Re:Who are you to tell me what license I should us
on
BSD vs GPL
·
· Score: 1
Exactly... depending on the application and the proprietary-ness of the code, how much of a "strategic asset" the code is, its up to you to decide what is best for you...
I put the Xerox copier/printer bit in there for that, in a sense... I can easily see some custom hardware and a driver for it (GPL'd of course) to run the copier/printer hardware (laser/scanner and photo-sensitive drum, paper feed, etc) which would of course be very hardware specific and not much use to the competition. And then the rasterizing software and such could be a "userland" app which (using LGPL'd libs) could be kept proprietary.
Probably the router example I gave could fall under this camp too.. I was trying to think of something that was low-level enough to need kernel mods that might also be proprietary "strategic asset" type code (good term for it, by the way).
The boundaries are sometimes vague... I was mainly trying to get the point across that there may be reasons a company might not want to use GPL'd code... and that, of course, as the author its always up to *me* as to how my code is licensed (unless I use other GPL'd or BSD'd code, in which case I have to follow those terms first). I may even choose to just give it away with no restrictions...
I hate seeing flame wars about GPL/BSD, Linux/BSD, because it just seems so rediculous... either or both sides jumping up and down like little kids yelling "mine, mine.. my way or the highway!" Like most adults, I am capable of making my own decisions on what is best for my code, or fits my ideals...
I personally get a good laugh sometimes out of it... I wonder how many of the Linux/BSD zealots who claim that "theirs is the best there is" have actually done any programming and have actually looked at the code, and have actually tried the other "product" before spouting their venom.
Who are you to tell me what license I should use?
on
BSD vs GPL
·
· Score: 2
I've read a lot of good arguments, and a lot of plainly pig-headed inflamatory arguments in this thread so far...
But, hey, like has been said before... it up to *me* to decide how I want *my* code to be used... so don't preach to me about how *your* license is best if it doesn't fit my idea about the future of *my* code.
I posted in a different spot, but the basic difference between BSD and GPL for company "X" that wants to use the code is as follows:
Company "X" is making a new high-performance router engine and wants to avoid re-inventing an OS or purchasing a commercial RTOS for big $$$ (and the obvious effect on the bottom line, always a big concern for the "shirts")...
Using BSD code, they take the OS and have some of their programmers spend a few hundred man-hours writing proprietary driver code and tweaking the kernel to get optimum performance for their product. They can sell the resulting product and not publish any of their proprietary code. They have a top-selling/top-performing product for a year before the competition catches up...
Using GPL code, they take the OS and have some of their programmers spend a few hundred man-hours writing proprietary driver code and tweaking the kernel to get optimum performance for their product. Again, they can sell the resulting product, but in keeping with GPL they have to publish their code (thus making it non-proprietary I guess)... Their competition can look at all the code, whip out a new product in a few months with little or *NO* software development costs (after all, company "X" paid *their* programmers to write it, but company's "Y" and "Z" get their work for free)... sell it for less money (less startup costs w/ no man-hours on programming) and steal business away from company "X". Worst case, company "X" goes out of business because they were *forced* to give away trade secrets by the GPL.
And you would wonder why a company would want to avoid GPL'd code?
Now, of course, the license you choose for your work depends on your own idealogical point of view. If you are a anti-"shirt" anti-business zealot, of course you choose GPL. If you don't like the idea of anybody using your code for profit, for whatever reason, you'd choose GPL. If you don't care that someone may profit from your code later, as long as they acknowledge your contribution, you'd probably choose BSD.
My personal choice is BSD... if its not yours, great! Personally, I write code because I like writing code, its fun and interesting and usually its something that is useful to me at the time... and if it can be useful to someone else, great! If you take my code, put some of your own novel ideas into it and sell it... hey, go for it! I'm not in it for the money, I have a job for that. If I was a good BS-artist "shirt" type I could probably sell some of the stuff myself for good money... but, then again, money isn't my prime motivator.
Maybe that isn't for you... great for you! You are entitled to whatever choice you make. I can choose to be Buddhist and you can choose to be Christian... great!! We could have some fun idealogical discussions, but don't get in my face about how your religion is "the only one for the world".
