Actually, git has a number of nice features for individual developers and even testers. There's a great thing called "bisect" for finding what broke the system. You tell it "this version worked, and that version didn't." Then it checks out for you another version to try. You build it, test it, and tell it if it works. Then it checks out another one for you. After O(log(n)) steps, it tells you which patch broke things. It's actually faster than trying to figure out which patch it was by thinking about the problem.
The reason this sort of thing is possible is that git is a free system, and people can write the tools they need to interact with it. It's true that most people didn't have to use BK, but git is useful in many more situations and there aren't the licensing reasons not to use it, so you're likely to see people tell you to use git just to move data around.
For that matter, git has extensions to the patch format, so you need to use it if you want to make certain types of patches in ways that are easy to review (such as a patch that renames a file; standard diff requires the reviewer to figure out what is different between the lines removed from the old name and the lines added to the new one).
Actually, software like databases is probably limited by memory bandwidth, not any execution unit in the processor. So hyperthreading won't help the database any, because it doesn't add any memory bandwidth. On the other hand, it probably does let other code running on the machine get more done while the database threads are waiting for data, gives more time to random Windows cruft at the expense of useful work in this case (that is, other processes get more quickly to the point of making memory accesses, and use more of the cache, since they can run while the database is waiting for data).
Ideally, get everybody development databases. Oracle basically lets you have them for free, and a small database on each workstation. If you can't run the complete system under single-user load on a workstation, there's no way you'll get it to run in production, unless you're Google.
If you expect unit tests to fail (especially common with XP, where you write tests for code you haven't implemented), make sure that reports don't list them as being a problem, because this just hides unexpected failures.
Language coding standards generally leave a bunch of flexibility. Eliminate all flexibility in your coding standards. There are better and worse options to choose, but non-uniformity is even worse than GNU^W^W^W^Wthe worst of all.
Don't comment your code. Comment your data structures. People can see what your code does by reading it (and can see what it should be doing by reading the spec). Your data structures just sit there and can be misinterpreted. Once you've explained in detail what your data means, you can comment tricky bits of code.
Use symbolic constants. If there are any two things which match by more than coincidence, you're missing a constant, and you're going to be in trouble when one of them changes. This includes comments.
Does your history teacher read the Globe? There were articles on the 25th (about Galvin opposing the decision), 29th (about Senator Pacheco's opposition), and 1st (about Quinn accepting responsibility for not getting buy-in from the disabled). Before that, there were presumably articles about the initial decision and Romney's support for it, but the search thing isn't going that far back. Even ignoring the actual issue entirely, I'm be surprised if a well-informed local wouldn't have noticed the fight over jurisdiction between the executive and legislative branches, or the posturing of the two most likely candidates for Governor next time.
This generates "weak collisions", which are where the attacker finds a pair of texts which hash to the same value, not "strong collisions", which are where the attacker finds a text that hashes to the same value as a text chosen by the user. So, someone could now set their password to some string, and then be able to type a different string to get in. (Except that neither are likely to be ascii text, making it a bit tough to type).
The actual issues are for document signing; the attacker could give you one document to sign, and use the signature on a different document with the same hash. There are smaller issues in the case where code expects no two documents to have the same hash. Obviously, collisions must exist, but the code to handle the case is likely not to be well-tested, since test cases were previously impractical to find.
In general, the maintainers want to avoid having to deal with a lot of stuff in a short time period, which means encouraging people to submit things at different times or have people not expecting a quick response. Linus is also partially making the point that if some feature gets excluded from a cycle, 6 weeks just isn't all that long to wait for it to get in the next cycle.
The first thing that came up is that Andrew Morton doesn't know which of the things in the testing tree are going in this cycle, because people haven't said anything in the middle of the window. So the first problem would be solved if everybody said at the beginning of the window that they're intending to submit something, but not quite yet.
The real issue, it turns out, is that James Bottomley (the SCSI maintainer) is in an untenable position. He's supposed to get patches from developers, integrate them, and submit them to Linus. But the developers waited for 2.6.14 to come out before they rebased their patches on it, and then took most of the window rebasing the patches, leaving James not enough time to determine for sure if all of these patches actually work correctly together before the window closes.
