Frankly, I couldn't give a **** if it's OS or not, because I'm never going to have time to read all the source and make sure I agree with it. I bet you aren't either.
The advantage of open source is not only that you can read the code, but that anybody can. Because of this, you can be somewhat more sure that development will continue for at least as long as you need it.
However, rather than assuming that anything closed source will be worse than the OS competition, I'm prepared to take a look at how it performs, evaluate it using meaningful criteria, and base my choices on the results.
I don't think that open source proponents automatically assume that a closed source alternative is worse in every way. They prefer open source for other reasons (one of them being the one stated above). I think that it is well known and accepted that GCC doesn't perform as well as the Intel or Microsoft compilers on the X86 architecture so nobody cares if you are "prepared to take a look at how it performs, [blah, blah, textbook definition of performance analysis, blah]".
More people die from traffic accidents than accidental shootings.
Um, this may have to do with the fact that cars are driven in public more often than guns are shot in public. But, hey, if it doesn't support your argument, then common sense be damned, right?
Wow. Now I have a term to describe Linux package management. Package managers, at least in my experience, are the epitome of leaky abstractions. I am now able to do the things that I want to do with my package system...but only after learning how the Debian packaging tools work (yeah, yeah, I'm sure Gentoo, RPM, Sorceror and BSD ports are all much easier). I've found that this is also true for Linux (probably Unix in general) configuration programs that try and add a later of abstraction above plain text files.
It is, in my opinion obviously, the most significant shortcoming of Linux distributions. We really need a comprehensive and airtight software/configuration management infrastructure for Unix-like systems. Um...and I nominate somebody else to do it:)
Yes, that would suck. I type weird things into documents and source code when I'm idle.
An application, however, could provide the option of encrypting the delta information. This option should probably be turned on by default to avoid mistakes such as the one you describe. Again, though, if the application isn't designed properly, just managing this crap could turn into another headache.
Xdelta is a program that can generate deltas on binary files. Since it uses the copy/insert method for generating deltas, it is often more efficient than normal diff (especially for transpositions, which are relatively common in source code and other regular text).
Since text files are a subset of binary files, Xdelta can handle these too. The reason people still use diff is that the output is readable. Xdelta output isn't readable, but readability doesn't matter if you just want to store revisions. A document editor (or whatever) can use the Xdelta method to store revisions and reproduce older versions. Then, it can use some custom highlighting or formatting to display the changes to the user. I've seen tools that do this with regular diff output to make it easier to understand.
What are you trying to say here? I'm assuming you're tone is sarcastic, but that doesn't seem to make sense because the parent post didn't blame Windows. If anything, the blame is placed on "software developers who aren't aware...". If you are not being sarcastic, then you are well on your way to becoming the stereotypical Slashdot poster.
I used to use vim before emacs, which is a very good editor as well, but I do not want to compile in some shell and follow the error line numbers manually.
While I don't mind using a separate shell to compile, Vim's ":make" command will run your Makefile and Vim can be set up to parse the output to let you jump between error-causing lines automatically.
You need only memorize or add keybindings for the commands you use most.
This is mostly the same for Vim. However, I've found Vim's help system easier to navigate through and find the command you're looking for.
I'm pretty sure C doesn't have compile-time dependencies in the sense that some files must be compiled before others (at least with everything I've done). Since the interface and implementation are in different files, only the contents of the header file are required to produce object files (and even this isn't even a strict requirement and apparently only warrants a warning). In a language like Java, however, there is such a dependency because of the mix of interface and implementation (I don't mean Java interfaces here, I mean method/variable signatures). I don't know about C++.
While linking, however, there are dependencies because that's when the actual function call addresses are plugged in (Java doesn't even do this till runtime which probably accounts for some of it's startup sluggishness).
I agree that Amdahl's law is completely true in it's stated context, but there are cases where, because of the increased total memory of a cluster, a clustered parallel implementation will achieve a speedup greater than N.
That said, I don't think that a distributed compile is such a case (though maybe optimization caching could take advantage of the extra memory).
100 MBit switch with 10 MBit NICs? Wouldn't that dumb the switch down? How much slower would a 10 MBit switch be? Please explain, I'm a low layer network newbie.
I think the reason that click-wrap agreements are annoying in commercial software is the fact that you are being presented with restrictions on something you have already paid for.
Yes...I know that open source doesn't necessarily mean free of charge, but a click-wrap agreement for these is more like "here's the terms we're offering the software on, take it or leave it". Not much harm done here.
Of course not having a click-wrap agreement would be least annoying, but apparently it has some legal advantages.
Regarding the specific case of the JDK, even before the shell script agreement, I think you have to accept another agreement on their website before you are even allowed to download it. This agreement could say something about modifying the downloaded program. I guess you could also modify the webpage but as someone already said, you can't get away with modifying a contract anyway.
