Although the article says that hard disks that held classified data will still be destroyed, there still is lots of information that is unclassified but is not supposed to be released, for example, private personal information, a large body of information that should have been classified because classified information can be derived from directly from it, etc. You can bet that the US military is much more familiar with this issue than the insulting "Pentagon Discovers dd" slashdot title implies.
Note to slashdot editors: if you value your readers' time, please expand all but the most common acronyms or abbreviations the first time you use them, even if you are quoting an email that you received. For example, the article that does not require registration does not mention an "Nd:YAG laser" and very cursory search on google did not turn up a definition (although it did find references to the term).
Might be more useful if AuDSL were freer
on
DSLBlaster?
·
· Score: 1
I think it's good when anyone writes software that is useful for some purpose, even if it's completely proprietary, so I'm not saying that the AuDSL authors have done a bad thing, but I'd like point out that AuDSL is not currently free software, according to the "COPYRIGHT" file in their source distribution:
5. Redistribution of the code as part of, or in association with,
a hardware product specifically designed to perform data
communications or audio signal signal processing by means of
the code is prohibited without the specific prior written
permission of Araneus Information Systems Oy.
Also, if you want to go for compatability with the GNU General Public License, the Free Software Foundation claims that these sorts of advertising restrictions are GPL-incompatible (item 3 in the AuDSL COPYRIGHT file).
If AuDSL were freed, I think their encoding implementations might be useful for the various free soft modem efforts, especially with the cheap Computer Telephony Integration risers available on some motherboards that basically connect the built-in AC97 sound hardware to the phone line. I also wonder if the techniques they use to get 96kbps might be usable to provide a "poor man's DSL," although I've heard that some semi-soft modem hardware that includes a Digital Signal Processor is theoretically capable of doing DSL, in which case there might not be anyone who would find this useful and rich enough to afford a premium phone service.
DDR has about the same latency as SDRAM (slightly worse latency in some configurations), but twice the bandwidth.
SDR and DDR both take about six cycles to set up the transaction, then SDR can do one transfer per clock and DDR can do two per clock. Virtually all motherboards have a 64-bit wide memory bus, so a Pentium 3's 32 byte cache line requires four transfers, an Athlon's 64 byte cache line requires eight transfers and an Pentium 4's 128 byte cache line requires sixteen. So, each SDR cache fill on a Pentium 4 should require 22 cycles (6 setup + 16 transfer), while each DDR cache line fill should require 14 (6 setup + 8 transfer), an improvement of 57%.
Nothing against installwatch, but just for everyone's information, logwrites does much the same thing and has been around much longer. logwrites has been available under GPL since May, 1996 and was announced on comp.os.linux.announce on June 18, 1996.
I think installwatch's distinct features are that it logs some operations that do not create or delete file names (like chown), has a handy shell script command and can use syslog. A distinct feature of logwrites is that it separates the system call traps from the logging into two libraries. The simple logging library can be used elsewhere for recording additional events relevent to installation history in the same format, without invoking the system call trapping. For example, we run modified cp, mv, and install to record where a file came from.
Both programs are pretty trivial, so the duplication of effort has been tiny. Nevertheless, this situation illustrates that it can save you time to check around for an existing piece of free software before you start coding. You may be surprised what people have already released.
I have no connection with E-Tenna thing or with EETimes. However, I would like to step forward and say that for years I have had more respect for EETimes than any of the numerous trade journal that I receive, bar none.
Just about every other trade journal will take news of a patent as a major technological breakthrough and proof of great technical leadership within a company. EETimes seems to scrutinize the actual technology and give at least some coverage to small companies and graduate student project. They also occasionally cover the arguments of those who identify patent and copyright abuse, attempts to make government works copyrighted, and export restrictions. EETimes does not claim to be a referreed technical journal or even "hard news" like The San Jose Mercury News, but I think it's probably the best trade journal that I have come across. I, for one, respect them as much and usually more than the sources on which Slashdot stories are based.
I think one of the greatest efficiencies of free software is in reducing friction in reuse of other work in ways that may not have been imagined originally. For this to occur easily enough so that this efficiency is not offset by some other inneficiency (like negiating license exceptions), mutually compatible copying conditions are necessary.
