She didn't have the fish water running through the CPU cooler. I was quite disappointed. She was concerned about detritus in the fish water clogging the water cooling, but a canister filter could take care of that and also serve as the pump.
You'd need a pretty big tank to use up a PCs heat load. Tank heaters might be specified at 4W/gallon but they are on thermostats & don't run continously in heated houses. A radiator and temperature controlled fan after the CPU to keep the water back to the tank from being over warm would fix that. You need to make the tank small enough that the PC at idle at night in winter (assuming set-back HVAC thermostat) will keep it warm enough. You could run folding so you're never idle....
The larger the tank, the longer it will take to see if one component is over or under sized. I just replace the mechanical thermostat on my 400 gallon hot tub with an electronic one & it takes *forever* to calibrate the thermistor probe.
But does the old Sequent contract have a similar exception for {Sequent authored features} in {Unix derived Dynix} not being controllable by {Unix's owner}?
(note that I said controllable, not owned. Derivative works are messy, eh?)
Maybe IBM's JFS is free & clear, but what about [originally Sequent] authored NUMA?
You: Ok, don't you want to be able to listen to your music collection all on five of the computers networked throughout your house? Caller: Uh, I only have two computers.
You: why two? Caller: well, I have an old Gateway with Win 98. When I bought my new XP Home Dell computer, I realized that I didn't have the CDs from some software anymore and didn't have the right keys for some other stuff that I'd bought as a download, so I keep the old computer around to run those programs.
"normal folks" are already exposed to DRM hassles, but on a much coarser basis: fewer DRM'd items & infrequent dealing with the problems that DRM causes. When they see the mess it'll cause with music, they'll recognize the beast.
What happens when your hard drive crashes the same week someone steals your identity? Good luck getting all your online-purchased music back without a long & painful customer service horror story.
If the world would keep scheduling things on the same day, mmm that would be a big help yesss. I ran over my Palm with my wife's SUV. Even a bumper won't protect it from that.
This means that according to their claim, Linux in fact does not have to be the same as System V, it only has to be derived from it for them to have ownership of (part of) the Linux source code.
While SCO have said many things that are wrong, one thing they got right was how derivative copyrights work.
They don't claim to own Linux, but only that they can exert control over it. This means that they can demand $699 from every user, but Linus could demand $6999 from every user (boycott SCO by taking Linux "off the market" for a week) and SCO couldn't do anything about it.
SCO is claiming not that they just own a portion of Linux, but that the entire "work" is a derviative work of their IP. They've stated that removing the infringing pieces once identified won't be possible, because their design is pervasive.
If entities collaborate, their work is controlled by both. If they can't agree on terms to distribute, neither party can distribute independently.
I suspect this is why it's so hard to get classic videogame ROMs released. Their ownership might be tied up with claims from Atari/Namco/Williams some of whom might have been through reorganization.
For an example of software resurrection gone very right, check out (Free) Blender.
if your program falls on its face without my library, then it's a derived work
That's not a very useful definition to me. While you might use it only for your program, it doesn't seem generally applicable. What application can run without an OS under it? What OS can run without driver to help it talk to hardware? What GUI app can run without widget/drawing libraries?
It is not the end. You cannot just replace the code retroactively. There has been a lot of people who have made money off of that code. They would need some sort of reperations. That would get messy and would hurt Linux's adoption rate.
That's my worry. What I'm afraid will happen is that somewhere in the mess of Project Monterey and Sequent/IBM's Unix licenses (and the Dynix transition from BSD- to Unix-based) there will be a shred of I.P. that's judged to be a misappropriated, how can SCO be compensated for the "damage?"
A judge facing this quandry might give in to SCO's request for money from every Linux user.
they're willing to take a much lower profit maragin to build market share: not a bad plan if you expect mini-iPod buyers to graduate to higher maragin products in a year or so
Have they ever done that? Apple tries to add value to justify higher prices rather than remove value & sell cheap.