If assert() were made part of the language, maybe the compiler could make sure that there are no side-effects in the expression. Java's assert is part of the language but unlike C++, it doesn't have the 'const' modifier to mark whether functions have side-effects or not. Could a compiler-checked assert be added to C++ (kind of like the format of printf is checked by the compiler, even though it's not really part of the language)?
String x = "anything";
if(x == "anything") doSomething();
Actually, the 'doSomething' method will be executed. Because Java guarantees (I think) that String literals of equal values will be assigned the same object, you can sometimes use '==' to speed things up. But yeah...if you read 'x' in from a file and compared it to a literal, it would be false.
MacOS has various ways of hiding and showing certain windows, including "window shading" and "pop up" windows. You can also properly HIDE applications entirely if you'd like.
Window shading is a dumb feature (yeah, yeah, my opinion). I think minimizing and using the taskbar is a much better way to do things. Do you really find shading/whole-app-only switching a good thing or did you just say that so that every point of mine was responded to?
Nope, you press the Control key to bring up contextual menus. Why is using click modifier keys such a pain? What are you doing with your other hand that doesn't involve the keyboard?
Regardless of what I'm doing with my other hand (you're a funny guy), there's a reason your mouse has a button in the first place. Why not just add another key to the keyboard? Because it's stupid. There are five fingers on your hand and while some are used for grip, I'm sure even apple users can handle two buttons. I think that it is natural to make implicit associations between the pointer and the mouse and so using the mouse buttons to control different types of clicks is just not as easy? I can't believe you prefer single button mice (unless, of course, you are impulsively arguing in defense of your cult leader:-).
MacOS has a device called the "application switcher" that tears off from that menu, and can be placed anywhere on the desktop as a floating palette.
Yeah, I've seen the application switcher. But I honestly do prefer the windows taskbar. It doesn't get in the way and, especially with browsers and telnet windows, it's nice to see right away what windows are present.
Merely holding the Option key will close the previous window automatically. You can combine this with tie window title bar hierarchy to move very swiftly and cleanly - much faster than Windows, in fact.
This I'll need to try. This window title bar heirarchy...is this at the top of the finder window? Is that what most people use because the alternative (opening window after window) really sucks. I'll have to try this next time I'm in the Mac labs. I can't imagine it being faster than the Start menu....but I'll try it before I say anything stupid:-).
I don't think Windows is the greatest UI in the world but I also don't think that the Apple UI is everything people seem to blindly proclaim it is. My limited usage of Macs has something to do with it but c'mon, one button mice?
MacOS 8 and 9 suck so hard. I don't know if the retarded behaviour can be configured out, but the defaults are ridiculous:
- Can't have a few windows from one application and a few from another at the top. If you have a bunch of browser windows, clicking on one brings them all to the top.
- The Start menu is so much better for launching programs. The idea of having to open folder after folder of windows that stay open is worse than the Win 3.1 Program Manager.
- Apple made a conscious decision to avoid 2-button mice for so long. The context menu right-click is so useful in windows. Yeah, yeah, you can press the apple key to fake it, but that is a pain.
- The lack of a taskbar (remember, we're talking about OS 8 and 9 here) makes it hard to switch between programs. The peice-of-junk menu at the top right is no substitute.
- Finder is much worse than Windows Explorer.
- Applications that don't quit even after closing all Windows is a weird feature. I can see how sometimes it may be useful but most of the time it just gets in the way. The OS even detects that and pops up a dialog to ask you if you want to close it. This is a sure sign of bad design.
This new OS isn't any different. The only difference is that it still hasn't been fully optimized yet, and so it continues to be speedier at each OS upgrade.
I don't mean to flame, but this is exactly what I dislike about some people in the Mac community. They regurgitate everything Jobs tells them as if speaking from first-hand experience.
Have you seen the source code to OS X and realized that there is a lot of room for optimization?
Do you realize that you are paying (in terms of hardware and time) for the slow and completely useless genie effect? Do you realize that Apple advertising is geared towards, well, whoever likes to watch colorful computers skate around on the TV screen?
Apple wants to be 'hip' and if that's all you expect from a computer company, that's fine. Going around calling PC's expensive, and basing comparisons on resale value is not right. I can see how old Macs may fetch a higher price: the key element, the appearance, doesn't degrade as quickly as the key element in PCs, which is relative performance.
And if the forwarding is well-written, most/all of it can be inlined and otherwise optimized away. Voila -- 15 containers, all sharing code.
But the process of inlining will generate a new copy of the generic code for every data type. If inlining is not done, you don't get the speedups associated with using templates over void* data types.
However, the instant you start giving copies to other people, you move into the realm of copyright infringement. The only thing that allows you to distribute copies is the GPL, which means you either distribute by its terms or don't distribute at all.
