The FSF has already stated that if the public has had the right to use the program under the GPL that it can't be revoked.
The FSF can state whatever it wants, but its not going to change the law. The reason that the GPL may be revocable, at least in the US, is that, as a matter of law, copyright licenses that do not meet the legal definition of contracts (among other things, requiring mutual consideration) are revocable at will by the licensor. Statements by the FSF, even where they are the copyright holder (which is often not the case on GPL software), have no bearing on that.
The FSF's statements may have bearing on whether or not promissory estoppel is invoked in particular cases by courts to limit the power of a copyright holder to seek a remedy for copyright violation after the GPL has been revoked, since they might have some bearing on what reasonable expectation the licensee had. But promissory estoppel is unlikely to be a complete and eternal bar to any action after a revocation; it is most likely to prevent things like the copyright holder revoking the license and then taking immediate action seeking damages for copyright infringement against someone distributing software consistent with the now-revoked license.
Like a lot of things in the USA, we're actually very fragmented. Remember, we don't have a national school system, we have 50 state level systems.
More than that in the case of higher ed, actually; as each private institution is its own unique institution, and some states have multiple, separately governed public systems (California has a Community College system, the California State University system, and the University of California system.)
Not unless they're prepared to rewrite or throw away any parts contributed by other authors under the GPL.
In order to offer both GPL and commercial licenses, MySQL already had to demand copyright assignment for any code they included; it wouldn't have been legal for them to use the dual-licensing model they did if they accepted contributions under the GPL. So no worries on that front.
This leads us to the only part of the GPL that I think is in any way legally questionable (IANAL). I'm not sure it is entirely legally clear if the copyright holder is allowed to revoke the GPL licensing terms or not, no matter what is said in the license.
I think its pretty well settled case law, in the US at least, that a gratuitous license is revocable at will regardless of what the license says (of course, if you are also a GPL licensee, revoking the downstream license of your derivative of a GPL-licensed work breaches the license you have, with all the usual consequences of such a breach, but if you require copyright assignment for all code that you include, then that's not an issue.)
If revoked, the consequences of the revocation might be limited to some degree, to prevent injustice, on a promissory estoppel basis, but that's a case-by-case, minimum-to-avoid-injustice equitable remedy, not a clear-cut, absolute legal right.
You know, I never understood the point of InnoDB. One may want a complete, fully functional DBMS, in that case, there is PostgreSQL, or one may want a lightning fast data indexing/accessing machine, and for that case there is MySQL.
The point of MySQL isn't a "lightning fast data indexing/accessing machine". The point of MySQL is the modular backends which enable it to serve as a common gateway to tables that each use the storage engine most appropriate to the way the table is used. (some of which may require a lightning fast data indexing/acessing engine and accept some risks to get it, some tables may not.)
The point of InnoDB (and, presumably, Falcon) is to support the kind of usage scenarios for which traditional RDBMS are designed, while the point of certain other MySQL table drivers is to support other types of loads.
If you do it based on a clean-room reimplementation of the database's communication protocol rather than, e.g., relying on the vendor's client libraries, as GP's post ("as long as you know the protocol...") suggest, the main barrier to the selling the client app is the challenge it poses to your ability to assure recipients that it won't get broken by future server upgrades.
Well, even if the maintainers have the copyrights, that only means future versions can be closed source. They can't terminate the already-outstanding licenses without a breach of terms.
If the "already-outstanding" licenses are license contracts, that would be true, but that is only the case where there is mutual consideration (i.e., where the licensee has given something to the licensor in exchange for the permissions in the license.) Gratuitous licenses (which most FOSS licenses would seem to be) are revocable at will, though the effect of revocation may be limited where a court would find that it result in an injustice on an estoppel theory.
The kind of legal uncertainty a license revocation would create for any fork would probably kill any serious business adoption and paid support opportunities until and unless the cloud could be cleared up, though personal users might not care.
Just because you can run your games of choice on an older system doesn't mean we're all tech-whores when we get newer components. Sometimes, it's needed.
