| |||||||||||||
safety-critical systems can use ROMsafety-critical systems can use ROMPosted Oct 18, 2006 7:00 UTC (Wed) by bojan (subscriber, #14302)In reply to: safety-critical systems can use ROM by jstAusr Parent article: FSF should separate GPLv3 changes (Linux.com)
> Which seems to answer your complaint about the first paragraph.
So, who gets the keys for a mobile phone? I would think according to GPLv3, that would be the end user (i.e. a person buying a mobile phone). Now if that's the case, can't they do whatever they like, including changing the output of the device, the frequencies it operates on etc.?
I don't think regulators would like that.
(Log in to post comments)
safety-critical systems can use ROM Posted Oct 18, 2006 7:31 UTC (Wed) by man_ls (subscriber, #15091) [Link] I don't think regulators would like that.Regulators don't like it either when people stick a knife into each other; you are still allowed to buy pointy cutlery. A silly but perhaps significant example. My wireless router (a 3COM WDR100) asks at initialization the country it will be used in (and warns that "it might be illegal to choose incorrectly). If I say US or Japan instead of Spain it might use an illegal part of the spectrum here. With the hacked WRT54GL I also have, it will probably give me that choice too (haven't checked really). Would it be illegal? Probably. Do regulators like it? Who cares, I'm a responsible person and promise solemnly not to do it.
safety-critical systems can use ROM Posted Oct 18, 2006 8:54 UTC (Wed) by edomaur (subscriber, #14520) [Link] >> Who cares, I'm a responsible person and promise solemnly not to do it.
IRL, this is not the case for every person who lives on this Earth.
safety-critical systems can use ROM Posted Oct 18, 2006 9:14 UTC (Wed) by man_ls (subscriber, #15091) [Link] You can only regulate so much -- I would say it is reasonable to do so against careless modification, but not to prevent knowledgeable people from doing so. Just requiring to update the firmware is a reasonable barrier of entry IMHO, at least for things such as wireless spectrum use. We are not talking about spectrum scanners or police radio receivers (and they would be hard to prevent anyway).
safety-critical systems can use ROM Posted Oct 18, 2006 10:25 UTC (Wed) by nix (subscriber, #2304) [Link] ... and if they maliciously pick a wrong value, then that's for the justice system to deal with. (Note that the justice system can distinguish between malicious and non-malicious intent, which code cannot. Hell, code can't even determine that you have specific permission to use some frequency, and bans you anyway: viz Alan Cox's complaints that frequency governors in some wireless systems forbid him from using frequencies which he *is* in fact allowed to use due to an amateur radio license...)
If everything followed your criteria, it would be impossible to invite a guest into your house (they're not the owner! they're probably a burglar!)
safety-critical systems can use ROM Posted Oct 18, 2006 10:19 UTC (Wed) by bojan (subscriber, #14302) [Link] > Regulators don't like it either when people stick a knife into each other; you are still allowed to buy pointy cutlery.
In some countries, it is illegal for manufacturers to ship devices that aren't "locked" to particular settings. And that's made illegal because the regulators think that users shouldn't be able to fiddle with devices in such manner. Whether this is "right" or not, is another matter.
So, if there was a device on such a market with GPLv3 software on it, the manufacturer would be forced to discontinue it (maybe even recall it), as it would not be compliant with regulations.
safety-critical systems can use ROM Posted Oct 18, 2006 14:36 UTC (Wed) by coriordan (guest, #7544) [Link] Nah, just stick the small amount of regulation-fettered logic into ROM, or some other modifiable technology, or don't give the users access to that part of the software storage - and put the rest of the code somewhere that the user can modify it.
Using telephones as an example, the software for setting the frequency etc. might have to go into ROM, but the rest could be left in user-modifiable storage.
safety-critical systems can use ROM Posted Oct 18, 2006 19:57 UTC (Wed) by RareCactus (guest, #41198) [Link] But what if the user lives in a country where he needs frequency X, but he only has a phone that is locked to frequency Y?Then the company that made the phone software is in violation of the end user clause of the GPLv3.