Currently, the largest set of mutually compatible free software available is the set defined by GPL compatibility. This includes the new BSD copying conditions, the "MIT style" copying conditions, public domain, and many other varieties. The GPL also covers many crucial system components, like gcc. So, for reasons of both the amount of software and the specific software covered, it is unlikely that a superior pool of mutually cominglable free software will be created in the near future that does not include some GPL'ed software.
So, currently, if you care about the network effects and efficiency benefits of Open Source,
even if you don't give a hoot about Free Software, your most effective maintenance strategy is to use GPL-compatible copying conditions (not necessarily the GPL), not because you think the GPL came from heaven, but primarily because the GPL has itself become a dominant technology compatibility standard.
This License Agreement shall be governed by and interpreted in all respects by the law of the State of Virginia, excluding conflict of law provisions.
I believe the term "this license agreement" refers only to the CNRI license, not to all other agreements that might be offered with respect to the copyrights of other works that Python comprises. For example, the BeOpen material is covered by a California choice-of-law provision.
As I read it, the CNRI and BeOpen conditions do not say that the GPL must be interpreted under Virginia or California law, nor do they (nor could they) require that litigation be conducted in California or Virginia.
Also, since CNRI knowingly published their works for years without a copyright notice and knew that these works were being treated as though they were covered by the CWI copying conditions, I think CNRI's copyrights are probably unenforceable because of the "sleeping on your rights" doctine.
On the other hand, I am not a lawyer, so don't take this as legal advice. In fact I'm not even taking my advice: just to be careful, we use a much older version of Python in the Yggdrasil build tree. (Yeah, I know its been years since a
release, but there are still nightly builds and now rsync'ing. You could arguably call me crazy for it.).
No string that begins with one or more zeroes can be encoded this way, so no prime number begins with one or more zeroes when written the normal decimal form (for that matter, no integer other than 0 begins with a zero when written this way).
I do not know if every finite string of digits can be found in e or pi, but I can show it is not true that every finite string of digits can be found in every irrational number (although the original author did not claim that). Consider the irrational number where the Nth decimal digit to the right of the decimal point is 1 if the N is a perfect square and 0 if not. That is:
1.10010000100000010...
That number is irrational, and no string of digits containing, say, a "2" every appears in it.
I guess, but it's quite conceivable that they could have, and would only have been (legally) granted the patent once it was done in
software.
You have not provided any reason why the patent office would refuse to grant a patent on a mechanical RSA device. Therefore, it still seems that a patent on RSA would be a patent on the "technological advance" and not just a patent on "the software."
If there are these supposed patents on just "the software" in the US that do not even theoretically cover a "technological advance", it should be easy to list some good examples. It appears that, so far, nobody has been able to find an example or even construct a hypothetical instance with an example of (hypothetical) infingement of just "the software".
As I read the British recommendation, the RSA algorithm would represent a "technological advance" and therefore be patentable. It's not like people built or wrote about mechanical public key cryptography devices using modular exponentiation before and all the RSA people did was implement it in software.
A patent can be granted in the US if the process that is happening in software is an old process, but not in software. And a
patent can be granted for "frivoulous" innovation because the line for innovation has not been clearly drawn.
This is how urban myths get started. I have been following software patents in the US for more than a decade. I am not a lawyer, but I have read through much of Title 17 (all the relevant parts, I believe), and many court decisions. If you cannot provide even a hypothetical example to illustrate this supposed distiction, can you identify a law or court decision (in the US) that makes the distinction you are supporting?
Please do not simply quote another unsubstantiated slashdot message. Again, that is how urban myths get started.
Could you provide an exmple (hyptothetical if necessary, but specific) of a situation where the "software" would be patented but the "technological innovation" in it would not be, and then give an example of infringement?
19. The Government's conclusion is thus to reaffirm the principle that patents are for technological innovations. Software should not be patentable
where there is no technological innovation, and
technological innovations should not cease to be patentable merely because the innovation lies in
software.
