There is nothing preventing Microsoft from taking Linux, slapping a binary copy of some of their shared libraries plus the Windows Explorer on top of it, "improving" the API's of Linux to support it, provided that they will provide the source for the bits they took from Linux for free.
There is no requirement for MS to wait for their patches to be accepted back into mainstream Linux, so hey presto: a split in Linux. It would not give anyone access to the source of the MS crown jewels, which are not integrated but are just distributed with it in binary.
The BSD/MIT and GPL license communities are hurting each other by disallowing developers to lift the good parts of competing implementations only for legalistic reasons. Fortunately, the sheer numbers of open source developers gives us the edge here, but it sure wastes my time. I contribute to both GPL, MPL/NPL and BSD licensed projects, and the time wasted on license issues could otherwise have gone to improving the state of the project. Sigh.
I'm not worried about Microsoft stealing BSD code because the license alows them. I'm worried about them writing shoddy re-implementations (MS TCP/IP stack, anyone?) rather than stealing good code. GPL is no deterrent against any big player messing up open source in a world where buying decisions are taken by cluely challenged executives, and nothing is going to change that reality.
Very well said. I've done my share of UI design, and well... it sucked, and only got better after listening to the end users (who were quite powerful and therefore able to force met to listen) to fix it.
The most important advantage open software has is the ability to fix things. In my experience, most developers (be they Gnome, KDE or app developers) do listen when improvements are suggested.
For example, another thread mentioned hierarchical menus which allow the user no slack in slightly moving off the menu without losing the menu. This is a major nuisance. If someone with a solid UI background mentions such a thing to the developers, they can put it on their list of things to do, and better: someone who is annoyed enough can actually write it him/herself and contribute it for all to use.
I am constantly amazed at the rapid pace with which user interface improvements find their way into toolkits. All it takes is the realization that something needs to be done. Coders are generally bad at identifying those points, but they do understand the importance and do take pride in implementing improvements (and that is what drives open source development, after all)
Hmmm, I must have missed the BSD bashing by Eric Raymond.
When I read The Cathedral and the Bazaar (the dead tree edition), it struck me how well he presented the BSD case in a world where few people look beyond the CNN headlines.
<ObTroll> I disagree with Eric Raymonds insinuation that the BSD license caused the fragmentation in the BSD camp. The legalistic battles don't deserve the credit this insinuation gives them. I've always written the fragmentation off to too many strong egos. <ObTroll>
Jordan Hubbard, as well as FreeBSD, deserve more credit than they sometimes get. Bickering about who was first to break the ice, who is more Open or more Free is entirely counterproductive. I'm still torn on the question of whether world dominance by Linux would be a good thing. On the one hand, I hate to see the fragmentation this world sees, with its duplicated efforts on writing device drivers and stuff. But I'm still leaning to the thought that survival of BSD as a significant underflow is a good thing, in that it broadens the gene pool without undermining Open Source.
Linux brought legitimacy to Open Source, which allows me as a BSD/OS user to run cool stuff like Gnumeric or GIMP. Enough folks from non-Linux communities help bear the weight of building and maintaining such tools, and the broad gene pool underneath such apps bring bugs to light that wouldn't have surfaced as fast if all the world ran Linux with glibc 2.1, which in turn eases the pain of future Linux upgrades.
Let me make a confession here. The reason that I contribute small, rounded off patches to Open Source projects (either a bug fix or one (1) feature), is that I know how much work goes into packaging something big, or just into merging patches like the ones I write (irrespective of how well these patches are done). There are too many projects I care about and work on, and too little time, and this must certainly be even more true for the slashdot crew.
I think Sun could put more effort into merging work on the JDK done by the community, but I wouldn't like to be in the shoes of the Sun manager who has got to get funding to hire folks to actually do it. And I have great respect for the folks at Netscape/Mozilla, who go through the daunting task of integrating patches in the face of criticism for not putting everything they get in on the same day.