Now, I would like to add that GPL'd code doesn't prevent someone from profiting... obviously RedHat and Caldera make money providing support and documentation and their own "userland". Yes, they have to give the code away, but they can still profit. And we have some Xerox printer/copier units here at work that the service guy says have embedded Linux/PC's in them as controllers. Obviously Xerox makes money selling the units... now, are they following the GPL and publishing the code? Got me... I'm not the GPL police and really don't care to be. And if they are breaking the GPL, who exactly is going to sue them to stop it? FSF? The authors (how many are there now??)??
Hmm... if I take your code, GPL'd or BSD'd, and stip off your license statement, and compile it into my product and put a licensing statement into it that its "illegal to disassemble/reverse-engineer this product" (as I've seen many times before)... how would anybody know anyways??? And if they did find out, what an interesting legal battle that would be... "you broke my license agreement in order to discover that I broke yours". I'd love to get a transcript of that case...:-)
At any rate... you make the choice for *your* code. GPL? Fine. BSD? Fine. Write your own? Fine. Doesn't matter to me... thats what freedom of choice is all about. Lots of Free Software, lots of licensing choices, and its all up to you to choose which you use... isn't freedom great?
Re:Why not contact the Author?!
on
BSD vs GPL
·
· Score: 2
...why not contact the author and see if he will allow a LGPL library of that 15%?
It seems so damn simple to me.
Yes, but its still up to that person, and beyond that... what happens if its a GPL'd library, and each routine is GPL'd by a different person? Should I then contact 30 different people to try and get them to LGPL their code for me? And if one says no, I have to rewrite his code anyways... or if its all part of one library, does that mean that even though 29 of 30 people said they'd LGPL their code, its all still GPL'd because of *one* person who won't LGPL??
Personally, I don't care if company X takes my code and makes a profit on it... I'm not writing it to make money, I'm writing it because: a) I like programming, its interesting & fun. b) Its probably something I need or is useful to me. c) After A & B, maybe someone else can use it.
The point of the BSD license is that if company X is building a new high-performance router engine and wants to use BSD code as a base, they can. They don't have to publish/provide their proprietary algorithms to the world just because they put some new drivers and kernel code into the system (which would give their competitors all their performance secrets and hurt their business). All they have to do is acknowledge that it uses BSD code.
In contrast, under GPL they would *have* to publish their mods, which their competition could then take for themselves, build a competing unit and sell at a cut-rate cost (gee... look ma, *no* software development costs for *our* unit!) and put the company and the programmers that wrote the code (worst case) out of business!
So you wonder why a company would rather use BSD code rather than GPL?
Re:I want to try BSD .. but which one?
on
*BSD News
·
· Score: 1
>
And why can't Linux or BSD do a binary emulation of my Concurrent 3200 OS/32 programs? Come to think of it, how about IBM 370 mainframe code???
Actually, there is a big difference between emulation and the ABI layer (app. binary interface). xBSD's on an i386 platform can run most of the i386 Linux apps, since the processor can run the actual code without "emulation", but xBSD's are not going to run Sparc SunOS code on an i386 platform, anymore than your average PC running M$ slow-doze is going to run Mac programs.. or your M$DUHS machine running Vax/VMS DCL scripts and executables.
An emulator would be painfully slow... but with an ABI layer you are running at the same speed as the processor, just adding and extra translation layer inbetween the running program and the OS. When properly handled the time taken by this layer will be very negligable. Obviously this will have a greater effect on programs that do a lot of I/O (lots of system calls) versus programs that do a large amount of calcuation (work in memory will run the same under the ABI as not, perhaps better than under the original OS if the ABI machine has better memory management).
Yup... I interviewed with a place once that runs a lot of the state lotteries. Mission Critical?!? They got *fined* $1000 per *minute* of downtime, since while they were down all the Lotto machines in the entire state were down (and thats actually probably a helluva lot *more* than $1000/min in lost revenue!!).
Mission Critical... 3 machines, one watching the other and if #1 fails #2 takes over within *seconds*, and then powers up #3 (which is normally off). If the power fails, they go to battery bacukp, and after 10 minutes on battery backup the diesel generator outside powers up to run the systems.