The Plan 9 license says:
Distribution of Licensed Software to third parties pursuant to this grant shall be subject to the same terms and conditions as set forth in this Agreement,
I still don't see how this clause is Bad since it only applies to redistribution. Somebody clue me in.
Does anybody remember when Intel tried to leverage the P2's image off of the average idiot's perception of the Internet. All their ads said that the P2 "let's you unleash the power of the internet". This was way back when most people still used dialup (when the bottleneck was definitely not the processor). Most sites still use plain old HTML (as opposed to the much hyped VRML) for everything.
Slashdot Mirror
The advantage of open source is not only that you can read the code, but that anybody can. Because of this, you can be somewhat more sure that development will continue for at least as long as you need it.
I don't think that open source proponents automatically assume that a closed source alternative is worse in every way. They prefer open source for other reasons (one of them being the one stated above). I think that it is well known and accepted that GCC doesn't perform as well as the Intel or Microsoft compilers on the X86 architecture so nobody cares if you are "prepared to take a look at how it performs, [blah, blah, textbook definition of performance analysis, blah]".
I'm pretty sure that the episode was a Segway parody. I think he called the device "it", the same thing Segway was referred previously reffered to as.
Um, this may have to do with the fact that cars are driven in public more often than guns are shot in public. But, hey, if it doesn't support your argument, then common sense be damned, right?
Wow. Now I have a term to describe Linux package management. Package managers, at least in my experience, are the epitome of leaky abstractions. I am now able to do the things that I want to do with my package system...but only after learning how the Debian packaging tools work (yeah, yeah, I'm sure Gentoo, RPM, Sorceror and BSD ports are all much easier). I've found that this is also true for Linux (probably Unix in general) configuration programs that try and add a later of abstraction above plain text files.
It is, in my opinion obviously, the most significant shortcoming of Linux distributions. We really need a comprehensive and airtight software/configuration management infrastructure for Unix-like systems. Um...and I nominate somebody else to do it :)
Yes, that would suck. I type weird things into documents and source code when I'm idle.
An application, however, could provide the option of encrypting the delta information. This option should probably be turned on by default to avoid mistakes such as the one you describe. Again, though, if the application isn't designed properly, just managing this crap could turn into another headache.
Xdelta is a program that can generate deltas on binary files. Since it uses the copy/insert method for generating deltas, it is often more efficient than normal diff (especially for transpositions, which are relatively common in source code and other regular text).
Since text files are a subset of binary files, Xdelta can handle these too. The reason people still use diff is that the output is readable. Xdelta output isn't readable, but readability doesn't matter if you just want to store revisions. A document editor (or whatever) can use the Xdelta method to store revisions and reproduce older versions. Then, it can use some custom highlighting or formatting to display the changes to the user. I've seen tools that do this with regular diff output to make it easier to understand.
What are you trying to say here? I'm assuming you're tone is sarcastic, but that doesn't seem to make sense because the parent post didn't blame Windows. If anything, the blame is placed on "software developers who aren't aware...". If you are not being sarcastic, then you are well on your way to becoming the stereotypical Slashdot poster.
While I don't mind using a separate shell to compile, Vim's ":make" command will run your Makefile and Vim can be set up to parse the output to let you jump between error-causing lines automatically.
This is mostly the same for Vim. However, I've found Vim's help system easier to navigate through and find the command you're looking for.
I'm pretty sure C doesn't have compile-time dependencies in the sense that some files must be compiled before others (at least with everything I've done). Since the interface and implementation are in different files, only the contents of the header file are required to produce object files (and even this isn't even a strict requirement and apparently only warrants a warning). In a language like Java, however, there is such a dependency because of the mix of interface and implementation (I don't mean Java interfaces here, I mean method/variable signatures). I don't know about C++.
While linking, however, there are dependencies because that's when the actual function call addresses are plugged in (Java doesn't even do this till runtime which probably accounts for some of it's startup sluggishness).
I agree that Amdahl's law is completely true in it's stated context, but there are cases where, because of the increased total memory of a cluster, a clustered parallel implementation will achieve a speedup greater than N.
That said, I don't think that a distributed compile is such a case (though maybe optimization caching could take advantage of the extra memory).
100 MBit switch with 10 MBit NICs? Wouldn't that dumb the switch down? How much slower would a 10 MBit switch be? Please explain, I'm a low layer network newbie.
Wow! Why aren't the parent post's links followed by "[domain.name]"? Is his karma super high.
10 seconds?
:)
Results 1 - 10 of about 526. Search took 0.13 seconds.