My prediction is that the mini iPod will be flash based for 3 1/2 reasons:
1) cost 2) vibration/shock resistance 3) size 4) and maybe power consumption (dunno 'bout the CSE's mW)
I'm guessing it'll be $129 not $99, but I think the suggestion made elsewhere that this might include some ITMS songs is right-on. They eventually put out a $99 model with the flash of the $129 & do a new $129 with double the old one.
You can jog with an iPod if you hold it in your hand, but you're pushing its ability to keep the buffer from emptying. Snowboarding, mountain biking, &c will be harder. I think they'll make it really small and emphasize the "active" aspect.
How is that any better than everyone else's flash based player? (that is Cornice would be a bigger differentiator) Good question, mmmm... 1) rides the coattails of its big brother 2) is incredibly small 3) iTunes & the store 4) some wildcard like Bluetooth.
1) travels around with me and is more likely to get dropped than my PC or CDs
2) has already crashed once (3 months old) requiring a "force disc mode." That's documented at Apple, not in the manual & I actually found out about it in the iPodlounge FAQ. This recovery mode requires a "restore" which reformats the disc.
My music is safely on a bunch of discs sitting in a changer. Now that I have an iPod, the only reason discs leave the changer is to be ripped (I'm only 1/2 done).
I don't want to lose the work I did ripping & encoding, so I'll burn 4 or 5 DVDs at some point.
The only way an iPod could save my music is if my house burned down while I was out and it didn't crash before I had a chance to back it up. I plan to store my DVD-R backupss at work for a better version of the same idea.
The trickiest part of GPL-like OSS licenses to me is figuring out where the line on "derivative works" ends.
This is really a problem of copyright law in general and companies that collaborate (source only open to the parties involved) can get into trouble deciding who owns the resultant work, but most closed-source licenses are pretty simple: I can distribute object form only of the Metaware libraries, I can do anything with the input or output of their compiler (my code) and my work which uses their libraries is not considered a derived work (even though they're statically linked).
Linus has clarified where he stands on userland use of the kernel, but lately has made some odd statements about kernel modules distributed in binary form. There's also issues with libraries distributed as GPL instead of LGPL (e.g. the MAD mp3 decoder). If MAD is dynamically linked to my app, but both are contained in a ROM and the "dynamic" loader uses the simplified BFLT format, is my app still independent & not derived (so I can release source to MAD but not my app)? Does statically linking my app mean I must release its source???
Don't forget, Unix System Laboratories vs. Berkeley Software Design Inc. was settled.
The only terms that weren't confidential are:
BSDI has agreed to substitute a port of the University of California's forthcoming new release to be known as 4.4 BSD(Lite) for BSD/386. For a limited period of time, BSDI may continue to distribute its BSD/386 product, although certain portions of the code may be distributed in binary form.
Most unsatisfying. I'm hoping for all complaints to reach judgement, but I expect some to be settled with a sealed agreement, leaving things most unsettled.
Much more customizable audio output support for one, can push out 24bit/96Khz if your card supports it, better dithering at 16 bits, etc. etc.
There's a MAD plugin for Winamp that decodes to 24 bits. The waveOut output plugin is supposed to carry those bits to the soundcard without rounding down to 16 bits.
Foobar can do gapless mp3 playback which according to hydrogenaudio discussions I've followed, looks to be pretty hard. The short gaps between songs that are gapless on CD when my iPod plays 'em back is quite disturbing. For instance, the Meridian Art Ensemble's Prime Meridian has Peaches En Regalia flow right into Let's Make The Water Turn Black and it's gorgeous, but the segue is wrecked on the iPod.
Foobar 2000 pushing the envelope doesn't let Apple get away with "it's too hard."
freedb.org haven't added a general purpose search to the cddbd server. It can only be searched by discid.
The full-text search is still labelled experimental in CVS, though the page at freedb.org doesn't warn about that anymore.
I've chosen to download the database to a local server and tweak the server code (in C-- I'm nowhere with Perl) to allow full-text searches. I ripped about 30 CDs with EAC before I knew how to set the tagging options right so I'm missing track numbers.
The term for what you're describing is misappropriation. A misappropriated trade secret is still considered secret and anyone that uses it is committing a crime. At least that was the theory before cases like DeCSS and Verance/SDMI/Felten where the trade secret holders are desperately trying to apply the old rules to the international internet.