Furthermore, Linus would like to get infrastructure changes in before driver changes, so that it's easier to debug. It's hard to figure out what's wrong if you get both an infrastructure change that shouldn't affect anything in the driver, but might due to a bug and a driver change that should affect the driver behavior. It's a lot nicer if you can say whether the infrastructure change had any effect at all on what the driver does, and know right away which part to debug. But that depends on getting things included in a particular order.
I suspect that part of the problem is that the new release cycle is new, and contributors are not yet convinced that they can rebase their patches against 2.6.14-rc3 and have them apply cleanly to 2.6.14, and it will not be such an issue once people have gotten used to it.
The reason that there are so many musical credits is that this includes a lot of spoofs of songs from other things. Elton John and Andrew Lloyd Webber didn't write anything (intentionally) for this show, either. This isn't exactly serious drama, even for a musical.
Except, of course, that the ROKR isn't under the iPod brand. Apple calls it exclusively the "Motorola ROKR". It's marketted entirely as a Motorola product, with the only Apple trademark being iTunes. It's officially entirely unrelated to the iPod, except that they share the same DRM.
Actually, heavy weapons would just make ships more valuable targets for pirates. Cruise ship crews just aren't going to be very attentive, especially since pirate attacks are rare, and once they're in boarding range, rockets aren't going to be a good idea. And rockets would work really well against the sorts of ships that pirates target, so they'd be eager to get them. It's better to have an LRAD, which is effective at repelling the pirates, but not a weapon they'd be able to use effectively.
The current state of Linux is largely due to corporations. Only a small fraction of development has been done lately by people who aren't employed full time by companies to do it. On the other hand, it's a consortium effort, like Apache. Organizations that would benefit from Linux being a bit better in particular ways put in effort on those areas.
Getting it started was done by tinkerers and hobbyists, but once it became sufficiently important to employers, corporate involvement increased to be much larger than individual involvement.
Of course, this doesn't really fit well with venture capital. What I'd like to see would be for venture firms that invest in a number of related companies (like they often do) to invest in a new company which will take their venture money and produce software that the current companies will benefit from. The open source software company obviously loses all of the investment, having no income, but it provides a good return for the venture firm's portfolio in general.
Re:master of the obvious
on
Java Puzzlers
·
· Score: 1
Actually, the thing that's surprising is that (c == b) is false. You'd expect that the implicit conversion for equality testing would match the explicit one, whatever it is that it gives. Of course, the real issue is that Java really shouldn't have a "char" primitive, and certainly not one that accepts casts from numeric types.
Actually, that should be int8_t and uint16_t. A "char" in C can be signed or unsigned, and can be a platform-dependant number of bits. It's 8 bits on all current platforms, but that's just because 9- and 12-bit platforms have all died out. And a "short" in C can be any length that's no longer than an "int" and no shorter than a "char" (nothing can be shorter than a "char", because sizeof is defined in units of the size of char). Java defines bit lengths and such for all of the primitive types, and C only does this for the C99 integer types.
For that matter, Jesus was a left-wing radical who chewed out the relatively conservative and fundamentalist heads of organized religion. His big thing was really, "Forget about all this religious law stuff. Just be good people." That's about the least fundamentalist possible position. I'd put the start of fundamentalism at the Council of Nicea in 325, where dogma became something to be officially decided rather than discussed endlessly for the edification of the participants.
It depends on the exact lawsuit they lose. If it's just a product liability lawsuit, it's not that big a deal. If the court (or the insurance company) is convinced that the software was designed to damage your computer, that's not covered by insurance. The court could also issue useful orders, such as prohibiting Sony from installing things on people's computers or using DRM; this is also a possible outcome of a settlement, and their insurance company would probably push for them to accept a settlement.
In his book, it's 10/10, but the slashdot review format doesn't allow for ratings other than 8/10. Reviewers ask for the variable rating feature every once in a while, but not often enough for it to actually get implemented.