Nor did I mean to suggest that, at all. However, I wasn't really clear about the point I was trying to make, which is this: it seems to me that, it is less the case now than it has been at any other time for quite some number of years that even new games really require the latest and the greatest hardware to run acceptably. There are, of course, quite a few games that do require high-end gear when they come out, and can benefit from gear that isn't even available when they are first release. But my perception is that the trend of requiring the latest and greatest is declining, even for gaming, so that expecting the current (and even moreso, past) degree to which being a gamer has meant needing the highest-end hardware on the market to continue without limit into the future is likely unrealistic.
Perhaps. But pre-SDI, there was no suggestion that defensive weapons could do more than mitigate the effect of a Soviet attack. By contrast SDI was supposed to be a "shield" that we could hide safely behind.
AFAICT, even (among serious participants in defense policy debate/analysis) the optimists who believed that the systems in SDI could in some reasonable time frame be deployed never believed that; insofar as such characterizations got made, they were political sales pitches. The real strategic purpose, as with earlier defense systems, was (1) to assure the capacity to retaliate against an enemy with more advanced weapons than earlier systems were meant to defend against, and thus deny the enemy the capacity (or even the ability to delude themselves into believing they had the capacity) for an overwhelming first strike, and (2) to raise the fear in any potential enemy (again, noting that our primary enemy had more sophisticated weapons than previously) that we might believe we had sufficient defense against any enemy retaliation to any first strike of ours that the enemy must take seriously our threat that we might use nuclear weapons in an escalating crisis, forcing them to back down without a war being initiated at all.
Essentially, all through the cold war everything with nuclear weapons centered on (1) deterring attack by assuring that no enemy could believe they could attack you with losses that would be acceptable to their leadership, and (2) convincing the enemy that you believed that you could attack them with results that would be acceptable to your leadership, so that they could not risk actions that might provoke an attack.
Of course, those strategies aren't the kind of things you can be completely open about with your own public most of the time.
The hardware manufacturers won't like this, as consumers will wind up going for a product that will last much longer for no particular price premium, but I'm not sure what they can do about it.
Hardware manufacturers will, I think, do fine, especially if free software takes off. Sure, the individual computers they sell may not be as expensive, and may be retained longer, but they'll sell more of them at the low end, and probably also sell more of the high-end servers (sure, home users may not have their own, but all those netbooks mean lots of users accessing the web and, if they aren't storing stuff on their home servers, probably using someone's servers in the 'cloud', too: and all those means demand for hardware -- servers, routers, etc.)
But "good enough" computing won't suffice for gamers.
It won't suffice for people for whom having the latest and greatest is a status thing; OTOH, I'm using a several year old Athlon with a fairly basic GPU (one of Nvidia's integrated products, not a dedicated board) for our home's gaming PC and its fine for most games, even the newer ones I play (e.g., Empire: Total War.)
The core of it is, even if people got to the point where they all flipped over to Linux, the "good enough" computers of today just wouldn't be acceptable for the kids of tomorrow.
Of course not: the kids of tomorrow will be using the "good enough" computers of tomorrow, which will be (adjusted for inflation) cheaper than those of today, and farther behind the "high end" of computers that are available then the "good enough" computers of today are. Which is the same thing that would be true if you compared "good enough" vs "high end" PCs today to those of 1995, or those of 1995 with those of 1985.
Parkinson's law is "Work expands to fill all available time". It applies to processing power too. What's "good enough" today won't be "good enough" tommorrow, because someone will invent some CPU-sucking memory-hogging disk-flogging killer app that everybody will want to have.
Sure, but over time the percentage of computers that are sold new, and in general use, that are significantly below the "top of the line" increases -- and that's not just a prediction of the future, but I think something that is true of the trend over time we've already seen. As networks become more pervasive, I would expect to accelerate that trend, because what you can use your computer to do becomes increasingly distinct from what your computer can do on its own.
Will my 'good enough' computer handle my photo library, my 32MP entry level camera, recognise the faces in my photo collection. This sound like far fetched stuff today, but as these technologies peculate down from high end systems and people get used to the computer doing more of their mind-numbing repetitive tasks, user expectation will adapt and want them in their 'good enough' computers.