This is just one example of why the GPLv3 is a terrible idea, and is going to hurt commercial adoption of open source software. Companies avoid legal grey areas like this like the plague, because they don't want to waste time and money on legal hassles.
Of course, RMS doesn't care about stuff like this. He's happy to sit in his ivory tower and tinker with HURD, which they rewrite every few months or so (heavily borrowing from the Linux sources of course.) RMS does not believe in choice-- he believes that all software should be open source, and that closed source software is immoral. I am NOT kidding about this, read his web page if you doubt me.
But Linus, who is a running a real project that is making a real difference in the world, recognizes that this license is a poison pill for open source projects, and is happy to avoid it. Good for him, and for us who use and contribute to Linux.
safety-critical systems can use ROM Posted Oct 18, 2006 20:40 UTC (Wed) by RareCactus (guest, #41198) [Link] Ok, I have been searching the text of the GPLv3 to see just how it proposes to enforce end-user modification rights. This paragraph at the end of section 2 seems relevant:The Corresponding Source also includes any encryption or authorization keys necessary to install and/or execute modified versions from source code in the recommended or principal context of use, such that they can implement all the same functionality in the same range of circumstances. (For instance, if the work is a DVD player and can play certain DVDs, it must be possible for modified versions to play those DVDs. If the work communicates with an online service, it must be possible for modified versions to communicate with the same online service in the same way such that the service cannot distinguish.) A key need not be included in cases where use of the work normally implies the user already has the key and can read and copy it, as in privacy applications where users generate their own keys. However, the fact that a key is generated based on the object code of the work or is present in hardware that limits its use does not alter the requirement to include it in the Corresponding Source. And section 3:
3. No Denying Users' Rights through Technical Measures. No covered work constitutes part of an effective technological "protection" measure under section 1201 of Title 17 of the United States Code. When you convey a covered work, you waive any legal power to forbid circumvention of technical measures that include use of the covered work, and you disclaim any intention to limit operation or modification of the work as a means of enforcing the legal rights of third parties against the work's users. So maybe I was incorrect in saying that using ROM to provide constraints on the program would be contrary to the license. I'm not sure-- I'm not a lawyer. :( In any case, there are still enough odious and ambiguous clauses in this license that I believe any sane company wouldn't touch it with a ten-foot pole.
safety-critical systems can use ROM Posted Oct 18, 2006 20:42 UTC (Wed) by coriordan (guest, #7544) [Link] I can't make sense of your scenario.
For one, I don't think any country sells telephones that don't work in other countries.
More to the point, the fact that it might be illegal for a company in whatever country to sell phones that broadcast outside of whatever range is not something that can be fixed by GPLv3.
If the company is required to lock down the frequency, they have to either put it in ROM, use DRM, or place a physical barrier (plastic casing or whatever) between the software container and the outside world. This is dictated by law, not by our licences.
GPLv3 says that DRM isn't an option, so the phone maker will have to go with ROM or a lump of plastic. The effects on phone buyers is the same.
safety-critical systems can use ROM Posted Oct 18, 2006 21:41 UTC (Wed) by bronson (subscriber, #4806) [Link] GPLv3 says that DRM isn't an option, so the phone maker will have to go with ROM or a lump of plastic. The effects on phone buyers is the same....Until the phone buyer needs to upgrade the firmware on his handset. Maybe he wants a fix for a manufacturer defect, or for his phone to follow the new bluetooth standards, or just add a feature. Happens all the time. Yet, if the software is in ROM, the user is SOL. How can anybody possibly think that my freedom is increased by putting the software that I use into ROM instead of Flash? This just boggles the mind.
safety-critical systems can use ROM Posted Oct 18, 2006 22:53 UTC (Wed) by man_ls (subscriber, #15091) [Link] Putting a small portion of software in ROM and letting the rest be user-serviceable helps freedom. Having a device which can be upgraded by the manufacturer but not by you does not help freedom; the software might as well be burnt in ROM and we would not have the illusion of freedom. Sometimes small locks and visible bars help freedom.