Java failed for applets because AWT was done without any clue about user interfaces or graphics, and was broken, underspecified, and hard to fit into programs. The problem wasn't getting data back and forth, or anything about openness. The problem was that having a user interact with your Java program horribly mangled both the program and the user, and you couldn't avoid it if you wanted to put Java on the client.
The smart thing might have been to do something closer to Ajax, and have the Java program send HTML to the browser, which would render it somewhat competently and at least consistantly with other things the user is looking at. Of course, browsers of the time were reasonably flaky (in particular, they weren't using an SGML/XML engine to get structured data, and would interpret some documents in ways inconsistant with the DTD, which would then screw up everything).
Personally, I just have the plugin removed normally. If I go to a site that requires flash, the pain of reinstalling the plugin is rarely as bad as the pain of using the site, so it's not a big deal to me. Or, if it's a good site, the flash thing is good enough to be worth dealing with the plugin. Of course, I have to remember to disable the plugin again when I'm done, but that's a minor issue.
They apply to somewhat different situations. UML is more for the case where you want to run a virtual environment under a normal environment, while Xen is more for when you want to run a bunch of parallel virtual environments. There's a certain amount of overlap in usage, but there are enough differences that both are worth having in the kernel, like, for instance, the loopback device and device mapper (both for the situation where you want a block device that isn't exactly a piece of hardware, but you don't build disk images with dm or do RAID on loopback). Ideally, UML and Xen will be able to share code for the common stuff (e.g., the virtual hardware).
Except, of course, that they've yet to release two versions of the OS that are actually compatible with each other. Try writing a shell script which changes the current directory and drive to the script's location and works on more than one version of Windows. It's possible, but only by ignoring syntax errors from commands that didn't work.
I think it's gotten somewhat better in the past few years, at least from what my DBA friends were telling me. Anyway, there are a number of database administration programs that work on an Oracle database in addition to other databases. IIRC, hsqldb's front end will work with any database that supports JDBC, and the professional DBA I know was using Toad to administer Oracle last time I checked. That's the nice thing about standard APIs: you can choose different things for each side without too much hassle.
It's worth mentioning that part of MA's plan has always been for everyone who already has MS Office to use it with a converter to generate suitable documents. So it's not only incorrect to say that the plan would require people to quit using MS Office, but the plan never even assumed that people would quit using it.
Another wrinkle here is that everybody hates Romney, especially the legislature. People are guessing that he'll mess up as much stuff as possible before he runs for president instead of running for re-election. Connecting the proponents of the plan "the Romney administration" is somewhat of an insult, even if it's technically standard to refer to the entire executive branch as the administration of its leader. But it's not surprising that, if the executive branch does anything, and says it doesn't need the legislature's approval, and Romney comes out in favor of it, and there's any contraversy at all about it, the legislature will want hearings and the candidates for governor will oppose it without even looking into the matter.
I bet those publishers don't accept unsolicited book proposals in LaTeX, though, even if it's a sample chapter or something. Or, at least, they probably shouldn't.
LaTeX is great for publication (including documents to be edited), but most business use of documents these days is communication with untrusted parties, so it's not really suitable as a format for document exchange.
I'd really like a LaTeX with TeX's layout engine, but without the Turing-complete and output-enabled scripting language accessible to documents. Really, all the exciting stuff should be in documentclasses and packages, and the particular document should be straightforward data (which includes calling some things, of course, but only the reasonable ones).
Those are, in fact, things that the South Korean government could demand if they found Microsoft to be a monopoly, and filings are supposed to mention such risks that investors take. As unlikely as it may seem to us, if Microsoft is prohibited by the South Korean government from selling Windows, or required to make modifications to the software before they sell it, they might actually comply with the order.
Actually, the problem with TeX is that it's hopelessly insecure; it is a programming language with no security model, giving arbitrary file access, with no support for sandboxing, or even code signing. It incidentally has optimized support for string constants and a library with a great layout engine. But the only reason it isn't banned as a security hole is that people don't generally accept LaTeX documents from other people, but rather have them written into PDF or PS.