There certainly will be things you want to do that the cheap ultraportable computer won't do, but you'll still have the capabilities, because the "good enough" computers that almost everyone will have at least one of will be something like a netbook you can take with you everywhere that can do basic tasks on its own, and hook up with a better computer -- perhaps your "latest and greatest" desktop that each household might have one of at home, or something in the cloud -- via pervasive wireless networking (whether more widely provisioned WiFi access points, or an evolution of cellular data systems) to handle those tasks it can't handle on its own.
There's a few areas that MySQL still wins:... Analytic functions and general SQL friendliness.
How so? PostgreSQL's available aggregate functions appear to be a superset of MySQL's, and neither MySQL nor PostgreSQL support, in their current general release versions, what I've ususally seen referred to as "analytic" functions (windowed aggregate functions) at all (though PostgreSQL has some support in the 8.4beta.)
Re:Postgres is looking better than ever
on
Oracle Buys Sun
·
· Score: 1
The case sensitivity isn't just about the equality operator; it's also about index structure and uniqueness constraint checking. Basically, if you declare a uniqueness constraint on a column with a case-sensitive collation, you're going to allow keys that differ on 'abc' vs. 'ABC', which trust me, will lead to duplicates.
It will lead to "duplicates" (not actual duplicates, but semantic duplicates) only if case is irrelevant to the semantics of the attribute the column represents. In which case, the uniqueness constraint should not be on the string representation, but on the function of the string representation which must be unique (either upper(column) or lower(column).)
Sure, making string equality case insensitive sometimes goes well with a poorly analyzed domain model to accidentally work correctly, but equally so its going to work wrong where you have an attribute where case is meaningful that needs to be unique and support all the possible distinct values.
From the other comments, it does seem like MySQL does have the correct solution here: per-column collations, so that "equality" stands for a user-chosen equivalence relation.
Yeah, and this is one of those areas where MySQL is ahead in terms of supporting standard functionality than PostgreSQL. (There are, of course, quite a number of those that work the other way.)
Who the hell modded this insightful? While parent seems to have an understanding of what it means, the light-year is an internationally-accepted unit of length. Everybody knows what it means. His pedantic rephrasing of it is redundant, if that...
The post it was responding to clearly interpreted "20 light-years" as 20 years travel time, so GP's explanation that it was 20 years if and only if travelling at the speed of light, something we are notably incapable of doing at the current time, was, taken in context, neither "pedantic" nor "redundant", but perfectly appropriate and relevant.
I'm not sure anything we've found suggests that our type of solar system is rare. The limitations of our detection method by and large assures we'd find systems different from our own first. Astrophysicists might not have expected to find gas giants very close in to stars, but if they exist, we were going to find those first.
To elaborate on that (you covered the distance part, yourself), the main factors is detecting exoplanets right now are (1) its easier to detect bigger exoplanets, and (2) its easier to detect exoplanets closer to the stars they orbit. So, gas giants orbitting close to the stars are comparatively easy to detect, anything smaller and/or more distant is harder.
You can't generalize well from the results of a highly-biased detection system.
To be fair, the Nike-Hercules missiles were among the last nuclear defenses intended to be employed. The first was to knock out air bases with nuclear strikes to prevent bombers from getting in the air in the first place. After that came air interception using missiles such as the AIR-2 Genie. Nuclear-tipped SAMs would attempt to intercept over the ocean or unpopulated territory where possible (the Nike-Hercules had a range of over 75 miles), and explode over populated territories only if nothing else worked.
All true. Then again, nuclear-bomb-pumped X-ray lasers weren't the sole defense envisioned in SDI, either. My point was mainly that (leaving aside the fact that getting anything like the envisioned system working in any reasonable number of decades where sheer fantasy), having a defensive system that had a component that relied on detonating nuclear weapons wasn't all that out of line with what we'd had prior to the time work started on SDI (I think there was a gap of a few years between the retirement of the last Nike-Hercules batteries and when "Star Wars" proposals came out, but not much.)
If they had actually deployed lasers like that one, I think I would have been more afraid of our missile defense than of any missiles.
