I don't know what the original poster is using as a linker, but the GNU linker won't even bother pulling in object files from archives / libraries unless they satisfy an undefined symbol dependency, even when linking statically. you can have tons of routines in your libraries, and only get the ones linked in that you directly or indirectly reference.
this only works if you have your routines split out one-per-file, otherwise the linker has to do extra work to hunt down dead routines...
Re:First Airport, now this...
on
The Guts Of An iPod
·
· Score: 2, Interesting
According to my anonymous source at PortalPlayer, it's based on Lineo's RXTC microkernel. Of course, the ``application'' side has been hacked on quite a bit and has had significant additions to the database and filesystem added, but it's still RXTC based.
Do you ever notice that most of the BSD posts are centered around "personality" issues.
Sure. <TROLL>BSD users know BSD (of whatever flavor) is technically superior to linux, therefore the only thing left to argue about is the politics!</TROLL>
All of this is due to the licensing issues that bit them in the behind WRT ipf.
Yet another shining example of Theo's stellar diplomatic skills. Theo's modus operandi in that situation was the same as well: make and act on assumptions, and if the license author tries to get clarification or disagrees with those assumptions, they are told to "fuck off".
There was concern about ipf in both the NetBSD and FreeBSD camps too, but they both talked with Darren and came to agreements, something Theo seems to be completely unable to do. His "act before discuss" behaviour is both his best and worst point!:)
Theo's the instigator and the one calling names. Dan is not completely innocent either, but Theo (as usual) could have definitely handled the licensing issue better. Something along the lines of "Hey Dan, I was looking over your licensing for qmail and djbdns, and I'm concerned that we may inadvertently be breaking it," and then working from there. Even if Dan had ended up saying "no, don't change my paths" at least there would have been a good faith effort to work together. Did Theo even bother to ask for clarification or get in contact with Dan before pulling DJB programs out of ports?
What a knee-jerk reactionist Theo is... he seems to go out of his way to piss people off and provoke confrontation. In this followup DJB clearly states:
I don't mind a BSD-style port that simply follows the installation
instructions. I have also explicitly granted permission for the distribution of precompiled packages that behave correctly. There's nothing stopping OpenBSD from distributing a qmail package.
So, as usual, Theo is blowing his stack over nothing and jumping to conclusions. He may be a good programmer, and he may be a good security expert, but he acts like a two-year-old.
I'd rather have a cluster of these little 3"x3"x3" aluminum strongarm-based boxes, made by the kanadian company Intrinsyc. you could always take handhelds and do similar things... plus they would be more portable.
It's great to see NetBSD technology catching up to the mid-1980s!
but I'll bet linux running on a sun2 would be big news...
NetBSD isn't catching up to the mid-1980s -- it is being extended to 1980s vintage machines -- there is a difference. heck, it runs on VAX 11/780s, many of which are older than some/. readers.
This is a reference platform, which is almost like open-source hardware. Motorola provides schematics, board artwork, VHDL code, bill-of-materials, everything so you can use the design in your own products.
and of course since all the hardware is open and documented, support should be good.
it comes down to licensing -- not technical issues. manufacturers to this day still have to pay the ``apple tax'' for each firewire port on their device, as opposed to USB where no such rediculous licensing is required.
MO never took off because the various disk manufacturers could never agree on common formats. ZIP was similarly doomed from the start since only one company manufactures it.
USB appears to be obsoleting serial and parallel for all practical purposes... it was showing up on PC motherboards before apple came out with the imac. it's becoming difficult to get a PC motherboard without USB. printers, scanners, mice, keyboards, even the beloved cd burners are now available in USB form, knocking out their bus-predecessors.
I personally think compactflash prices and densities will eventually improve to the point that they will replace the floppy. it just needs to get cheaper.
apparently nobody has been reading the small print at the bottom of the digittrax website:
Yes, in the strictest sense of the word LocoNet is a proprietary system. In order to maintain a system as complex as LocoNet someone has to be "in charge" so that valuable system resources are not used unwisely. Digitrax has
decided to maintain LocoNet professionally for the benefit of the hobby & to work with other competent DCC developers so that they can include it in their systems. The non-disclosure agreements & licensing agreements & fees for the use of LocoNet are not intended to discourage anyone from using the system but merely to insure that system resources are used prudently & to cover our cost to maintain the system.
sounds like a project that needs open-sourcing to me. while the protocols for DCC aren't freely copyable, at least there don't appear to be any non-disclosure agreements necessary to develop or use it.