The key cases on the misappropriation doctrine seems to be E. I. DuPont de Nemours Powder Co. v. Masland where a competitor took aerial photos of a chemical plant & claimed it was reverse engineering. The reasonableness of a measure to discover inner workings was at issue.
It's still a wierd area. Printed circuit board layouts can be copyrighted but you can look at one, extract a schematic and lay out your own board. The circuit design could receive patent protection, but not trade secret protection.
SCO (or Novell) owns copyrights on Unix. IBM created a derived work called AIX, which they couldn't distribute without SCO's permission. That permission is what the contract that is now in dispute grants.
SCO representatives say a lot of things that are just plain nuts, but one thing Darl said that was true and helpfully clarifying is that SCO doesn't own AIX, but they get to exert some control over it. Not releasing source containing SCO (or Novell's) trade secrets is part of that control.
Now they're trying to extend that control to Linux claiming that IBM's contributions are also somehow derived works of Unix. It's a strange claim to me because if you look at the history of NUMA for example, it started with Sequent as a feature on BSD *ix and only migrated to Unix (R) during the AT&T v. Berkeley mess when it seems everyone took out a Sytem V license in case things went against them (SunOS moved from BSD to SysV-based Solaris around the same time).
It's clear to me that NUMA was an independent feature that might depend on *ix-ish features, but not Unix ones. Same thing with JFS on OS/2 before AIX.
Even pretty tightly integrated things like Linux on top of Mach (MkLinux) vs. BSD on top of Mach (Darwin) seem to be pretty independent of a particular flavor of *ix. Those operating systems concepts that originated at Bell Labs are part of the general know-how of Computer Scientists now, thanks in large part to all the Prentice-Hall and Addison-Wesley books by the Bell guys and the spreading around of what's now called Ancient Unix source code. I feel that no one owns it anymore than anyone owns the z plane. Maybe we could use attribution to make inventors feel better. As we have Einstein's Theories of Relativity and a Riemann Sphere, how about the Thompson/Ritchie/McBride OS (Darl doesn't deserve credit, but if it'll make him drop the suits, let's give it to him).
2) Stop selling the product until the GPLed code was replaced with proprietary code, and re-release it,
Under some definitions of derived work, it might not be possible or might be very hard to write new modules to replace the GPL-licensed imported ones without the new code being a derived work of that GPL'd code.
For instance, Linus claims that if you include the kernel headers, you're creating a derived work.
I'll accept that, but if I use methods I learned from GPL'd source code to write my own, I should be in the clear. It's the expression of an idea, not the idea itself, that can be protected.
However, GPL'd code authors aren't like corporate coders who have a legal staff to handle IP infractions. Joe invents something very cool and tricky that's widely admired. I adapt his method, but not his implementation and use it in a product which makes my company millions. Joe and his admirers might be very upset with me and while Joe might not be able to sue, we'll be exposed to a big P.R. mess.
That said, anyone have an opinion on sneaking a look at/usr/src/linux/fs/binfmt_flat.c while writing a loader that swallows the output of elf2flt? The modified source of the latter would be released to OEMs writing dynamically loaded modules, but the loader would be in a closed-source ROM.
I think the best approach is to look at binfmt_flat.c while writing a spec of how to load a flat binary, then put away the source while I write my own implementation or even better, hand the spec to another coder. It's not as simple, though, as some suggest to comply with the edge cases of the GPL and the penalty is not as simple as some claim.
You ought to be able to hire a lawyer for one hour (e.g. $200) to get your facts straight, and then not have to pay another dollar (unless your case really needs it) until after the judge/jury has given the verdict.
I did this once. The lawyer charged by the 10th of an hour (6 minutes) at $125/hour or something. I met with him in his office for two hours, then called him for another 10 minutes to clarify something.
I had about 30 pages of stock option contracts I wanted an opinion on. It would have been a fortune to have him read the whole thing, so I pointed out the sections that confused me & he clarified the contract law issues.