Mission Critical... anyone want to fly in the space shuttle if they switch over the 5 on-board computers to Linux?? Not me. I'd feel safer with custom VxWorks (unix derived RTOS) software that was *simulated* and *validated* to work consistently every time, and to handle hardware failures in a graceful and controlled fashion. (I'd be scared of NT for this also, definitely not a M$ fan).
Linux may work fine for a *lot* of applications, especially as a desktop OS or for a deparmental server... but don't act like it will solve all the worlds problems. Like any OS, it has its plusses and its minuses...
Here it is again... "FreeBSD is dying".. and what, prey tell, great all-knowing source of information are you getting your "facts" from?? Other than that Linux's user base is increasing, as told in many IT magazines in many forms, where did you see some numbers that show that FreeBSD's user base is decreasing?? I very rarely see the BSD's mentioned in the mainstream press... much less any mention of number of current users of any of the BSD's.
Sorry, but it sure seems to me that there are calm and rational responses to this thread from both the Linux and *BSD people, and then there's people like *you*... from both sides, I admit... that lash out with "venomous" "name calling and childish attacks on the message bearers" as you so aptly put it.
Its time for *you* to face reality and grow up. You may drive a Ford, I drive a 4WD Chevy... if you think your Ford is God's gift to the earth, so be it.. but I like my car, and don't try to shove yours down my throat. There's room in this world for both. I'm not going to use my old '82 280ZX to haul a trailer, and my 4WD truck isn't exactly a nice "night on the town" car. Each has their uses, and there is room for both in my life.
I use Linux, NetBSD, and OpenBSD for UN*X boxes at home, plus a Mac/OS system and a Win/95 system. Each has there purpose, and each does some things better than others.
NetBSD is by far the most portable, and widely ported, although Linux is getting better. Before someone jumps down my throat over this, I should add that I have two VaxStation 3100's, two MicroVAX's, several DecStations, two 68K Mac's, a Sun Sparc 1+, and several old Sun/3's, and NetBSD supports them all, today. Yeah, I can wait for a working Linux/Vax to show up someday... I can wait for a Linux that runs on my DecStation 5000... or Linux support for my Sun 3/260... or I can run NetBSD now, full multi-user support.
Linux is more mainstream, and has more available apps for it (although a good # of them run under the BSD's compatibility mode). Like M$ Windows, more available apps does *not* necessarily mean a better OS, although I hate to make that comparison since Linux is a far better OS than M$'s. My biggest beef against Linux to date is not the speed, the underlying code, or anything to do with the software, its the "Linux zealots" that think that their "Ford" is the best there is.
I just installed OpenBSD a few weeks ago, and haven't had time to play with it much, but the codebase forked from NetBSD, so I presume its fairly close (w/ better crypto, being out of the US). My NetBSD systems have been rock-solid.
I haven't tried FreeBSD, but from the "non-zealot" comments I read on these threads, I may give it a try. The comments about better VM usage seemed knowledgeable.
To those of you who want to blow off "zealot" steam at me about how Linux is the "best" kernel, I have only one question:
How many of you have a programming background and have looked at the actual kernel code for Linux and the BSD's? (I have).
I feel obliged to comment on this topic... I have to agree w/ CJS on his topic of what constitutes a "port", versus what I would term a "porting effort". Linux/VAX certainly would not qualify as a "port" in my book, if it does not even boot to a shell, nor would Linux/68000 (referring to the page given of "Linux ports") when I see on their page the comment:
"Stay tuned while I try to build a libc.a and an/etc/init, and maybe even a/bin/sh binary"
hmm... am I to consider this a "port" when they don't even have a working init process??? Sounds somewhat like "vaporware" to me... "I have this great new 3D game, blows Quake away!!! Stay tuned while I write the 'main()' routine and a mouse input routine... and *maybe* some display code."
But, go to the Linux/PA-RISC page and you'll see they have a bootloader that can actually load a dummy kernel!! No kernel, but they *DO* have a large web site full of info about it. Maybe they should spend time writing code instead of web pages??