The "St. Louis Business Journal" Article that
Slashdot points to has a
link
at the botton labelled "Click for permission to reprint",
and the page that it takes you to claims to be able to charge $5 for the right to link, as if copyright could restrict that. As I see it, this makes the "St. Louis Business Journal" a scam intended to take advantage of people who don't understand copyright and fair use, and, therefore, not a trustable source of news.
It's only social pressure, but I would like to recommend that slashdot label unconfirmed "St. Louis Business Journal" stories as rumour, since their misleading statements about copyright show that publication by the "St. Louis Business Journal" does not make it particularly likely that a statement is true. Also, keeping the "St. Louis Business Journal" in quotation marks helps make it clear that that site is not acknowledged as the journal of record for business in St. Louis.
I am not a lawyer, so don't rely on this as legal advice, but I believe that in the US, amusingly named the Supreme Court case Hustler v. Moral Majority (1988?) would apply.
In that case, the Moral Majority had sent a copyrighted letter to some of its members complaining about Hustler Magazine. Hustler then reproduced the letter in its entirety and distributed it to a much wider audience to solicit financial contributions in support of freedom of speech. The Moral Majority sued for copyright infringement, and the Supreme Court sided with Hustler, ruling that this kind of reproduction was fair use, even though the document was reproduced verbatim in its entirety and used for financial gain.
This type of free speech is very important in maintaining accountability. There are so many times when I have seen subtle threats or other underhanded tactics perpetrated by email and am glad that the victim of those attacks posts this "private" email publicly. The possibility that such email abuse may be made public by the victim provides an important check. This is especially important where the perpetrator knows that the victim is unlikely to have the resources to spend ~$100k per instance to spend on the next best recourse, litigation.
For instance, if we did not have this right, you can bet that a lot of spam would be written to fit whatever the legal criteria of "private" was to prevent victims from forwarding the evidence.
I remember seeing a documentary in the last year about Battlestar Galictica that mentioned that the fellow who played Commander Apollo had started shooting a sequel to it but was stopped by some obnoxious copyright claims by one of the original producers. The documentary did show some of a trailer for the new movie and that piece of the trailer was literally Star Wars level. It was incredible.
The BSD distributions no longer promote GPL-incompatible copying conditions by default. The new BSD copying conditions are the result of cosultation to achieve GPL compatiblity in the eyes of the FSF. The Univesity of California also agreed to the change for the BSD software that they own the copyrights to.
Whether or not OpenSSH is legally required to
change their name given the prior permissions on earlier versions of SSH is a minor issue, since only the Linux/BSD distribution makers need to track the change and the updated software will continue to ship in all of the distributions.
What is of more concern to me is whether the shell command "ssh" will have to change, since that would actually make a lot shell scripts and other programs incompatble.
Ricochet modems also have a packet-based peer-to-peer "STAR" mode which Linux supports
through the
strip.o kernel module, basically
acting like a low bandwidth but longer range wireless ethernet, at, I believe, the same legally limited 1 watt of power used by the Metricom pole top repeaters. Stanford University has a network of these things called
MosquitoNet.
At ~10X the range, and therefore ~100X the coverage area of 802.11b wireless ethernet, the 128kbps $99 metricom units could easily be used by
nerds or local ISP's to blanket most metropolitan areas with their own wireless internet service.
By the way, since metricom modem cards are made by
separate companies like Novatel and Sierra Wireless (don't know about the external modems) and the ISP's are also independent
companies, I think Metricom-based
networks would find a way to continue if,
heaven forbid, Metricom were to go under. I
certainly hope the Metricom people make a fortune. They have made a great product, which I use every day.
Your project, slashcode, sourceforge, parts of gnu mailman, etc. all point to the need to a standard file layout for web server programs, so that people could install these big services on the web servers as easily as they today unpack tar files from / that put things in/usr/bin,/usr/lib,/var/man, etc.
OpenDivx is not
Open Source, at least as opensource.org defines the term. For example, you are forbidden by the
DivX copyright terms to use the software in a way that does not implement MPEG-4, and you are forbidden from using it where "Encoded Content is a primary or substantial product."