I started using NetBSD in 1996 after a surplus decstation 5000/240 was given to me by a friend. since then I've been won over by NetBSD's emphasis on portability and full support of platforms linux is still struggling on. (alpha 3000/xxx, decstation, and vax.)
the NetBSD Goals page really lays out the reasons I like NetBSD.
NeXTstep had quad fat binaries for years... a single binary runs on 68k, sparc, hppa, and x86. all the APIs are consistent across platforms, so all adobe or whoever has to do is port once to cocoa (or whatever the hell the NeXTstep stuff is being called now) and make sure the proper compiler is installed.
the extra code in the binaries could be stripped upon installation...
between ECOS, GNU HURD, RTEMS and others, there is a plethora of open-source hard real-time and microkernel systems available to the open-source hacker. closed source systems may be slick, but you will always be at the mercy of the closed-source owners.
NetBSD can upgrade installed ports. It's called 'make update'. it updates the current pkg and all the packages that depend on it. the only linux packager I know that does this is debian's apt (or whatever it's called.) I'm sure {Free,Open}BSD have this feature too...
courageous folks are working hard to port NetBSD to VAX architecture.
NetBSD already runs on VAXen. It has for quite a while, and gets better all the time. Just a few weeks ago, VAX ELF binaries were demonstrated, so it probably won't be too long before GNU/Linux runs on them too.
I think you will find more embedded systems devoplers going to X because is allow them to remotly display debugging back to the desktop.
microwindows has a client / server network transparent architecture, similar to X. it's a heck of a lot smaller than X, too.
there is also RTEMS, which is a real-time multiprocessor-supporting OS, if you ever want to ditch vxworks... (there is also ecos which looks interesting. but this is getting off topic.)
Developing/studying systems that can be proved secure (buffer overflow wrapper where?)
you could have runtime protection automatically inserted by the compiler, like stackguard but it'd probably be better in the long term to use languages that have strict bounds and type checking. (modula3?)
Beta Video tapes, DAT Audio, Digital AM/FM Radio, all have been "the best" and all have simply died.
beta survives as ``betamax'' and is used heavily in the broadcasting industry, and DATs are still used by tapers and for mastering in the pro-audio world. the algorithms developed for digital radio survive in MPEG2 layer III...
and btw... walnut creek and BSDi merged back in March, so FreeBSD is effectively backed by BSDi now.
Perhaps the best thing that WinCE^WPocket PC will do is drive the cost down for high-powered portable hardware. Who says you need to run a Microsoft OS on a Pocket PC anyway? NetBSD/hpcmips already runs on most of the Pocket PCs...
Microsoft is spurring the market for smaller, faster, cheaper hardware. That's fine with me.
Slashdot Mirror
I don't know what the original poster is using as a linker, but the GNU linker won't even bother pulling in object files from archives / libraries unless they satisfy an undefined symbol dependency, even when linking statically. you can have tons of routines in your libraries, and only get the ones linked in that you directly or indirectly reference.
this only works if you have your routines split out one-per-file, otherwise the linker has to do extra work to hunt down dead routines...
According to my anonymous source at PortalPlayer, it's based on Lineo's RXTC microkernel. Of course, the ``application'' side has been hacked on quite a bit and has had significant additions to the database and filesystem added, but it's still RXTC based.
Sure. <TROLL>BSD users know BSD (of whatever flavor) is technically superior to linux, therefore the only thing left to argue about is the politics!</TROLL>
Yet another shining example of Theo's stellar diplomatic skills. Theo's modus operandi in that situation was the same as well: make and act on assumptions, and if the license author tries to get clarification or disagrees with those assumptions, they are told to "fuck off".
There was concern about ipf in both the NetBSD and FreeBSD camps too, but they both talked with Darren and came to agreements, something Theo seems to be completely unable to do. His "act before discuss" behaviour is both his best and worst point! :)
Theo's the instigator and the one calling names. Dan is not completely innocent either, but Theo (as usual) could have definitely handled the licensing issue better. Something along the lines of "Hey Dan, I was looking over your licensing for qmail and djbdns, and I'm concerned that we may inadvertently be breaking it," and then working from there. Even if Dan had ended up saying "no, don't change my paths" at least there would have been a good faith effort to work together. Did Theo even bother to ask for clarification or get in contact with Dan before pulling DJB programs out of ports?