Now, don't get the idea that I don't like Linux.. I certainly think it is a good OS, but I don't like the way some of the "Linux ports" (if you can call them that) sell themselves. Maybe I've been jaded by too many of the vendors I've dealt with over the years selling goods they never deliver. I still have a "caching SCSI controller" (5+ years old now) that the vendor advertized for over a year as "WinNT drivers coming soon!" (NT 3.1 at the time) and they never made true on that. One of our minicomputer vendors advertised an "ODBC interface" for their proprietary database... a nice glossy flier... and when I called they were "in negotiations with a 3rd party to have them port their code" to their platform. It took two years, but they did finally release them... if you want to buy a new $10K intelligent ethernet controller and $10K worth of software (for 10 clients).
I have nothing wrong with hyping the porting effort to a new platform, but please don't group it in with all the "working" ports as if its functionally crippled.
NetBSD seems to have a good way of doing it. They seperate out their "ports" into "formal release ports" and then "experimental ports", which are basically working systems but may be buggy or don't yet support a large range of hardware, and then there are "not yet integrated porting efforts" which are works in basic development. Seems like a much more professional and controlled method of development...
I'm mainly a NetBSD user(I do have a Linux machine and an OpenBSD machine).. why?? Because I happen to have two VaxStations, two MicroVAX II's, a DECstation 5000/125, several Mac's, a Sun 3/260 and several 3/60's plus two Sparc's, and an HP475T system... and NetBSD runs on them all, fully supported in a "formal" release. I could hook any one of them up and run it as a web server, today! Now, *that* I consider having a real working "port".
One of my projects I am working on is porting NetBSD to the Apollo DNxxxx series workstations. I don't have any flashy web pages, just a little blurb in the "projects" section of the NetBSD site to let people know I'm working on it. Maybe when I get a bootloader and a kernel going (working on the bootloader now) and can get a single-user shell on one system, I'll try to get listed as a "porting effort" on the NetBSD site. I have 5 different models of Apollo's to test with (every SAU level I'm planning on supporting) and when I get it working on all of them and get a working userland toolset going, then maybe I'll become an "experimental" port...
I'm not looking for an ego boost from having my name all over web pages that I'm working on a new "port"... I'm just looking to get a reasonably up to date OS running on some old machine's I've collected. If someone else benefits, great! If noone else is still running one, at least I get a decent OS out of it.
I'm sure I could go on here, but I've got an "appointment" to go play pool and have a good time, and then maybe write some code for my bootloader tonight...
Slashdot Mirror
Exactly... depending on the application and the proprietary-ness of the code, how much of a "strategic asset" the code is, its up to you to decide what is best for you...
I put the Xerox copier/printer bit in there for that, in a sense... I can easily see some custom hardware and a driver for it (GPL'd of course) to run the copier/printer hardware (laser/scanner and photo-sensitive drum, paper feed, etc) which would of course be very hardware specific and not much use to the competition. And then the rasterizing software and such could be a "userland" app which (using LGPL'd libs) could be kept proprietary.
Probably the router example I gave could fall under this camp too.. I was trying to think of something that was low-level enough to need kernel mods that might also be proprietary "strategic asset" type code (good term for it, by the way).
The boundaries are sometimes vague... I was mainly trying to get the point across that there may be reasons a company might not want to use GPL'd code... and that, of course, as the author its always up to *me* as to how my code is licensed (unless I use other GPL'd or BSD'd code, in which case I have to follow those terms first). I may even choose to just give it away with no restrictions...
I hate seeing flame wars about GPL/BSD, Linux/BSD, because it just seems so rediculous... either or both sides jumping up and down like little kids yelling "mine, mine.. my way or the highway!" Like most adults, I am capable of making my own decisions on what is best for my code, or fits my ideals...
I personally get a good laugh sometimes out of it... I wonder how many of the Linux/BSD zealots who claim that "theirs is the best there is" have actually done any programming and have actually looked at the code, and have actually tried the other "product" before spouting their venom.
I've read a lot of good arguments, and a lot of plainly pig-headed inflamatory arguments in this thread so far...
:-)
But, hey, like has been said before... it up to *me* to decide how I want *my* code to be used... so don't preach to me about how *your* license is best if it doesn't fit my idea about the future of *my* code.
I posted in a different spot, but the basic difference between BSD and GPL for company "X" that wants to use the code is as follows:
Company "X" is making a new high-performance router engine and wants to avoid re-inventing an OS or purchasing a commercial RTOS for big $$$ (and the obvious effect on the bottom line, always a big concern for the "shirts")...