So, to counter Eric Raymonds assertion that releasing early & often is the ticket to success, I'd like to quote Wietze Venema (I hope; apologies if I goof up the quote): "It is ready when it's ready".
On a few occasions, I've been referred to slashdot by people I was surprised to find had Linux on their radar. That in itself is a major achievement, and if the only price I pay for that service is to put some brainpower in myself when I need to build a user community site in Internet time (which could also be phrased as "ripping off slashdot in a hurry"), so be it.
The question was not who contributed most code, or who has maximized the integral over time of code contributions. RMS is a major contender and probable even winner, and the computer industry wouldn't have been where it is now if it weren't for the two great tools of GCC and Emacs; tools that have made my life as a developer and sysadmin bearable over the last ten years.
The question was: who was the best Open Source advocate. IMHO, RMS is a loose cannon who is very good at galvanizing the body of like minded persons (think rebel leader in a divided nation), but a danger to their acceptance (think foreign politics).
Perhaps an example from personal experience can paint the picture more clearly from my viewpoint. I once took a day off and drove the length of our country (not that it's that big) to hear RMS speak in Amsterdam on patents. This was in the early days of the revolt against patents on software, and I agreed with every word he said. But his mode of speach was such that I decided there and then not to associate myself with the LPF, because I did not wanted to be painted with the same brush that he would be painted with.
My non-membership of the LPF was (and is) no big deal. I publicly and vocally express the LPF view at many occasions. But at a time when we need division between equally minded people like we need holes in our heads, putting RMS in the limelight for his advocacy does not bode well for freely available software.
I wished the lawyers would stay out of it. The whole mess we got into was caused by someone slapping a copyright on code labeled Public Domain. If just a percentage of all the productivity lost in the world over license debates since would've been put into lobbying (or outright buying) Congress to change the laws to make the term Public Domain mean what every programmer understood it to mean back then, Linux would've taken over the world five years ago. Except it may have been called BSD or Minix in that case.
Well, the thing that always gets me is that the age old friends of the sysadmin, the curses based tools like top, are consistently overlooked. Yes, ktop and friends are kewl, but if the system is really screwed and you're happy to squeeze a few cycles or bytes out, top really saves the day.
Same for ntop. If you're remotely logging into a site behind an overloaded circuit and a few bytes of telnet are the max you can squeeze through it, you still can figure out which connection is hogging the link.
In years of systems and network monitoring, I've yet to encounter a more effective user interface.
Slashdot Mirror
There is no requirement for MS to wait for their patches to be accepted back into mainstream Linux, so hey presto: a split in Linux. It would not give anyone access to the source of the MS crown jewels, which are not integrated but are just distributed with it in binary.
The BSD/MIT and GPL license communities are hurting each other by disallowing developers to lift the good parts of competing implementations only for legalistic reasons. Fortunately, the sheer numbers of open source developers gives us the edge here, but it sure wastes my time. I contribute to both GPL, MPL/NPL and BSD licensed projects, and the time wasted on license issues could otherwise have gone to improving the state of the project. Sigh.
I'm not worried about Microsoft stealing BSD code because the license alows them. I'm worried about them writing shoddy re-implementations (MS TCP/IP stack, anyone?) rather than stealing good code. GPL is no deterrent against any big player messing up open source in a world where buying decisions are taken by cluely challenged executives, and nothing is going to change that reality.
The most important advantage open software has is the ability to fix things. In my experience, most developers (be they Gnome, KDE or app developers) do listen when improvements are suggested.
For example, another thread mentioned hierarchical menus which allow the user no slack in slightly moving off the menu without losing the menu. This is a major nuisance. If someone with a solid UI background mentions such a thing to the developers, they can put it on their list of things to do, and better: someone who is annoyed enough can actually write it him/herself and contribute it for all to use.
