicle.
-
The information which permits the above identity association must have effect only for a specified window of log data.
-
There has to be an efficient way to search the transmission logs for messages with a given identity.
It's obvious that, underlying all of this, the vehicle owner has to be given a digital identity as well, in order to provide authentication for all the other functions. I think that would address your concern about the unit being compromised, though I'm not sure how those would arise.Let's see, what else? There is a trust issue here that can't be entirely avoided. No matter what the manufacturer may claim about the data remaining anonymous, there is no design which prevents transmission of identifying data as part of the payload. About the only way to mitigate this risk is to have the system periodically inspected by a trusted third party. That would also address your concern about network identifiers lower in the stack finding their way into the transmission logs where they could potentially be exploited by the manufacturer. Any other party with access to the identifiers, such as a telecommunications carrier, would make not be able to correlate them with transmitted data since that is encrypted.
Your purchase of a book is a contract, because it meets this test. A gift is not a contract. The "cease and desist" letter is a gift.
Copyright is something else entirely, related to ownership of title in a creative work, under the terms of the Berne Convention or local law.
Whether it's a purchase or a gift, the author of a "cease and desist" letter has title to it. However, not every such work is subject to copyright. For example, advertizing and other promotional materials are excluded. But the strongest point is that copyright law explicitly permits fair use of a work "for purposes such as criticism." (15 USCS Section 107).
When a party acts to put unsolicited material into my hands, I must reasonably conclude that they are granting me unconditional use of same.
I always said that Microsoft would never successfully migrate from 16 to 32 bits...
Me!
And I put a parabolic reflector behind it for good measure.
Yep, it sure does hurt when you shoot yourself in the foot like that.
The specific capability that is being exercised here is to capture traffic at the subscriber edge of the domestic network.
Carnivore and Echelon are placed on the core networks, not the edge. They're only good for intercepting traffic that passes through the core, for example, international traffic.
CALEA, by contrast, mandates the capture of domestic traffic. Users of the domestic network are, by definition, individuals located inside the US.
Or they might be so dynamically unstable that the system would thrash itself a couple of times and then end up jammed in the opposite state from that intended.
The only control input is the price per copy, which as formulated has a destabilizing effect on market share. Charge more when there are more copies being purchased? That's not your usual supply and demand economics, certainly not with nonrival goods where there is no scarcity of supply.
Conversely, it allows a real momentum to develop around the network effects of illegal copying. As you point out, that's not only driven directly by popularity (not a control input) but also amplified by the pricing model itself, which premises that there is a scarcity of supply. The problem is that illegal copies disprove the premise by providing a very competitive supply of product. And once the effect gets going to the point where illegal copies are commonplace, they become an extremely difficult brand to dislodge. Moreover, there's really no control input to stop the process. What, lower prices? Bit late for that, mate.
It's just not a good business strategy to drive your customers into the arms of the competition. So I don't get it. I hate the notion of DRM, but this isn't a viable alternative. You have to make it cheap and obvious in order for the masses to go there and sustain a network effect.
Which is a shame for me, personally, because my musical tastes are rarefied and I'm going to be paying extra to get at that super cool material that nobody has heard of. But the higher prices make it worthwhile for distribution to take place, or a little bit of piracy for that matter. Maybe that would take away some of the pain.
- Mozilla 1.7.13
- Sun JDK 1.5.0_02
- Red Hat Linux release 9 (Shrike)
The Java popup works as advertized in Firefox 2.0.0.6, same environment in other respects.Here's an example of why it's useful to support older versions of critical software in your environment. I know, it's all very conservative, but sometimes it's really nice to be able to roll back.
It would be more practical and less risky to migrate to Open Office as the first step.
Get through the process of deploying it and supporting it. Let people become familiar with using it. Trial it initially to be sure that you can comprehensively edit your existing corpus of MS documents. You're likely to turn up a few which need to be cleaned up, and you should budget for this modest effort when considering how to promote the migration.
Once you've got the necessary software integrated and accepted in the workplace, you've overcome the technical and administrative obstacles. Conversion to ODF then essentially becomes a policy decision. Sweet!
I was with you up to there. Yes, in fact you do get to install what software you want. Put it in your home directory, duh. Open Office has no problem with that, neither does most other Unix software.