Using BSD code, they take the OS and have some of their programmers spend a few hundred man-hours writing proprietary driver code and tweaking the kernel to get optimum performance for their product. They can sell the resulting product and not publish any of their proprietary code. They have a top-selling/top-performing product for a year before the competition catches up...
Using GPL code, they take the OS and have some of their programmers spend a few hundred man-hours writing proprietary driver code and tweaking the kernel to get optimum performance for their product. Again, they can sell the resulting product, but in keeping with GPL they have to publish their code (thus making it non-proprietary I guess)... Their competition can look at all the code, whip out a new product in a few months with little or *NO* software development costs (after all, company "X" paid *their* programmers to write it, but company's "Y" and "Z" get their work for free)... sell it for less money (less startup costs w/ no man-hours on programming) and steal business away from company "X". Worst case, company "X" goes out of business because they were *forced* to give away trade secrets by the GPL.
And you would wonder why a company would want to avoid GPL'd code?
Now, of course, the license you choose for your work depends on your own idealogical point of view. If you are a anti-"shirt" anti-business zealot, of course you choose GPL. If you don't like the idea of anybody using your code for profit, for whatever reason, you'd choose GPL. If you don't care that someone may profit from your code later, as long as they acknowledge your contribution, you'd probably choose BSD.
My personal choice is BSD... if its not yours, great! Personally, I write code because I like writing code, its fun and interesting and usually its something that is useful to me at the time... and if it can be useful to someone else, great! If you take my code, put some of your own novel ideas into it and sell it... hey, go for it! I'm not in it for the money, I have a job for that. If I was a good BS-artist "shirt" type I could probably sell some of the stuff myself for good money... but, then again, money isn't my prime motivator.
Maybe that isn't for you... great for you! You are entitled to whatever choice you make. I can choose to be Buddhist and you can choose to be Christian... great!! We could have some fun idealogical discussions, but don't get in my face about how your religion is "the only one for the world".
Now, I would like to add that GPL'd code doesn't prevent someone from profiting... obviously RedHat and Caldera make money providing support and documentation and their own "userland". Yes, they have to give the code away, but they can still profit. And we have some Xerox printer/copier units here at work that the service guy says have embedded Linux/PC's in them as controllers. Obviously Xerox makes money selling the units... now, are they following the GPL and publishing the code? Got me... I'm not the GPL police and really don't care to be. And if they are breaking the GPL, who exactly is going to sue them to stop it? FSF? The authors (how many are there now??)??
Hmm... if I take your code, GPL'd or BSD'd, and stip off your license statement, and compile it into my product and put a licensing statement into it that its "illegal to disassemble/reverse-engineer this product" (as I've seen many times before)... how would anybody know anyways??? And if they did find out, what an interesting legal battle that would be... "you broke my license agreement in order to discover that I broke yours". I'd love to get a transcript of that case...
At any rate... you make the choice for *your* code. GPL? Fine. BSD? Fine. Write your own? Fine. Doesn't matter to me... thats what freedom of choice is all about. Lots of Free Software, lots of licensing choices, and its all up to you to choose which you use... isn't freedom great?
...why not contact the author and see if he will allow a LGPL library of that 15%?
It seems so damn simple to me.
Yes, but its still up to that person, and beyond that... what happens if its a GPL'd library, and each routine is GPL'd by a different person? Should I then contact 30 different people to try and get them to LGPL their code for me? And if one says no, I have to rewrite his code anyways...
or if its all part of one library, does that mean that even though 29 of 30 people said they'd LGPL their code, its all still GPL'd because of *one* person who won't LGPL??
Personally, I don't care if company X takes my code and makes a profit on it... I'm not writing it to make money, I'm writing it because:
a) I like programming, its interesting & fun.
b) Its probably something I need or is useful to me.
c) After A & B, maybe someone else can use it.
The point of the BSD license is that if company X is building a new high-performance router engine and wants to use BSD code as a base, they can. They don't have to publish/provide their proprietary algorithms to the world just because they put some new drivers and kernel code into the system (which would give their competitors all their performance secrets and hurt their business). All they have to do is acknowledge that it uses BSD code.