The main legal doctrine I learned about was that when a point in a contract is ambiguous, courts are supposed to favor the interpretation of the party who didn't write the contract.
I didn't have to put him on retainer, but if things had not turned out my way (the company hadn't bought back my stock when I quit), he would have been available for more questions & either filing suit or referring to someone who could have.
That referral mechanism is the key-- I called this particular lawyer from a listing (not an ad) in the yellow pages. Some of the lawyers I called never returned my messages. It's an insular club that if you don't have a brother who has a friend who plays raquetball with a lawyer, can be hard to penetrate (unless it's injury claims).
By the way, I think the rumor that I'd consulted a lawyer helped to make management take the whiners more seriously. In retrospect, I wish we'd buckled down & made the best damn code around, but the "lottery factor" (possibility that you might get a whole lot of something for nearly nothing) clouded our judgement.
I started with a K-1000, but when it was $130 from K-mart in 1984. They aren't made anymore and are more expensive than warranted due to (overblown) reputation.
Yes, they're tough (mine still worked fine 4 years ago with no CLA (clean, lubricate and adjust) when I traded up to a Super Program), but they're lacking:
1) Crappy meter. Slow to react and wierdly non-linear at low light levels, so not good for existing light photography with an f/1.4 50mm lens & 400ASA film.
2) No depth of field (a.k.a. depth of focus) preview. This is a hard feature to learn how to use, but control of DOF is a big part of learning photography and one area where 35mm kicks the crap out of point-and-shoot digicams (which have small sensors, short focal lengths and deep DOF so hard to knock the background out of focus for portraits).
3) Slow flash sync (X) speed. 1/60th, right? Once you learn manual existing-light photography, you might want to try manual (guide number/focus distance) flash photography. For fill-flash (lighting up a face shaded by a hat brim or eyes shaded by brow), faster sync gives you flexibility.
[I actually don't recommend trying to learn to use bounce & other tricks to make flash look more natural on anything but digital unless you have a darkroom. Too much lag between exposure & result to figure out what you're doing]
4) rubberized-cloth fully mechanical shutter. This means the battery only powers the meter & the camera will work with no battery at all. However, it isn't as accurate as quartz-controlled metal blade shutters like in the SuperProgram.
That said, the Pentax line is nice because the lenses work on the new bodies (including their digital *ist), though sometimes metering doesn't work. Nikon is the only other mfg. that kept the mount the same when they went autofocus-- Canon & Minolta changed. Minolta still makes their manual focus cameras, though. Canon manuals are orphaned with parts getting harder to find.
---
Make sure you get a "fast" lens. 85mm or 100mm f/2 or 50mm f/1.4. It's damned hard to focus an f/2 50mm lens (which came on my K1000 originally) because the DOF wide-open is too deep to give you a "snappy" focus.
You could use these to take drugs from one place to another.
Makes a lot more sense than search-and-rescue. For that application, I'd want a Moto-Guzzi powered super-slow UAV.
Drugs are a great civilian app for cruise missles, though: low altitude and high speed help evade detection and value of goods makes up for high cost of jet fuel and low payload capacity.
Maybe a more acceptable app would be delivering *legal* drugs (antibiotics, HIV triple-cocktail) to regions with poor transport infrastructure, esp. with ongoing civil wars (I've been watching too much e.r.).
Slashdot Mirror
Sun is again a boring name... i'm sure other Suns exist other than the sun we know and love.
m l
Sunn makes (made?) great bass amps. Owned by Fender now (what isn't owned by Gibson or Fender these days?)
Sunn Microsystems? Are they competing with Pignose now?
www.sunnbicycle.com
www.sunnbattery.com
Newspapers
www.thesun.co.uk
www.sunspot.net
Son House
www.slidingdelta.com/bluesmen/sonhouse2.ht
I think a picture of him is going on our server room door.
Sun studios (Elvis)
www.sunstudio.com
The original Sun
sohowww.nascom.nasa.gov
Where's the item in the moderate listbox for "insightful flamebait"?
She didn't have the fish water running through the CPU cooler. I was quite disappointed. She was concerned about detritus in the fish water clogging the water cooling, but a canister filter could take care of that and also serve as the pump.