Slashdot Mirror
Actually, git has a number of nice features for individual developers and even testers. There's a great thing called "bisect" for finding what broke the system. You tell it "this version worked, and that version didn't." Then it checks out for you another version to try. You build it, test it, and tell it if it works. Then it checks out another one for you. After O(log(n)) steps, it tells you which patch broke things. It's actually faster than trying to figure out which patch it was by thinking about the problem.
The reason this sort of thing is possible is that git is a free system, and people can write the tools they need to interact with it. It's true that most people didn't have to use BK, but git is useful in many more situations and there aren't the licensing reasons not to use it, so you're likely to see people tell you to use git just to move data around.
For that matter, git has extensions to the patch format, so you need to use it if you want to make certain types of patches in ways that are easy to review (such as a patch that renames a file; standard diff requires the reviewer to figure out what is different between the lines removed from the old name and the lines added to the new one).
Actually, software like databases is probably limited by memory bandwidth, not any execution unit in the processor. So hyperthreading won't help the database any, because it doesn't add any memory bandwidth. On the other hand, it probably does let other code running on the machine get more done while the database threads are waiting for data, gives more time to random Windows cruft at the expense of useful work in this case (that is, other processes get more quickly to the point of making memory accesses, and use more of the cache, since they can run while the database is waiting for data).
Ideally, get everybody development databases. Oracle basically lets you have them for free, and a small database on each workstation. If you can't run the complete system under single-user load on a workstation, there's no way you'll get it to run in production, unless you're Google.
If you expect unit tests to fail (especially common with XP, where you write tests for code you haven't implemented), make sure that reports don't list them as being a problem, because this just hides unexpected failures.
Language coding standards generally leave a bunch of flexibility. Eliminate all flexibility in your coding standards. There are better and worse options to choose, but non-uniformity is even worse than GNU^W^W^W^Wthe worst of all.
Don't comment your code. Comment your data structures. People can see what your code does by reading it (and can see what it should be doing by reading the spec). Your data structures just sit there and can be misinterpreted. Once you've explained in detail what your data means, you can comment tricky bits of code.
Use symbolic constants. If there are any two things which match by more than coincidence, you're missing a constant, and you're going to be in trouble when one of them changes. This includes comments.
I prefer:
//checks to see if x is equal to 256. If it is, then the code within is executed
if x==456 then
That way, there's less chance for confusion if the code gets modified in the future.
Does your history teacher read the Globe? There were articles on the 25th (about Galvin opposing the decision), 29th (about Senator Pacheco's opposition), and 1st (about Quinn accepting responsibility for not getting buy-in from the disabled). Before that, there were presumably articles about the initial decision and Romney's support for it, but the search thing isn't going that far back. Even ignoring the actual issue entirely, I'm be surprised if a well-informed local wouldn't have noticed the fight over jurisdiction between the executive and legislative branches, or the posturing of the two most likely candidates for Governor next time.
This generates "weak collisions", which are where the attacker finds a pair of texts which hash to the same value, not "strong collisions", which are where the attacker finds a text that hashes to the same value as a text chosen by the user. So, someone could now set their password to some string, and then be able to type a different string to get in. (Except that neither are likely to be ascii text, making it a bit tough to type).
The actual issues are for document signing; the attacker could give you one document to sign, and use the signature on a different document with the same hash. There are smaller issues in the case where code expects no two documents to have the same hash. Obviously, collisions must exist, but the code to handle the case is likely not to be well-tested, since test cases were previously impractical to find.
In general, the maintainers want to avoid having to deal with a lot of stuff in a short time period, which means encouraging people to submit things at different times or have people not expecting a quick response. Linus is also partially making the point that if some feature gets excluded from a cycle, 6 weeks just isn't all that long to wait for it to get in the next cycle.
The first thing that came up is that Andrew Morton doesn't know which of the things in the testing tree are going in this cycle, because people haven't said anything in the middle of the window. So the first problem would be solved if everybody said at the beginning of the window that they're intending to submit something, but not quite yet.