In contrast, under GPL they would *have* to publish their mods, which their competition could then take for themselves, build a competing unit and sell at a cut-rate cost (gee... look ma, *no* software development costs for *our* unit!) and put the company and the programmers that wrote the code (worst case) out of business!
So you wonder why a company would rather use BSD code rather than GPL?
>
And why can't Linux or BSD do a binary emulation of my Concurrent 3200 OS/32 programs? Come to think of it, how about IBM 370 mainframe code???
Actually, there is a big difference between emulation and the ABI layer (app. binary interface). xBSD's on an i386 platform can run most of the i386 Linux apps, since the processor can run the actual code without "emulation", but xBSD's are not going to run Sparc SunOS code on an i386 platform, anymore than your average PC running M$ slow-doze is going to run Mac programs.. or your M$DUHS machine running Vax/VMS DCL scripts and executables.
An emulator would be painfully slow... but with an ABI layer you are running at the same speed as the processor, just adding and extra translation layer inbetween the running program and the OS. When properly handled the time taken by this layer will be very negligable. Obviously this will have a greater effect on programs that do a lot of I/O (lots of system calls) versus programs that do a large amount of calcuation (work in memory will run the same under the ABI as not, perhaps better than under the original OS if the ABI machine has better memory management).
>
Yup... I interviewed with a place once that runs a lot of the state lotteries. Mission Critical?!? They got *fined* $1000 per *minute* of downtime, since while they were down all the Lotto machines in the entire state were down (and thats actually probably a helluva lot *more* than $1000/min in lost revenue!!).
Mission Critical... 3 machines, one watching the other and if #1 fails #2 takes over within *seconds*, and then powers up #3 (which is normally off). If the power fails, they go to battery bacukp, and after 10 minutes on battery backup the diesel generator outside powers up to run the systems.
Mission Critical... anyone want to fly in the space shuttle if they switch over the 5 on-board computers to Linux?? Not me. I'd feel safer with custom VxWorks (unix derived RTOS) software that was *simulated* and *validated* to work consistently every time, and to handle hardware failures in a graceful and controlled fashion.
(I'd be scared of NT for this also, definitely not a M$ fan).
Linux may work fine for a *lot* of applications, especially as a desktop OS or for a deparmental server... but don't act like it will solve all the worlds problems. Like any OS, it has its plusses and its minuses...
Here it is again... "FreeBSD is dying".. and what, prey tell, great all-knowing source of information are you getting your "facts" from?? Other than that Linux's user base is increasing, as told in many IT magazines in many forms, where did you see some numbers that show that FreeBSD's user base is decreasing?? I very rarely see the BSD's mentioned in the mainstream press... much less any mention of number of current users of any of the BSD's.
Sorry, but it sure seems to me that there are calm and rational responses to this thread from both the Linux and *BSD people, and then there's people like *you*... from both sides, I admit... that lash out with "venomous" "name calling and childish attacks on the message bearers" as you so aptly put it.
Its time for *you* to face reality and grow up. You may drive a Ford, I drive a 4WD Chevy... if you think your Ford is God's gift to the earth, so be it.. but I like my car, and don't try to shove yours down my throat. There's room in this world for both. I'm not going to use my old '82 280ZX to haul a trailer, and my 4WD truck isn't exactly a nice "night on the town" car. Each has their uses, and there is room for both in my life.
I use Linux, NetBSD, and OpenBSD for UN*X boxes at home, plus a Mac/OS system and a Win/95 system. Each has there purpose, and each does some things better than others.
NetBSD is by far the most portable, and widely ported, although Linux is getting better. Before someone jumps down my throat over this, I should add that I have two VaxStation 3100's, two MicroVAX's, several DecStations, two 68K Mac's, a Sun Sparc 1+, and several old Sun/3's, and NetBSD supports them all, today. Yeah, I can wait for a working Linux/Vax to show up someday... I can wait for a Linux that runs on my DecStation 5000... or Linux support for my Sun 3/260... or I can run NetBSD now, full multi-user support.