You'd need a pretty big tank to use up a PCs heat load. Tank heaters might be specified at 4W/gallon but they are on thermostats & don't run continously in heated houses. A radiator and temperature controlled fan after the CPU to keep the water back to the tank from being over warm would fix that. You need to make the tank small enough that the PC at idle at night in winter (assuming set-back HVAC thermostat) will keep it warm enough. You could run folding so you're never idle....
The larger the tank, the longer it will take to see if one component is over or under sized. I just replace the mechanical thermostat on my 400 gallon hot tub with an electronic one & it takes *forever* to calibrate the thermistor probe.
(sorry, should have previewed)
But does the old Sequent contract have a similar exception for {Sequent authored features} in {Unix derived Dynix} not being controllable by {Unix's owner}?
(note that I said controllable, not owned. Derivative works are messy, eh?)
Maybe IBM's JFS is free & clear, but what about [originally Sequent] authored NUMA?
But does the old Sequent contract have a similar exception for to not being controllable by ?
Maybe IBM's JFS (which others have pointed out was developed OS-independent and deployed 1st on OS/2) is free & clear, but what about NUMA?
You: Ok, don't you want to be able to listen to your music collection all on five of the computers networked throughout your house?
Caller: Uh, I only have two computers.
You: why two?
Caller: well, I have an old Gateway with Win 98. When I bought my new XP Home Dell computer, I realized that I didn't have the CDs from some software anymore and didn't have the right keys for some other stuff that I'd bought as a download, so I keep the old computer around to run those programs.
"normal folks" are already exposed to DRM hassles, but on a much coarser basis: fewer DRM'd items & infrequent dealing with the problems that DRM causes. When they see the mess it'll cause with music, they'll recognize the beast.
What happens when your hard drive crashes the same week someone steals your identity? Good luck getting all your online-purchased music back without a long & painful customer service horror story.
January 23rd is also the beginning of World Rally Championship season with the Rallye Automobile Monte Carlo
If the world would keep scheduling things on the same day, mmm that would be a big help yesss. I ran over my Palm with my wife's SUV. Even a bumper won't protect it from that.
This means that according to their claim, Linux in fact does not have to be the same as System V, it only has to be derived from it for them to have ownership of (part of) the Linux source code.
While SCO have said many things that are wrong, one thing they got right was how derivative copyrights work.
They don't claim to own Linux, but only that they can exert control over it. This means that they can demand $699 from every user, but Linus could demand $6999 from every user (boycott SCO by taking Linux "off the market" for a week) and SCO couldn't do anything about it.
SCO is claiming not that they just own a portion of Linux, but that the entire "work" is a derviative work of their IP. They've stated that removing the infringing pieces once identified won't be possible, because their design is pervasive.
If entities collaborate, their work is controlled by both. If they can't agree on terms to distribute, neither party can distribute independently.
I suspect this is why it's so hard to get classic videogame ROMs released. Their ownership might be tied up with claims from Atari/Namco/Williams some of whom might have been through reorganization.
For an example of software resurrection gone very right, check out (Free) Blender.
NiCad batteries do indeed last longer when allowed to mostly discharge then be recharged.
They also might last longer when put through fewer cycles on a charger that overcharges them.
If designers cared for NiCads as carefully as they have to with Li-Ion, the NiCads might last a little longer.
Linux is used for oil exploration research.
Gates and McBride are in a different circle of hell than Cheney and Perle.
if your program falls on its face without my library, then it's a derived work
That's not a very useful definition to me. While you might use it only for your program, it doesn't seem generally applicable. What application can run without an OS under it? What OS can run without driver to help it talk to hardware? What GUI app can run without widget/drawing libraries?
It is not the end. You cannot just replace the code retroactively. There has been a lot of people who have made money off of that code. They would need some sort of reperations. That would get messy and would hurt Linux's adoption rate.
That's my worry. What I'm afraid will happen is that somewhere in the mess of Project Monterey and Sequent/IBM's Unix licenses (and the Dynix transition from BSD- to Unix-based) there will be a shred of I.P. that's judged to be a misappropriated, how can SCO be compensated for the "damage?"