Cheers.
It's outrageous for a person in his position to misstate such an elementary security principle. Fine, if he doesn't know about security then he can just keep modestly quiet. The creepy thing is that he pretends to know. We could use a lot less of that.
Exactly, like a double blind.
Not so, fortunately. Think about it. You can verify a signed object against a public key without knowing who owns the corresponding private key. There is nothing in the key pair itself which carries identity.
And if you make use of a certificate infrastructure, you can verify that the public half of the key pair was signed by an authority whose identity you do know.
Certificates can be used to carry many sorts of identity, including anonymized identities. It's not the case that they must do so. This property makes them very useful for voting systems.
But just as the generalist role has been underconstrained, now the specialist role is overconstrained as well as subordinate. And because there's very little common ground between the two roles, they don't end up communicating well. You can see this in business and government both.
In my opinion, it all began with a false dichotomy in the first place. You need someone who is respected in both communities. We used to be really good at advancing such people to positions of high office, but not so much lately. Some organizations still have it, thankfully, though as a professional you have to really shop around to find it.
In government? Well, cynicism is easy. It's necessary to cultivate some optimism about that or you'll slit your wrists. But yes, it's possible to find excellent and able people in high places. And I think we just have to keep on pushing for that and not sit back cynically and watch it all run down.
But that makes the challenge doubly hard. It still does no good to bring some unqualified individual in from out of nowhere. That just serves to shuffle the oligarchy around a bit without adding any merit. The ideal individual has to be expert in how the present system works, and also somehow untainted by it.
Now if you could parachute in some respected executive from some other patent jurisdiction, that might do it. There would still be significant friction and mistrust at first, but the move would also generate optimism and a flow of new ideas.
She has a generally strong political background but appears to be qualified in no particular whatsoever. That's a far cry from being "totally" qualified.
Totally qualified would mean qualified in all respects, not just maybe a vague hand wave because she knows how to work the political machine.
That appears to be the essential problem. People somehow have developed the hubris that any sufficiently talented generalist can master a new field in short order and lead it to success. Not so. A senior position in any field requires someone who knows the field. And at our present level of complexity, that degree of mastery takes a lifetime.
You can settle for less in the name of political expediency, but don't fool yourself that it's just as good an appointment as someone who actually knew the field well. A totally qualified candidate would know the field and be able to get along with the political establishment. You're getting half a slice of pie.
Anecdotal though this observation may be, I've been following the practice for about thirty years now and it's never failed me yet. Best of all, it minimizes mental fatigue. In fact, for much of the time it's almost therapeutic. It progresses gradually from rote activity to insightful redesign.
Setting aside the various possible refinements such as versioning, group walkthroughs, and regression testing, it goes basically like this:
- Reformat the code for consistent indentation, layout, use of whitespace and so on, with the intention of clearly expressing lexical structure. Make no other changes.
- Change symbol names for consistent capitalization, name length, symmetries in naming style. Add definitions for constants. Remove unused declarations. The intention is primarily to improve readability, but you'll probably notice that you've improved meaning as a side effect. At this point, isolated parts of the code are starting to look competently written, and you have at least moderate confidence that you understand how those parts work.
- Move things around for consistent grouping. What exactly this means will depend a lot on the size and complexity of the application, as well as how it fits into the larger corpus of software to be maintained. This is essentially an exercise in librarianship, of making things easy to find.
- Make what we might call syntactic structural changes. Replace weird tests and loop constructs with obvious ones, make more sensible variable assignments. You'll probably find that all sorts of little things get changed at this stage, and some disappear outright. This is where you have to stay sharp and do the most aggressive testing and debugging, because it's easy to introduce obscure errors or to remove something that looks weird but is responsible for correct behavior. In short, this is the dirtiest part of the exercise right here.
- Begin semantic restructuring without changing the architecture. The work should be getting easier again, because you're only looking to replace messy semantics with clean semantics. For example, you might recognize that several functions that do nearly the same thing could be merged, that similar methods could be moved into a superclass, that the code should not have raw library calls scattered about but should be passing them through a slightly more abstract layer.
- Refactor the architecture where it has become obvious that refactoring is called for.