Is "symantics" fair game? (not being strictly part of the joke and all)
I think the reason that click-wrap agreements are annoying in commercial software is the fact that you are being presented with restrictions on something you have already paid for. Yes...I know that open source doesn't necessarily mean free of charge, but a click-wrap agreement for these is more like "here's the terms we're offering the software on, take it or leave it". Not much harm done here. Of course not having a click-wrap agreement would be least annoying, but apparently it has some legal advantages.
Regarding the specific case of the JDK, even before the shell script agreement, I think you have to accept another agreement on their website before you are even allowed to download it. This agreement could say something about modifying the downloaded program. I guess you could also modify the webpage but as someone already said, you can't get away with modifying a contract anyway.
If assert() were made part of the language, maybe the compiler could make sure that there are no side-effects in the expression. Java's assert is part of the language but unlike C++, it doesn't have the 'const' modifier to mark whether functions have side-effects or not. Could a compiler-checked assert be added to C++ (kind of like the format of printf is checked by the compiler, even though it's not really part of the language)?
Actually, the 'doSomething' method will be executed. Because Java guarantees (I think) that String literals of equal values will be assigned the same object, you can sometimes use '==' to speed things up. But yeah...if you read 'x' in from a file and compared it to a literal, it would be false.
But isn't that what temporary-insanity-like defences bet on?
Window shading is a dumb feature (yeah, yeah, my opinion). I think minimizing and using the taskbar is a much better way to do things. Do you really find shading/whole-app-only switching a good thing or did you just say that so that every point of mine was responded to?
Regardless of what I'm doing with my other hand (you're a funny guy), there's a reason your mouse has a button in the first place. Why not just add another key to the keyboard? Because it's stupid. There are five fingers on your hand and while some are used for grip, I'm sure even apple users can handle two buttons. I think that it is natural to make implicit associations between the pointer and the mouse and so using the mouse buttons to control different types of clicks is just not as easy? I can't believe you prefer single button mice (unless, of course, you are impulsively arguing in defense of your cult leader
Yeah, I've seen the application switcher. But I honestly do prefer the windows taskbar. It doesn't get in the way and, especially with browsers and telnet windows, it's nice to see right away what windows are present.
This I'll need to try. This window title bar heirarchy...is this at the top of the finder window? Is that what most people use because the alternative (opening window after window) really sucks. I'll have to try this next time I'm in the Mac labs. I can't imagine it being faster than the Start menu....but I'll try it before I say anything stupid
I don't think Windows is the greatest UI in the world but I also don't think that the Apple UI is everything people seem to blindly proclaim it is. My limited usage of Macs has something to do with it but c'mon, one button mice?
MacOS 8 and 9 suck so hard. I don't know if the retarded behaviour can be configured out, but the defaults are ridiculous:
- Can't have a few windows from one application and a few from another at the top. If you have a bunch of browser windows, clicking on one brings them all to the top.
- The Start menu is so much better for launching programs. The idea of having to open folder after folder of windows that stay open is worse than the Win 3.1 Program Manager.
- Apple made a conscious decision to avoid 2-button mice for so long. The context menu right-click is so useful in windows. Yeah, yeah, you can press the apple key to fake it, but that is a pain.
- The lack of a taskbar (remember, we're talking about OS 8 and 9 here) makes it hard to switch between programs. The peice-of-junk menu at the top right is no substitute.
- Finder is much worse than Windows Explorer.
- Applications that don't quit even after closing all Windows is a weird feature. I can see how sometimes it may be useful but most of the time it just gets in the way. The OS even detects that and pops up a dialog to ask you if you want to close it. This is a sure sign of bad design.
I don't mean to flame, but this is exactly what I dislike about some people in the Mac community. They regurgitate everything Jobs tells them as if speaking from first-hand experience.
Have you seen the source code to OS X and realized that there is a lot of room for optimization?
Do you realize that you are paying (in terms of hardware and time) for the slow and completely useless genie effect? Do you realize that Apple advertising is geared towards, well, whoever likes to watch colorful computers skate around on the TV screen?
Apple wants to be 'hip' and if that's all you expect from a computer company, that's fine. Going around calling PC's expensive, and basing comparisons on resale value is not right. I can see how old Macs may fetch a higher price: the key element, the appearance, doesn't degrade as quickly as the key element in PCs, which is relative performance.
But the process of inlining will generate a new copy of the generic code for every data type. If inlining is not done, you don't get the speedups associated with using templates over void* data types.
The Plan 9 license says:
I still don't see how this clause is Bad since it only applies to redistribution. Somebody clue me in.
Does anybody remember when Intel tried to leverage the P2's image off of the average idiot's perception of the Internet. All their ads said that the P2 "let's you unleash the power of the internet". This was way back when most people still used dialup (when the bottleneck was definitely not the processor). Most sites still use plain old HTML (as opposed to the much hyped VRML) for everything.