A judge facing this quandry might give in to SCO's request for money from every Linux user.
they're willing to take a much lower profit maragin to build market share: not a bad plan if you expect mini-iPod buyers to graduate to higher maragin products in a year or so
Have they ever done that? Apple tries to add value to justify higher prices rather than remove value & sell cheap.
My prediction is that the mini iPod will be flash based for 3 1/2 reasons:
1) cost
2) vibration/shock resistance
3) size
4) and maybe power consumption (dunno 'bout the CSE's mW)
I'm guessing it'll be $129 not $99, but I think the suggestion made elsewhere that this might include some ITMS songs is right-on. They eventually put out a $99 model with the flash of the $129 & do a new $129 with double the old one.
You can jog with an iPod if you hold it in your hand, but you're pushing its ability to keep the buffer from emptying. Snowboarding, mountain biking, &c will be harder. I think they'll make it really small and emphasize the "active" aspect.
How is that any better than everyone else's flash based player? (that is Cornice would be a bigger differentiator) Good question, mmmm... 1) rides the coattails of its big brother 2) is incredibly small 3) iTunes & the store 4) some wildcard like Bluetooth.
Safeguard $3k worth of music
Uh, no.
My iPod
1) travels around with me and is more likely to get dropped than my PC or CDs
2) has already crashed once (3 months old) requiring a "force disc mode." That's documented at Apple, not in the manual & I actually found out about it in the iPodlounge FAQ. This recovery mode requires a "restore" which reformats the disc.
My music is safely on a bunch of discs sitting in a changer. Now that I have an iPod, the only reason discs leave the changer is to be ripped (I'm only 1/2 done).
I don't want to lose the work I did ripping & encoding, so I'll burn 4 or 5 DVDs at some point.
The only way an iPod could save my music is if my house burned down while I was out and it didn't crash before I had a chance to back it up. I plan to store my DVD-R backupss at work for a better version of the same idea.
The trickiest part of GPL-like OSS licenses to me is figuring out where the line on "derivative works" ends.
This is really a problem of copyright law in general and companies that collaborate (source only open to the parties involved) can get into trouble deciding who owns the resultant work, but most closed-source licenses are pretty simple: I can distribute object form only of the Metaware libraries, I can do anything with the input or output of their compiler (my code) and my work which uses their libraries is not considered a derived work (even though they're statically linked).
Linus has clarified where he stands on userland use of the kernel, but lately has made some odd statements about kernel modules distributed in binary form. There's also issues with libraries distributed as GPL instead of LGPL (e.g. the MAD mp3 decoder). If MAD is dynamically linked to my app, but both are contained in a ROM and the "dynamic" loader uses the simplified BFLT format, is my app still independent & not derived (so I can release source to MAD but not my app)? Does statically linking my app mean I must release its source???
The only terms that weren't confidential are:
Most unsatisfying. I'm hoping for all complaints to reach judgement, but I expect some to be settled with a sealed agreement, leaving things most unsettled.
2.4.x was better because it thinned out use of the BKL
Yeah, OK. I've only been following Kernel Traffic for a while, so I didn't recognize this either.
googling.... sourceforging....
BKL = Big Kernel Lock. Here's a nice paper on global locking The Linux Scalability Effort seems to have coordinated some of the effort to remove it and also introduce other features which help SMP. Not too surprising that a lot of contributions come from IBM & SGI folk.
Much more customizable audio output support for one, can push out 24bit/96Khz if your card supports it, better dithering at 16 bits, etc. etc.
There's a MAD plugin for Winamp that decodes to 24 bits. The waveOut output plugin is supposed to carry those bits to the soundcard without rounding down to 16 bits.
Foobar can do gapless mp3 playback which according to hydrogenaudio discussions I've followed, looks to be pretty hard. The short gaps between songs that are gapless on CD when my iPod plays 'em back is quite disturbing. For instance, the Meridian Art Ensemble's Prime Meridian has Peaches En Regalia flow right into Let's Make The Water Turn Black and it's gorgeous, but the segue is wrecked on the iPod.