It's a somewhat slow and contemplative process, rather like spring cleaning. It takes time, but it's not stressful, and except for wbat I've called syntactic restructuring, it doesn't require a great deal of alertness. On the contrary, you're transforming the code so that it will appear obvious even to a not very alert person. And, amazingly, it usually is obvious at the end of the process.You're talking about the Lisp Machine, my man, speaking of MIT. Dynamic type checking, atom as a base type, a great platform for symbolic math and lazy evaluation.
That project drew some outstanding architects and programmers too, speaking of what does it take to make a great computer scientist. But in commercialization it couldn't keep up with the mainstream, fell behind in performance and economies of scale. It embodied good ideas, though evidently not sufficiently compelling to alter the prevailing fashion.
My hope is that, if those ideas are truly sound, they'll make a reappearance eventually. And take heart that in the larger scheme of things the Lisp Machine is still very much a conservative Von Neumann architecture. More radical alternatives may make an appearance in their own time.
I take your point that simply because something lies within the current fashion is not much of a claim. But people seem fond of making such claims anyway. [sigh]
-
Situational assessment and action by urgency
-
Scheduled commitments
-
Creative time such as project work, planning, research, checking in with colleagues
In terms of responsible time management, that's good enough. A good technical manager encourages his staff to do the same, and applies the second item with restraint.However, I think it needs to be said that science is not the exclusive club you make it out to be. The universals of human behavior and social elitism notwithstanding, anyone can be a scientist. All they have to do is play by the rules. Yes, if you wanted to make the case that there is a pecking order for access to the Hubble space telescope or the Large Hadron Collider, you'd get no argument from me. But nowadays the tools of a computer scientist are available to anyone.
Computer science is not a pure meritocracy either, but I'd argue that it contains fewer inherent barriers to entry than most scientific disciplines, and being a young discipline, perhaps even fewer social barriers. But that doesn't mean that any twit with a loud voice is going to get equal consideration. I disagree with your claim that we dissuade people from trying, at least no more than any other humans do in comparable settings in science, the humanities, the arts, even at a barbeque competition. If someone wants to go off and do their own thing, fine, I don't see anyone stopping them. But I also think it's reasonable to require some demonstration of merit for an idea to gain acceptance within a community of practice.
However we may resolve the foregoing question, conditions don't seem to be stopping people from doing all kinds of crazy shit with computers. I don't know if encouraging more of this is required in order for it to become a socially or scientifically fruitful activity. Its inherent randomness is hard to organize purposefully. Yes, in principle more monkeys with typewriters could produce more works of Shakespeare, and likewise we could set up some kind of random funding institute. I had a similar idea when I was a teenager of wrapping many millions of joints in hundred-dollar bills and then shovelling them out of light aircraft over populated areas. Surely that would have produced some positive, though highly unpredictable, creative outcomes. But I could never get support from the government despite the liberal funding climate of the day.
I'd also like to respond to the suggestion that the Linux community is perhaps being hypocritical or splitting hairs when rejecting alternative approaches to the application of computer science. I think that, to be credible as a constructive suggestion, it has to be made with respect to particular examples and not as a sweeping generalization. In other words, your dissing of a hypothesized dissing falls to its own argument. And of course it's a strawman argument in any case. The Linux community does not stand in place of the computer science community at large. What could be said is that they're both meritocracies involving the use of computers. I see that as a common success, not as a common failing.
No doubt things are missing from the corpus of identified things that could engage computer science. But which in particular are important? Hypothetical lack is a fine notion, but on an equally sound philosophical footing, I think it is reasonable to assume that things are okay as they are, in the absence of evidence to the contrary. I don't share your view that there is a gap which needs to be closed, but again, it would help if the discussion could focus on particulars.
In particular terms, I know no more dedicated group of professionals than the dozens -- or perhaps by now hundreds -- of computer scientists I've seen acting on thesis committees and with granting agencies, sitting on conference panels, editing research journals and reviewing submissions, supervising grad students, teaching undergraduates, developing curricula, and all of the other duties expected of someone working in science. These are people who encounter and debate new and sometimes outrageously alternative ideas every day of the week, and who I've often often seen make the point of asking, in these settings, whether compute