The real issue, it turns out, is that James Bottomley (the SCSI maintainer) is in an untenable position. He's supposed to get patches from developers, integrate them, and submit them to Linus. But the developers waited for 2.6.14 to come out before they rebased their patches on it, and then took most of the window rebasing the patches, leaving James not enough time to determine for sure if all of these patches actually work correctly together before the window closes.
Furthermore, Linus would like to get infrastructure changes in before driver changes, so that it's easier to debug. It's hard to figure out what's wrong if you get both an infrastructure change that shouldn't affect anything in the driver, but might due to a bug and a driver change that should affect the driver behavior. It's a lot nicer if you can say whether the infrastructure change had any effect at all on what the driver does, and know right away which part to debug. But that depends on getting things included in a particular order.
I suspect that part of the problem is that the new release cycle is new, and contributors are not yet convinced that they can rebase their patches against 2.6.14-rc3 and have them apply cleanly to 2.6.14, and it will not be such an issue once people have gotten used to it.
The reason that there are so many musical credits is that this includes a lot of spoofs of songs from other things. Elton John and Andrew Lloyd Webber didn't write anything (intentionally) for this show, either. This isn't exactly serious drama, even for a musical.
Except, of course, that the ROKR isn't under the iPod brand. Apple calls it exclusively the "Motorola ROKR". It's marketted entirely as a Motorola product, with the only Apple trademark being iTunes. It's officially entirely unrelated to the iPod, except that they share the same DRM.
Actually, heavy weapons would just make ships more valuable targets for pirates. Cruise ship crews just aren't going to be very attentive, especially since pirate attacks are rare, and once they're in boarding range, rockets aren't going to be a good idea. And rockets would work really well against the sorts of ships that pirates target, so they'd be eager to get them. It's better to have an LRAD, which is effective at repelling the pirates, but not a weapon they'd be able to use effectively.
The current state of Linux is largely due to corporations. Only a small fraction of development has been done lately by people who aren't employed full time by companies to do it. On the other hand, it's a consortium effort, like Apache. Organizations that would benefit from Linux being a bit better in particular ways put in effort on those areas.
Getting it started was done by tinkerers and hobbyists, but once it became sufficiently important to employers, corporate involvement increased to be much larger than individual involvement.
Of course, this doesn't really fit well with venture capital. What I'd like to see would be for venture firms that invest in a number of related companies (like they often do) to invest in a new company which will take their venture money and produce software that the current companies will benefit from. The open source software company obviously loses all of the investment, having no income, but it provides a good return for the venture firm's portfolio in general.
Actually, the thing that's surprising is that (c == b) is false. You'd expect that the implicit conversion for equality testing would match the explicit one, whatever it is that it gives. Of course, the real issue is that Java really shouldn't have a "char" primitive, and certainly not one that accepts casts from numeric types.
Actually, that should be int8_t and uint16_t. A "char" in C can be signed or unsigned, and can be a platform-dependant number of bits. It's 8 bits on all current platforms, but that's just because 9- and 12-bit platforms have all died out. And a "short" in C can be any length that's no longer than an "int" and no shorter than a "char" (nothing can be shorter than a "char", because sizeof is defined in units of the size of char). Java defines bit lengths and such for all of the primitive types, and C only does this for the C99 integer types.
For that matter, Jesus was a left-wing radical who chewed out the relatively conservative and fundamentalist heads of organized religion. His big thing was really, "Forget about all this religious law stuff. Just be good people." That's about the least fundamentalist possible position. I'd put the start of fundamentalism at the Council of Nicea in 325, where dogma became something to be officially decided rather than discussed endlessly for the edification of the participants.
It depends on the exact lawsuit they lose. If it's just a product liability lawsuit, it's not that big a deal. If the court (or the insurance company) is convinced that the software was designed to damage your computer, that's not covered by insurance. The court could also issue useful orders, such as prohibiting Sony from installing things on people's computers or using DRM; this is also a possible outcome of a settlement, and their insurance company would probably push for them to accept a settlement.
In his book, it's 10/10, but the slashdot review format doesn't allow for ratings other than 8/10. Reviewers ask for the variable rating feature every once in a while, but not often enough for it to actually get implemented.