Foobar 2000 pushing the envelope doesn't let Apple get away with "it's too hard."
freedb.org haven't added a general purpose search to the cddbd server. It can only be searched by discid.
The full-text search is still labelled experimental in CVS, though the page at freedb.org doesn't warn about that anymore.
I've chosen to download the database to a local server and tweak the server code (in C-- I'm nowhere with Perl) to allow full-text searches. I ripped about 30 CDs with EAC before I knew how to set the tagging options right so I'm missing track numbers.
The term for what you're describing is misappropriation. A misappropriated trade secret is still considered secret and anyone that uses it is committing a crime. At least that was the theory before cases like DeCSS and Verance/SDMI/Felten where the trade secret holders are desperately trying to apply the old rules to the international internet.
The key cases on the misappropriation doctrine seems to be E. I. DuPont de Nemours Powder Co. v. Masland where a competitor took aerial photos of a chemical plant & claimed it was reverse engineering. The reasonableness of a measure to discover inner workings was at issue.
It's still a wierd area. Printed circuit board layouts can be copyrighted but you can look at one, extract a schematic and lay out your own board. The circuit design could receive patent protection, but not trade secret protection.
ObLink
SCO [...] do not have a copyright over IBM's work
SCO (or Novell) owns copyrights on Unix. IBM created a derived work called AIX, which they couldn't distribute without SCO's permission. That permission is what the contract that is now in dispute grants.
SCO representatives say a lot of things that are just plain nuts, but one thing Darl said that was true and helpfully clarifying is that SCO doesn't own AIX, but they get to exert some control over it. Not releasing source containing SCO (or Novell's) trade secrets is part of that control.
Now they're trying to extend that control to Linux claiming that IBM's contributions are also somehow derived works of Unix. It's a strange claim to me because if you look at the history of NUMA for example, it started with Sequent as a feature on BSD *ix and only migrated to Unix (R) during the AT&T v. Berkeley mess when it seems everyone took out a Sytem V license in case things went against them (SunOS moved from BSD to SysV-based Solaris around the same time).
It's clear to me that NUMA was an independent feature that might depend on *ix-ish features, but not Unix ones. Same thing with JFS on OS/2 before AIX.
Even pretty tightly integrated things like Linux on top of Mach (MkLinux) vs. BSD on top of Mach (Darwin) seem to be pretty independent of a particular flavor of *ix. Those operating systems concepts that originated at Bell Labs are part of the general know-how of Computer Scientists now, thanks in large part to all the Prentice-Hall and Addison-Wesley books by the Bell guys and the spreading around of what's now called Ancient Unix source code. I feel that no one owns it anymore than anyone owns the z plane. Maybe we could use attribution to make inventors feel better. As we have Einstein's Theories of Relativity and a Riemann Sphere, how about the Thompson/Ritchie/McBride OS (Darl doesn't deserve credit, but if it'll make him drop the suits, let's give it to him).
2) Stop selling the product until the GPLed code was replaced with proprietary code, and re-release it,
/usr/src/linux/fs/binfmt_flat.c while writing a loader that swallows the output of elf2flt? The modified source of the latter would be released to OEMs writing dynamically loaded modules, but the loader would be in a closed-source ROM.
Under some definitions of derived work, it might not be possible or might be very hard to write new modules to replace the GPL-licensed imported ones without the new code being a derived work of that GPL'd code.
For instance, Linus claims that if you include the kernel headers, you're creating a derived work.
I'll accept that, but if I use methods I learned from GPL'd source code to write my own, I should be in the clear. It's the expression of an idea, not the idea itself, that can be protected.
However, GPL'd code authors aren't like corporate coders who have a legal staff to handle IP infractions. Joe invents something very cool and tricky that's widely admired. I adapt his method, but not his implementation and use it in a product which makes my company millions. Joe and his admirers might be very upset with me and while Joe might not be able to sue, we'll be exposed to a big P.R. mess.