I'm not saying that releasing OpenDivX with a promotional copyright is itself harmful or immoral, but their misapplication of the "Open Source" term to a scheme that shares few of the public benefits that attract developers is misleading.
In practice, free software has solved the "software recycling problem" in a way that object oriented programming proponents promised and generally failed to do, because it is much easier to reuse free software in ways that were not originally anticipated, and it is economically unoptimal and often impossible to anticipate all of those ways.
For example, people adapt hardware drivers (kernel, X windows, etc.) to new hardware and there are frequently features that require internal changes that could not have been easily anticipated in the original programming interface.
This is one reason why mutually compatible copying conditions are so important to free software.
I also think that embarrassment over this why one well known OO proponent
denounced free software.
This is not to say that inheritance, data abstraction, and calls of the form object->function(object,...) are useless. They are useful at times, and sometimes even in the peculiar combination called "object oriented programming", but their benefits have been oversold and the demand for free software is a testament to the fact that OO failed to remove most of the barriers to reuse that make the proprietary model so much less efficient.
If you are claiming that specific patent, trademarks, copyright, or nondisclosure agreements would have to be executed, please provide identify them. Otherwise, what legal restriction are you referring to when you say that one would have to "license the technology?"
Slashdot Mirror
With specialized equipment, you can often read data that was overwritten on a hard disk, so dd is often not enough. See Secure Deletion of Data from Magnetic and Solid-State Memory, by Peter Gutmann of the University of Auckland. There is also a previous slashdot article on this subject.
Although the article says that hard disks that held classified data will still be destroyed, there still is lots of information that is unclassified but is not supposed to be released, for example, private personal information, a large body of information that should have been classified because classified information can be derived from directly from it, etc. You can bet that the US military is much more familiar with this issue than the insulting "Pentagon Discovers dd" slashdot title implies.
Note to slashdot editors: if you value your readers' time, please expand all but the most common acronyms or abbreviations the first time you use them, even if you are quoting an email that you received. For example, the article that does not require registration does not mention an "Nd:YAG laser" and very cursory search on google did not turn up a definition (although it did find references to the term).
I think it's good when anyone writes software that is useful for some purpose, even if it's completely proprietary, so I'm not saying that the AuDSL authors have done a bad thing, but I'd like point out that AuDSL is not currently free software, according to the "COPYRIGHT" file in their source distribution:
5. Redistribution of the code as part of, or in association with, a hardware product specifically designed to perform data communications or audio signal signal processing by means of the code is prohibited without the specific prior written permission of Araneus Information Systems Oy.
Also, if you want to go for compatability with the GNU General Public License, the Free Software Foundation claims that these sorts of advertising restrictions are GPL-incompatible (item 3 in the AuDSL COPYRIGHT file).
If AuDSL were freed, I think their encoding implementations might be useful for the various free soft modem efforts, especially with the cheap Computer Telephony Integration risers available on some motherboards that basically connect the built-in AC97 sound hardware to the phone line. I also wonder if the techniques they use to get 96kbps might be usable to provide a "poor man's DSL," although I've heard that some semi-soft modem hardware that includes a Digital Signal Processor is theoretically capable of doing DSL, in which case there might not be anyone who would find this useful and rich enough to afford a premium phone service.
DDR has about the same latency as SDRAM (slightly worse latency in some configurations), but twice the bandwidth.
SDR and DDR both take about six cycles to set up the transaction, then SDR can do one transfer per clock and DDR can do two per clock. Virtually all motherboards have a 64-bit wide memory bus, so a Pentium 3's 32 byte cache line requires four transfers, an Athlon's 64 byte cache line requires eight transfers and an Pentium 4's 128 byte cache line requires sixteen. So, each SDR cache fill on a Pentium 4 should require 22 cycles (6 setup + 16 transfer), while each DDR cache line fill should require 14 (6 setup + 8 transfer), an improvement of 57%.
Nothing against installwatch, but just for everyone's information, logwrites does much the same thing and has been around much longer. logwrites has been available under GPL since May, 1996 and was announced on comp.os.linux.announce on June 18, 1996.