What a knee-jerk reactionist Theo is... he seems to go out of his way to piss people off and provoke confrontation. In this followup DJB clearly states:
So, as usual, Theo is blowing his stack over nothing and jumping to conclusions. He may be a good programmer, and he may be a good security expert, but he acts like a two-year-old.
not exactly. if you've read any of William Gibson's Neuromancer books, and saw my basement, it would make more sense to you...
Try Richard Crandall's Projects in Scientific Computation perhaps? It covers all of these and more in one book.
I'd rather have a cluster of these little 3"x3"x3" aluminum strongarm-based boxes, made by the kanadian company Intrinsyc. you could always take handhelds and do similar things... plus they would be more portable.
but I'll bet linux running on a sun2 would be big news...
NetBSD isn't catching up to the mid-1980s -- it is being extended to 1980s vintage machines -- there is a difference. heck, it runs on VAX 11/780s, many of which are older than some /. readers.
This is a reference platform, which is almost like open-source hardware. Motorola provides schematics, board artwork, VHDL code, bill-of-materials, everything so you can use the design in your own products.
and of course since all the hardware is open and documented, support should be good.
it comes down to licensing -- not technical issues. manufacturers to this day still have to pay the ``apple tax'' for each firewire port on their device, as opposed to USB where no such rediculous licensing is required.
MO never took off because the various disk manufacturers could never agree on common formats. ZIP was similarly doomed from the start since only one company manufactures it.
USB appears to be obsoleting serial and parallel for all practical purposes... it was showing up on PC motherboards before apple came out with the imac. it's becoming difficult to get a PC motherboard without USB. printers, scanners, mice, keyboards, even the beloved cd burners are now available in USB form, knocking out their bus-predecessors.
I personally think compactflash prices and densities will eventually improve to the point that they will replace the floppy. it just needs to get cheaper.
apparently nobody has been reading the small print at the bottom of the digittrax website:
sounds like a project that needs open-sourcing to me. while the protocols for DCC aren't freely copyable, at least there don't appear to be any non-disclosure agreements necessary to develop or use it.
NetBSD has already forked. Where do you think OpenBSD came from?
I started using NetBSD in 1996 after a surplus decstation 5000/240 was given to me by a friend. since then I've been won over by NetBSD's emphasis on portability and full support of platforms linux is still struggling on. (alpha 3000/xxx, decstation, and vax.)
the NetBSD Goals page really lays out the reasons I like NetBSD.
NeXTstep had quad fat binaries for years... a single binary runs on 68k, sparc, hppa, and x86. all the APIs are consistent across platforms, so all adobe or whoever has to do is port once to cocoa (or whatever the hell the NeXTstep stuff is being called now) and make sure the proper compiler is installed. the extra code in the binaries could be stripped upon installation...
NetBSD can upgrade installed ports. It's called 'make update'. it updates the current pkg and all the packages that depend on it. the only linux packager I know that does this is debian's apt (or whatever it's called.) I'm sure {Free,Open}BSD have this feature too...
NetBSD already runs on VAXen. It has for quite a while, and gets better all the time. Just a few weeks ago, VAX ELF binaries were demonstrated, so it probably won't be too long before GNU/Linux runs on them too.
microwindows has a client / server network transparent architecture, similar to X. it's a heck of a lot smaller than X, too.
there is also RTEMS, which is a real-time multiprocessor-supporting OS, if you ever want to ditch vxworks... (there is also ecos which looks interesting. but this is getting off topic.)
you could have runtime protection automatically inserted by the compiler, like stackguard but it'd probably be better in the long term to use languages that have strict bounds and type checking. (modula3?)
beta survives as ``betamax'' and is used heavily in the broadcasting industry, and DATs are still used by tapers and for mastering in the pro-audio world. the algorithms developed for digital radio survive in MPEG2 layer III...
and btw... walnut creek and BSDi merged back in March, so FreeBSD is effectively backed by BSDi now.
usenet has had distributed decentralized-server forums since before TCP/IP was widespread... the more things change, the more people seem to forget.
Perhaps the best thing that WinCE^WPocket PC will do is drive the cost down for high-powered portable hardware. Who says you need to run a Microsoft OS on a Pocket PC anyway? NetBSD/hpcmips already runs on most of the Pocket PCs...
Microsoft is spurring the market for smaller, faster, cheaper hardware. That's fine with me.
they have USB... and of course NetBSD supports USB ethernet. So you could have that diskless X terminal after all.