No one uses ROM anymore, get over it already Posted Oct 19, 2006 1:59 UTC (Thu) by svkelley (guest, #37299) [Link] What people don't seem to understand is that no one uses ROM any more in embedded devices. It is all programmable flash. What is clear is that the people working on the GPLv3 draft really lack any knowledge of modern embedded systems and the components that make them up.
Sean
Not all flash is updateable Posted Oct 19, 2006 7:23 UTC (Thu) by man_ls (subscriber, #15091) [Link] I would imagine that not all flash memory inside a device must be made user-serviceable. Even that it takes some effort to make it updateable from software. So, set the flash contents in the factory and just avoid upgrades on the field, and effectively you have a ROM, right?
Not all flash is updateable Posted Nov 2, 2006 17:27 UTC (Thu) by wookey (subscriber, #5501) [Link] Not really. Both nor and nand flash are intrinsically read/write. You could wire up a flash chip with the write line tied down so it couldn't be used bu then there is a problem about how to get the code into the device in the first place. Myabe you could do it with JTAG, but normally you have use JTAG on the CPU which then used the write line to get data into the chip.
In theory you could put some content in the chip before soldering it down, but the whole production process is now set up assuming that you don't have to do this sort of thing any more (and we all save money because of it).
So the 'just put it in ROM' is not a trivial thing. It requires significant design and production changes, if it is possible at all.
safety-critical systems can use ROM Posted Oct 18, 2006 23:12 UTC (Wed) by bojan (subscriber, #14302) [Link] > GPLv3 says that DRM isn't an option, so the phone maker will have to go with ROM or a lump of plastic. The effects on phone buyers is the same.
I don't think it's the same. Manufacturers prefer options that are cheap, because consumers prefer to buy cheaper products. In a mass production scenario (and all "consumer" devices are such), the emphasis is low cost. Putting yet another protection mechanism in place increases the cost and complexity for the manufacturer, not to mention reduces flexibility with the ROM option. Instead, they can use this money to purchase proprietary software that doesn't have the "restrictions" that this hypothetical GPLv3 software has. And they get where they want to go with less hassle.
The other player here, of course, is the mobile phone (or other service type) company providing the service. They may be inclined to like manufacturers of "flexible" but "locked" phones better than the ones that need physical intervention in case something goes wrong. After all, the user has a contract that defines conditions of entry to the network. The "locked" software here provides a convenient way for the service provider to have an easy upgrade path (in case of errors in software, changed regulation, changed contract conditions etc.), while having reasonably difficult to "hack" technical measures in place against potential disruptions on the network by users modifying devices in order to go around contract conditions.
We need to understand that it's not going to be engineers making those decisions. It's going to be accountants. The end effect would most likely be that such software would not be used in such devices. Whether this is good or bad for FOSS remains to be seen.
safety-critical systems can use ROM Posted Oct 20, 2006 8:59 UTC (Fri) by coriordan (guest, #7544) [Link] I don't think the numbers will square up. The cost of using a ROM chip, or of adding some tamper-proof seal, is probably few cents in a 100 euro phone. Whatever the cost is, I'm sure it's less than the point at which hardware manufacturers round out the figures. I don't know the marketing terms, but what I means is that if the phone plus a standard profit margin yields a price of 98 euro or 101 euro, the manufacturer will round those numbers up or down to 100 euro. I think the cost of using a ROM chip in mass production will be certainly less than 1 euro. Or whatever the cost is, it will be significantly less than having two computing systems in one - something that Motorolla find cost effective just to have a strong separation between modifiable and non-modifiable bits.