Considering that our pre-Star Wars anti-bomber defenses included preparing to toss up missiles with nuclear warheads in the midst of bomber formations, often necessarily over populated areas (as with Nike-Hercules), its not like the bomb-pumped lasers to defend against ballistic missiles would have been all that out of line with what preceded them (had they, you know, been practical to deploy.)
...And done again. The problem isn't having a way to communicate 3D content with the browser in a form which supports interaction, the problem is coming up with something worth using it for.
You might want to check the legal definition of "monopoly", as opposed to its etymology.
I'm quite familiar with the legal definition, which centers around market power rather than marketshare, but there is an intrinsic linkage between the two, and the latter can be pretty strong evidence for or against the former. Of course, if you think there is countervailing evidence that Oracle + MySQL would credibly constitute a monopoly, you are free to present your evidence here rather than just post throwaway comments with neither argument nor evidence.
That doesn't mean the free fork will stay the same as, or even compatible with, the "official" version.
Oracle would be kind of dumb not to keep up Java development open, since it would open up the field to some competitor to provide a commercially-supported-but-open-source fork that would harness the open-source community and be attractive to enterprises (and, since it could be as compatible with legacy Java apps as any closed version that Oracle produced, Oracle wouldn't have any compatibility advantage.) And there is at least one major competitor of Oracle's well positioned to do that.
You do realize that there are snide believers in evolution, also...right?
So? If there weren't snide evolutionists asking questions at the event on which the description was based, the fact that they exist elsewhere is irrelevant. The use of a adjective along with a noun usually suggests that the adjective is adding additional information that is not implicit in the noun, and also usually does not mean that there are no instances where the same adjective would apply to other nouns, even ones that are semantically opposed to the one being used in the present sentence.
Just being a creationist doesn't make someone "snide".
If it did, the phrase "snide creationist" would be redundant, so the mere use of the phrase suggests, indeed, that the GP realizes that.
Not looking for an argument, by any means...just pointing that out.
...in a context in which it is completely irrelevant. If you aren't trolling, exactly what are you doing?
Slashdot Mirror
The FSF can state whatever it wants, but its not going to change the law.
The reason that the GPL may be revocable, at least in the US, is that, as a matter of law, copyright licenses that do not meet the legal definition of contracts (among other things, requiring mutual consideration) are revocable at will by the licensor. Statements by the FSF, even where they are the copyright holder (which is often not the case on GPL software), have no bearing on that.
The FSF's statements may have bearing on whether or not promissory estoppel is invoked in particular cases by courts to limit the power of a copyright holder to seek a remedy for copyright violation after the GPL has been revoked, since they might have some bearing on what reasonable expectation the licensee had. But promissory estoppel is unlikely to be a complete and eternal bar to any action after a revocation; it is most likely to prevent things like the copyright holder revoking the license and then taking immediate action seeking damages for copyright infringement against someone distributing software consistent with the now-revoked license.
More than that in the case of higher ed, actually; as each private institution is its own unique institution, and some states have multiple, separately governed public systems (California has a Community College system, the California State University system, and the University of California system.)
In order to offer both GPL and commercial licenses, MySQL already had to demand copyright assignment for any code they included; it wouldn't have been legal for them to use the dual-licensing model they did if they accepted contributions under the GPL. So no worries on that front.
I think its pretty well settled case law, in the US at least, that a gratuitous license is revocable at will regardless of what the license says (of course, if you are also a GPL licensee, revoking the downstream license of your derivative of a GPL-licensed work breaches the license you have, with all the usual consequences of such a breach, but if you require copyright assignment for all code that you include, then that's not an issue.)
If revoked, the consequences of the revocation might be limited to some degree, to prevent injustice, on a promissory estoppel basis, but that's a case-by-case, minimum-to-avoid-injustice equitable remedy, not a clear-cut, absolute legal right.
The point of MySQL isn't a "lightning fast data indexing/accessing machine". The point of MySQL is the modular backends which enable it to serve as a common gateway to tables that each use the storage engine most appropriate to the way the table is used. (some of which may require a lightning fast data indexing/acessing engine and accept some risks to get it, some tables may not.)
The point of InnoDB (and, presumably, Falcon) is to support the kind of usage scenarios for which traditional RDBMS are designed, while the point of certain other MySQL table drivers is to support other types of loads.