I am constantly amazed at the rapid pace with which user interface improvements find their way into toolkits. All it takes is the realization that something needs to be done. Coders are generally bad at identifying those points, but they do understand the importance and do take pride in implementing improvements (and that is what drives open source development, after all)
When I read The Cathedral and the Bazaar (the dead tree edition), it struck me how well he presented the BSD case in a world where few people look beyond the CNN headlines.
<ObTroll>
I disagree with Eric Raymonds insinuation that the BSD license caused the fragmentation in the BSD camp. The legalistic battles don't deserve the credit this insinuation gives them. I've always written the fragmentation off to too many strong egos.
<ObTroll>
Jordan Hubbard, as well as FreeBSD, deserve more credit than they sometimes get. Bickering about who was first to break the ice, who is more Open or more Free is entirely counterproductive. I'm still torn on the question of whether world dominance by Linux would be a good thing. On the one hand, I hate to see the fragmentation this world sees, with its duplicated efforts on writing device drivers and stuff. But I'm still leaning to the thought that survival of BSD as a significant underflow is a good thing, in that it broadens the gene pool without undermining Open Source.
Linux brought legitimacy to Open Source, which allows me as a BSD/OS user to run cool stuff like Gnumeric or GIMP. Enough folks from non-Linux communities help bear the weight of building and maintaining such tools, and the broad gene pool underneath such apps bring bugs to light that wouldn't have surfaced as fast if all the world ran Linux with glibc 2.1, which in turn eases the pain of future Linux upgrades.
United we stand. Divided we fall.
Let me make a confession here. The reason that I contribute small, rounded off patches to Open Source projects (either a bug fix or one (1) feature), is that I know how much work goes into packaging something big, or just into merging patches like the ones I write (irrespective of how well these patches are done). There are too many projects I care about and work on, and too little time, and this must certainly be even more true for the slashdot crew.
I think Sun could put more effort into merging work on the JDK done by the community, but I wouldn't like to be in the shoes of the Sun manager who has got to get funding to hire folks to actually do it. And I have great respect for the folks at Netscape/Mozilla, who go through the daunting task of integrating patches in the face of criticism for not putting everything they get in on the same day.
So, to counter Eric Raymonds assertion that releasing early & often is the ticket to success, I'd like to quote Wietze Venema (I hope; apologies if I goof up the quote): "It is ready when it's ready".
On a few occasions, I've been referred to slashdot by people I was surprised to find had Linux on their radar. That in itself is a major achievement, and if the only price I pay for that service is to put some brainpower in myself when I need to build a user community site in Internet time (which could also be phrased as "ripping off slashdot in a hurry"), so be it.
It's ready when it's ready. That's my motto.
The question was: who was the best Open Source advocate. IMHO, RMS is a loose cannon who is very good at galvanizing the body of like minded persons (think rebel leader in a divided nation), but a danger to their acceptance (think foreign politics).
Perhaps an example from personal experience can paint the picture more clearly from my viewpoint. I once took a day off and drove the length of our country (not that it's that big) to hear RMS speak in Amsterdam on patents. This was in the early days of the revolt against patents on software, and I agreed with every word he said. But his mode of speach was such that I decided there and then not to associate myself with the LPF, because I did not wanted to be painted with the same brush that he would be painted with.
My non-membership of the LPF was (and is) no big deal. I publicly and vocally express the LPF view at many occasions. But at a time when we need division between equally minded people like we need holes in our heads, putting RMS in the limelight for his advocacy does not bode well for freely available software.
I wished the lawyers would stay out of it. The whole mess we got into was caused by someone slapping a copyright on code labeled Public Domain. If just a percentage of all the productivity lost in the world over license debates since would've been put into lobbying (or outright buying) Congress to change the laws to make the term Public Domain mean what every programmer understood it to mean back then, Linux would've taken over the world five years ago. Except it may have been called BSD or Minix in that case.
What's in a name?
Same for ntop. If you're remotely logging into a site behind an overloaded circuit and a few bytes of telnet are the max you can squeeze through it, you still can figure out which connection is hogging the link.
In years of systems and network monitoring, I've yet to encounter a more effective user interface.