Linux is more mainstream, and has more available apps for it (although a good # of them run under the BSD's compatibility mode). Like M$ Windows, more available apps does *not* necessarily mean a better OS, although I hate to make that comparison since Linux is a far better OS than M$'s. My biggest beef against Linux to date is not the speed, the underlying code, or anything to do with the software, its the "Linux zealots" that think that their "Ford" is the best there is.
I just installed OpenBSD a few weeks ago, and haven't had time to play with it much, but the codebase forked from NetBSD, so I presume its fairly close (w/ better crypto, being out of the US). My NetBSD systems have been rock-solid.
I haven't tried FreeBSD, but from the "non-zealot" comments I read on these threads, I may give it a try. The comments about better VM usage seemed knowledgeable.
To those of you who want to blow off "zealot" steam at me about how Linux is the "best" kernel, I have only one question:
How many of you have a programming background and have looked at the actual kernel code for Linux and the BSD's? (I have).
One of these days I'll try out FreeBSD...
I feel obliged to comment on this topic...
/etc/init, and maybe even a /bin/sh binary"
I have to agree w/ CJS on his topic of what constitutes a "port", versus what I would term a "porting effort". Linux/VAX certainly would not qualify as a "port" in my book, if it does not even boot to a shell, nor would Linux/68000 (referring to the page given of "Linux ports") when I see on their page the comment:
"Stay tuned while I try to build a libc.a and an
hmm... am I to consider this a "port" when they don't even have a working init process??? Sounds somewhat like "vaporware" to me... "I have this great new 3D game, blows Quake away!!! Stay tuned while I write the 'main()' routine and a mouse input routine... and *maybe* some display code."
But, go to the Linux/PA-RISC page and you'll see they have a bootloader that can actually load a dummy kernel!! No kernel, but they *DO* have a large web site full of info about it. Maybe they should spend time writing code instead of web pages??
Now, don't get the idea that I don't like Linux.. I certainly think it is a good OS, but I don't like the way some of the "Linux ports" (if you can call them that) sell themselves. Maybe I've been jaded by too many of the vendors I've dealt with over the years selling goods they never deliver. I still have a "caching SCSI controller" (5+ years old now) that the vendor advertized for over a year as "WinNT drivers coming soon!" (NT 3.1 at the time) and they never made true on that. One of our minicomputer vendors advertised an "ODBC interface" for their proprietary database... a nice glossy flier... and when I called they were "in negotiations with a 3rd party to have them port their code" to their platform. It took two years, but they did finally release them... if you want to buy a new $10K intelligent ethernet controller and $10K worth of software (for 10 clients).
I have nothing wrong with hyping the porting effort to a new platform, but please don't group it in with all the "working" ports as if its functionally crippled.
NetBSD seems to have a good way of doing it. They seperate out their "ports" into "formal release ports" and then "experimental ports", which are basically working systems but may be buggy or don't yet support a large range of hardware, and then there are "not yet integrated porting efforts" which are works in basic development. Seems like a much more professional and controlled method of development...
I'm mainly a NetBSD user(I do have a Linux machine and an OpenBSD machine).. why?? Because I happen to have two VaxStations, two MicroVAX II's, a DECstation 5000/125, several Mac's, a Sun 3/260 and several 3/60's plus two Sparc's, and an HP475T system... and NetBSD runs on them all, fully supported in a "formal" release. I could hook any one of them up and run it as a web server, today! Now, *that* I consider having a real working "port".
One of my projects I am working on is porting NetBSD to the Apollo DNxxxx series workstations. I don't have any flashy web pages, just a little blurb in the "projects" section of the NetBSD site to let people know I'm working on it. Maybe when I get a bootloader and a kernel going (working on the bootloader now) and can get a single-user shell on one system, I'll try to get listed as a "porting effort" on the NetBSD site. I have 5 different models of Apollo's to test with (every SAU level I'm planning on supporting) and when I get it working on all of them and get a working userland toolset going, then maybe I'll become an "experimental" port...
I'm not looking for an ego boost from having my name all over web pages that I'm working on a new "port"... I'm just looking to get a reasonably up to date OS running on some old machine's I've collected. If someone else benefits, great! If noone else is still running one, at least I get a decent OS out of it.
I'm sure I could go on here, but I've got an "appointment" to go play pool and have a good time, and then maybe write some code for my bootloader tonight...
Pete