If you do it based on a clean-room reimplementation of the database's communication protocol rather than, e.g., relying on the vendor's client libraries, as GP's post ("as long as you know the protocol...") suggest, the main barrier to the selling the client app is the challenge it poses to your ability to assure recipients that it won't get broken by future server upgrades.
If the "already-outstanding" licenses are license contracts, that would be true, but that is only the case where there is mutual consideration (i.e., where the licensee has given something to the licensor in exchange for the permissions in the license.) Gratuitous licenses (which most FOSS licenses would seem to be) are revocable at will, though the effect of revocation may be limited where a court would find that it result in an injustice on an estoppel theory.
The kind of legal uncertainty a license revocation would create for any fork would probably kill any serious business adoption and paid support opportunities until and unless the cloud could be cleared up, though personal users might not care.
Nor did I mean to suggest that, at all. However, I wasn't really clear about the point I was trying to make, which is this: it seems to me that, it is less the case now than it has been at any other time for quite some number of years that even new games really require the latest and the greatest hardware to run acceptably. There are, of course, quite a few games that do require high-end gear when they come out, and can benefit from gear that isn't even available when they are first release. But my perception is that the trend of requiring the latest and greatest is declining, even for gaming, so that expecting the current (and even moreso, past) degree to which being a gamer has meant needing the highest-end hardware on the market to continue without limit into the future is likely unrealistic.
AFAICT, even (among serious participants in defense policy debate/analysis) the optimists who believed that the systems in SDI could in some reasonable time frame be deployed never believed that; insofar as such characterizations got made, they were political sales pitches. The real strategic purpose, as with earlier defense systems, was (1) to assure the capacity to retaliate against an enemy with more advanced weapons than earlier systems were meant to defend against, and thus deny the enemy the capacity (or even the ability to delude themselves into believing they had the capacity) for an overwhelming first strike, and (2) to raise the fear in any potential enemy (again, noting that our primary enemy had more sophisticated weapons than previously) that we might believe we had sufficient defense against any enemy retaliation to any first strike of ours that the enemy must take seriously our threat that we might use nuclear weapons in an escalating crisis, forcing them to back down without a war being initiated at all.
Essentially, all through the cold war everything with nuclear weapons centered on (1) deterring attack by assuring that no enemy could believe they could attack you with losses that would be acceptable to their leadership, and (2) convincing the enemy that you believed that you could attack them with results that would be acceptable to your leadership, so that they could not risk actions that might provoke an attack.
Of course, those strategies aren't the kind of things you can be completely open about with your own public most of the time.
Hardware manufacturers will, I think, do fine, especially if free software takes off. Sure, the individual computers they sell may not be as expensive, and may be retained longer, but they'll sell more of them at the low end, and probably also sell more of the high-end servers (sure, home users may not have their own, but all those netbooks mean lots of users accessing the web and, if they aren't storing stuff on their home servers, probably using someone's servers in the 'cloud', too: and all those means demand for hardware -- servers, routers, etc.)
It won't suffice for people for whom having the latest and greatest is a status thing; OTOH, I'm using a several year old Athlon with a fairly basic GPU (one of Nvidia's integrated products, not a dedicated board) for our home's gaming PC and its fine for most games, even the newer ones I play (e.g., Empire: Total War.)
Of course not: the kids of tomorrow will be using the "good enough" computers of tomorrow, which will be (adjusted for inflation) cheaper than those of today, and farther behind the "high end" of computers that are available then the "good enough" computers of today are. Which is the same thing that would be true if you compared "good enough" vs "high end" PCs today to those of 1995, or those of 1995 with those of 1985.
Sure, but over time the percentage of computers that are sold new, and in general use, that are significantly below the "top of the line" increases -- and that's not just a prediction of the future, but I think something that is true of the trend over time we've already seen. As networks become more pervasive, I would expect to accelerate that trend, because what you can use your computer to do becomes increasingly distinct from what your computer can do on its own.