safety-critical systems can use ROM Posted Oct 18, 2006 22:42 UTC (Wed) by man_ls (subscriber, #15091) [Link] RMS does not believe in choice-- he believes that all software should be open source, and that closed source software is immoral.Stallman is known for his strong opinions on this matter, yes. Surprise surprise: believing that something is immoral is not the same as believing that something is evil, shoud be banned or the perpetrators should be executed on the spot. We people do immoral things all the time and we live on. By the way, if you said that to Stallman's face you would be treated to one of his "free software is different from open source" speeches. You would probably want to avoid that, that (and not morality) might be the real reason why people who have met Stallman or just seen him in action avoid the phrase "open source". Of course, RMS doesn't care about stuff like this. He's happy to sit in his ivory tower [...] But Linus, who is a running a real project that is making a real difference in the world, recognizes that this license is a poison pill for open source projects, and is happy to avoid it.You might be surprised to learn that GNU software (built by Stallman and accolytes) is used even more broadly than the Linux kernel (built by Torvals and company): it is also used in the *BSD family and on multiple proprietary Unices, and also on Windows and Mac OS X. It is hard to find a computer anywhere that cannot run any GNU software, and most do run it. By the way, from his ivory tower he created a license that governs now about 350 million lines of code (conservative estimate by Ingo Molnar, should be closer to a billion). To put this in perspective, it is about 70 to 200 times the size of the Linux kernel. To each his own; you may not like Stallman, but I would say he is acutely aware of actual computer programming issues. That is why he is designing the GPLv3. It is on purpose. Yes, really.
FSF software used widely? Posted Oct 19, 2006 20:33 UTC (Thu) by vonbrand (subscriber, #4458) [Link] Come on, what is used widely is BSD (and similar)-licensed stuff, i.e., sendmail, (La)TeX, X, apache, and a long list of other stuff. What the FSF really has built is a tiny fraction of open source software, and that (together with most software available freely in source code) was propelled to center stage by Linux. Before around '97, the whole GPL code was stuff that was played around with at universities and at best a curiosity outside, the only notable exception being the GCC stuff (courtesy of Cygnus, building on rather primitive FSF beginnings), and perhaps emacs (mostly in the form of xemacs).
FSF software used quite widely Posted Oct 19, 2006 23:09 UTC (Thu) by man_ls (subscriber, #15091) [Link] Come on, what is used widely is BSD (and similar)-licensed stuff, i.e., sendmail, (La)TeX, X, apache, and a long list of other stuff.Just so you broaden the spectrum of software you use every day, let me point you to some useful links: try bash as your command shell. Once you have your BSD-licensed X server running, be sure to try out GNU Object Model Environment, better known as GNOME, as your desktop: it runs on some tens of Linux distributions apart from commercial Unices and BSD variants. Also try GIMP (GNU's image manipulation program) to manipulate your images. Now that I think about it, just browse the or visit your favorite mirror, for such gems as glibc, ghostscript, gawk, wget, patch or GNU tar. They are quite useful if you ever want to put together a Linux distribution, or even a *BSD variant. You might have enough with Linux and BusyBox though, if you don't want a graphical environment; if you do, be sure to get acquainted to GTK. It is quite popular; used in Dia, Gnumeric, GnuCash and a thousand other programs.
Yes, you will probably need BSD-licensed software. Lots of it. I'm glad it is there too, and I'm thankful for the people who wrote it. There is no point in diminishing their good work. I found some references saying that a Linux distro is 3% Linux, 28% GNU software. They are from 1999 though; I haven't found anything more recent. I would venture that Linux is still playing catch up in 2006, but you seem really knowledgeable and will surely be able to supply better figures. :P What the FSF really has built is a tiny fraction of open source softwareThe FSF (and in particular Stallman) wrote the GPL. The estimate of 350+ million lines of code under the GPL comes from Ingo Molnar, kernel developer who is not so fond of the FSF; still he would probably bet for a billion rather. I wouldn't say that this is "a tiny fraction of open source software" unless I was trying to discredit the FSF. Of course not all of it was built by the FSF, but the authors liked the license enough that they generously put their work at your disposal under its conditions. Not that I want to confuse both things (code written by the FSF and code under the GPL), but since you speak about "the whole GPL code" later on, I take it that you noticed that it is an important contribution of the FSF. Even if I was trying to discredit the FSF, I would choose a different field, really. Even in a discussion about the GPLv3: I would try to dispute other facts, not the influence of the FSF in libre software.
|
|||||||||||||