Java failed for applets because AWT was done without any clue about user interfaces or graphics, and was broken, underspecified, and hard to fit into programs. The problem wasn't getting data back and forth, or anything about openness. The problem was that having a user interact with your Java program horribly mangled both the program and the user, and you couldn't avoid it if you wanted to put Java on the client.
The smart thing might have been to do something closer to Ajax, and have the Java program send HTML to the browser, which would render it somewhat competently and at least consistantly with other things the user is looking at. Of course, browsers of the time were reasonably flaky (in particular, they weren't using an SGML/XML engine to get structured data, and would interpret some documents in ways inconsistant with the DTD, which would then screw up everything).
Personally, I just have the plugin removed normally. If I go to a site that requires flash, the pain of reinstalling the plugin is rarely as bad as the pain of using the site, so it's not a big deal to me. Or, if it's a good site, the flash thing is good enough to be worth dealing with the plugin. Of course, I have to remember to disable the plugin again when I'm done, but that's a minor issue.
They apply to somewhat different situations. UML is more for the case where you want to run a virtual environment under a normal environment, while Xen is more for when you want to run a bunch of parallel virtual environments. There's a certain amount of overlap in usage, but there are enough differences that both are worth having in the kernel, like, for instance, the loopback device and device mapper (both for the situation where you want a block device that isn't exactly a piece of hardware, but you don't build disk images with dm or do RAID on loopback). Ideally, UML and Xen will be able to share code for the common stuff (e.g., the virtual hardware).
Except, of course, that they've yet to release two versions of the OS that are actually compatible with each other. Try writing a shell script which changes the current directory and drive to the script's location and works on more than one version of Windows. It's possible, but only by ignoring syntax errors from commands that didn't work.
I think it's gotten somewhat better in the past few years, at least from what my DBA friends were telling me. Anyway, there are a number of database administration programs that work on an Oracle database in addition to other databases. IIRC, hsqldb's front end will work with any database that supports JDBC, and the professional DBA I know was using Toad to administer Oracle last time I checked. That's the nice thing about standard APIs: you can choose different things for each side without too much hassle.
It's worth mentioning that part of MA's plan has always been for everyone who already has MS Office to use it with a converter to generate suitable documents. So it's not only incorrect to say that the plan would require people to quit using MS Office, but the plan never even assumed that people would quit using it.
Another wrinkle here is that everybody hates Romney, especially the legislature. People are guessing that he'll mess up as much stuff as possible before he runs for president instead of running for re-election. Connecting the proponents of the plan "the Romney administration" is somewhat of an insult, even if it's technically standard to refer to the entire executive branch as the administration of its leader. But it's not surprising that, if the executive branch does anything, and says it doesn't need the legislature's approval, and Romney comes out in favor of it, and there's any contraversy at all about it, the legislature will want hearings and the candidates for governor will oppose it without even looking into the matter.
I bet those publishers don't accept unsolicited book proposals in LaTeX, though, even if it's a sample chapter or something. Or, at least, they probably shouldn't.
LaTeX is great for publication (including documents to be edited), but most business use of documents these days is communication with untrusted parties, so it's not really suitable as a format for document exchange.
I'd really like a LaTeX with TeX's layout engine, but without the Turing-complete and output-enabled scripting language accessible to documents. Really, all the exciting stuff should be in documentclasses and packages, and the particular document should be straightforward data (which includes calling some things, of course, but only the reasonable ones).
Those are, in fact, things that the South Korean government could demand if they found Microsoft to be a monopoly, and filings are supposed to mention such risks that investors take. As unlikely as it may seem to us, if Microsoft is prohibited by the South Korean government from selling Windows, or required to make modifications to the software before they sell it, they might actually comply with the order.
Actually, the problem with TeX is that it's hopelessly insecure; it is a programming language with no security model, giving arbitrary file access, with no support for sandboxing, or even code signing. It incidentally has optimized support for string constants and a library with a great layout engine. But the only reason it isn't banned as a security hole is that people don't generally accept LaTeX documents from other people, but rather have them written into PDF or PS.