That said, anyone have an opinion on sneaking a look at
I think the best approach is to look at binfmt_flat.c while writing a spec of how to load a flat binary, then put away the source while I write my own implementation or even better, hand the spec to another coder. It's not as simple, though, as some suggest to comply with the edge cases of the GPL and the penalty is not as simple as some claim.
You ought to be able to hire a lawyer for one hour (e.g. $200) to get your facts straight, and then not have to pay another dollar (unless your case really needs it) until after the judge/jury has given the verdict.
I did this once. The lawyer charged by the 10th of an hour (6 minutes) at $125/hour or something. I met with him in his office for two hours, then called him for another 10 minutes to clarify something.
I had about 30 pages of stock option contracts I wanted an opinion on. It would have been a fortune to have him read the whole thing, so I pointed out the sections that confused me & he clarified the contract law issues.
The main legal doctrine I learned about was that when a point in a contract is ambiguous, courts are supposed to favor the interpretation of the party who didn't write the contract.
I didn't have to put him on retainer, but if things had not turned out my way (the company hadn't bought back my stock when I quit), he would have been available for more questions & either filing suit or referring to someone who could have.
That referral mechanism is the key-- I called this particular lawyer from a listing (not an ad) in the yellow pages. Some of the lawyers I called never returned my messages. It's an insular club that if you don't have a brother who has a friend who plays raquetball with a lawyer, can be hard to penetrate (unless it's injury claims).
By the way, I think the rumor that I'd consulted a lawyer helped to make management take the whiners more seriously. In retrospect, I wish we'd buckled down & made the best damn code around, but the "lottery factor" (possibility that you might get a whole lot of something for nearly nothing) clouded our judgement.
I started with a K-1000, but when it was $130 from K-mart in 1984. They aren't made anymore and are more expensive than warranted due to (overblown) reputation.
Yes, they're tough (mine still worked fine 4 years ago with no CLA (clean, lubricate and adjust) when I traded up to a Super Program), but they're lacking:
1) Crappy meter. Slow to react and wierdly non-linear at low light levels, so not good for existing light photography with an f/1.4 50mm lens & 400ASA film.
2) No depth of field (a.k.a. depth of focus) preview. This is a hard feature to learn how to use, but control of DOF is a big part of learning photography and one area where 35mm kicks the crap out of point-and-shoot digicams (which have small sensors, short focal lengths and deep DOF so hard to knock the background out of focus for portraits).
3) Slow flash sync (X) speed. 1/60th, right? Once you learn manual existing-light photography, you might want to try manual (guide number/focus distance) flash photography. For fill-flash (lighting up a face shaded by a hat brim or eyes shaded by brow), faster sync gives you flexibility.
[I actually don't recommend trying to learn to use bounce & other tricks to make flash look more natural on anything but digital unless you have a darkroom. Too much lag between exposure & result to figure out what you're doing]
4) rubberized-cloth fully mechanical shutter. This means the battery only powers the meter & the camera will work with no battery at all. However, it isn't as accurate as quartz-controlled metal blade shutters like in the SuperProgram.
That said, the Pentax line is nice because the lenses work on the new bodies (including their digital *ist), though sometimes metering doesn't work. Nikon is the only other mfg. that kept the mount the same when they went autofocus-- Canon & Minolta changed. Minolta still makes their manual focus cameras, though. Canon manuals are orphaned with parts getting harder to find.
---
Make sure you get a "fast" lens. 85mm or 100mm f/2 or 50mm f/1.4. It's damned hard to focus an f/2 50mm lens (which came on my K1000 originally) because the DOF wide-open is too deep to give you a "snappy" focus.
---
Oh and KEH for mail-order used.
You could use these to take drugs from one place to another.
Makes a lot more sense than search-and-rescue. For that application, I'd want a Moto-Guzzi powered super-slow UAV.
Drugs are a great civilian app for cruise missles, though: low altitude and high speed help evade detection and value of goods makes up for high cost of jet fuel and low payload capacity.
Maybe a more acceptable app would be delivering *legal* drugs (antibiotics, HIV triple-cocktail) to regions with poor transport infrastructure, esp. with ongoing civil wars (I've been watching too much e.r.).