There certainly will be things you want to do that the cheap ultraportable computer won't do, but you'll still have the capabilities, because the "good enough" computers that almost everyone will have at least one of will be something like a netbook you can take with you everywhere that can do basic tasks on its own, and hook up with a better computer -- perhaps your "latest and greatest" desktop that each household might have one of at home, or something in the cloud -- via pervasive wireless networking (whether more widely provisioned WiFi access points, or an evolution of cellular data systems) to handle those tasks it can't handle on its own.
How so? PostgreSQL's available aggregate functions appear to be a superset of MySQL's, and neither MySQL nor PostgreSQL support, in their current general release versions, what I've ususally seen referred to as "analytic" functions (windowed aggregate functions) at all (though PostgreSQL has some support in the 8.4beta.)
It will lead to "duplicates" (not actual duplicates, but semantic duplicates) only if case is irrelevant to the semantics of the attribute the column represents. In which case, the uniqueness constraint should not be on the string representation, but on the function of the string representation which must be unique (either upper(column) or lower(column).)
Sure, making string equality case insensitive sometimes goes well with a poorly analyzed domain model to accidentally work correctly, but equally so its going to work wrong where you have an attribute where case is meaningful that needs to be unique and support all the possible distinct values.
Yeah, and this is one of those areas where MySQL is ahead in terms of supporting standard functionality than PostgreSQL. (There are, of course, quite a number of those that work the other way.)
The post it was responding to clearly interpreted "20 light-years" as 20 years travel time, so GP's explanation that it was 20 years if and only if travelling at the speed of light, something we are notably incapable of doing at the current time, was, taken in context, neither "pedantic" nor "redundant", but perfectly appropriate and relevant.
To elaborate on that (you covered the distance part, yourself), the main factors is detecting exoplanets right now are (1) its easier to detect bigger exoplanets, and (2) its easier to detect exoplanets closer to the stars they orbit. So, gas giants orbitting close to the stars are comparatively easy to detect, anything smaller and/or more distant is harder.
You can't generalize well from the results of a highly-biased detection system.
All true. Then again, nuclear-bomb-pumped X-ray lasers weren't the sole defense envisioned in SDI, either. My point was mainly that (leaving aside the fact that getting anything like the envisioned system working in any reasonable number of decades where sheer fantasy), having a defensive system that had a component that relied on detonating nuclear weapons wasn't all that out of line with what we'd had prior to the time work started on SDI (I think there was a gap of a few years between the retirement of the last Nike-Hercules batteries and when "Star Wars" proposals came out, but not much.)
Considering that our pre-Star Wars anti-bomber defenses included preparing to toss up missiles with nuclear warheads in the midst of bomber formations, often necessarily over populated areas (as with Nike-Hercules), its not like the bomb-pumped lasers to defend against ballistic missiles would have been all that out of line with what preceded them (had they, you know, been practical to deploy.)
The 3D web worked fine with VRML (and continued to work fine with X3D, AFAIK.) The problem is that no one really had any work for it to do.
I'm quite familiar with the legal definition, which centers around market power rather than marketshare, but there is an intrinsic linkage between the two, and the latter can be pretty strong evidence for or against the former. Of course, if you think there is countervailing evidence that Oracle + MySQL would credibly constitute a monopoly, you are free to present your evidence here rather than just post throwaway comments with neither argument nor evidence.
Oracle would be kind of dumb not to keep up Java development open, since it would open up the field to some competitor to provide a commercially-supported-but-open-source fork that would harness the open-source community and be attractive to enterprises (and, since it could be as compatible with legacy Java apps as any closed version that Oracle produced, Oracle wouldn't have any compatibility advantage.) And there is at least one major competitor of Oracle's well positioned to do that.
So? If there weren't snide evolutionists asking questions at the event on which the description was based, the fact that they exist elsewhere is irrelevant. The use of a adjective along with a noun usually suggests that the adjective is adding additional information that is not implicit in the noun, and also usually does not mean that there are no instances where the same adjective would apply to other nouns, even ones that are semantically opposed to the one being used in the present sentence.
If it did, the phrase "snide creationist" would be redundant, so the mere use of the phrase suggests, indeed, that the GP realizes that.
If they said "Java" instead of "MySQL", that would be closer to being credible, but I doubt that Oracle spent the money for any one piece of IP.