I think installwatch's distinct features are that it logs some operations that do not create or delete file names (like chown), has a handy shell script command and can use syslog. A distinct feature of logwrites is that it separates the system call traps from the logging into two libraries. The simple logging library can be used elsewhere for recording additional events relevent to installation history in the same format, without invoking the system call trapping. For example, we run modified cp, mv, and install to record where a file came from.
Both programs are pretty trivial, so the duplication of effort has been tiny. Nevertheless, this situation illustrates that it can save you time to check around for an existing piece of free software before you start coding. You may be surprised what people have already released.
I have no connection with E-Tenna thing or with EETimes. However, I would like to step forward and say that for years I have had more respect for EETimes than any of the numerous trade journal that I receive, bar none.
Just about every other trade journal will take news of a patent as a major technological breakthrough and proof of great technical leadership within a company. EETimes seems to scrutinize the actual technology and give at least some coverage to small companies and graduate student project. They also occasionally cover the arguments of those who identify patent and copyright abuse, attempts to make government works copyrighted, and export restrictions. EETimes does not claim to be a referreed technical journal or even "hard news" like The San Jose Mercury News, but I think it's probably the best trade journal that I have come across. I, for one, respect them as much and usually more than the sources on which Slashdot stories are based.
I think one of the greatest efficiencies of free software is in reducing friction in reuse of other work in ways that may not have been imagined originally. For this to occur easily enough so that this efficiency is not offset by some other inneficiency (like negiating license exceptions), mutually compatible copying conditions are necessary.
Currently, the largest set of mutually compatible free software available is the set defined by GPL compatibility. This includes the new BSD copying conditions, the "MIT style" copying conditions, public domain, and many other varieties. The GPL also covers many crucial system components, like gcc. So, for reasons of both the amount of software and the specific software covered, it is unlikely that a superior pool of mutually cominglable free software will be created in the near future that does not include some GPL'ed software.
So, currently, if you care about the network effects and efficiency benefits of Open Source, even if you don't give a hoot about Free Software, your most effective maintenance strategy is to use GPL-compatible copying conditions (not necessarily the GPL), not because you think the GPL came from heaven, but primarily because the GPL has itself become a dominant technology compatibility standard.
This License Agreement shall be governed by and interpreted in all respects by the law of the State of Virginia, excluding conflict of law provisions.
I believe the term "this license agreement" refers only to the CNRI license, not to all other agreements that might be offered with respect to the copyrights of other works that Python comprises. For example, the BeOpen material is covered by a California choice-of-law provision. As I read it, the CNRI and BeOpen conditions do not say that the GPL must be interpreted under Virginia or California law, nor do they (nor could they) require that litigation be conducted in California or Virginia.
Also, since CNRI knowingly published their works for years without a copyright notice and knew that these works were being treated as though they were covered by the CWI copying conditions, I think CNRI's copyrights are probably unenforceable because of the "sleeping on your rights" doctine.
On the other hand, I am not a lawyer, so don't take this as legal advice. In fact I'm not even taking my advice: just to be careful, we use a much older version of Python in the Yggdrasil build tree. (Yeah, I know its been years since a release, but there are still nightly builds and now rsync'ing. You could arguably call me crazy for it.).
No string that begins with one or more zeroes can be encoded this way, so no prime number begins with one or more zeroes when written the normal decimal form (for that matter, no integer other than 0 begins with a zero when written this way).
I do not know if every finite string of digits can be found in e or pi, but I can show it is not true that every finite string of digits can be found in every irrational number (although the original author did not claim that). Consider the irrational number where the Nth decimal digit to the right of the decimal point is 1 if the N is a perfect square and 0 if not. That is:
1.10010000100000010...That number is irrational, and no string of digits containing, say, a "2" every appears in it.
I guess, but it's quite conceivable that they could have, and would only have been (legally) granted the patent once it was done in software.
You have not provided any reason why the patent office would refuse to grant a patent on a mechanical RSA device. Therefore, it still seems that a patent on RSA would be a patent on the "technological advance" and not just a patent on "the software."
If there are these supposed patents on just "the software" in the US that do not even theoretically cover a "technological advance", it should be easy to list some good examples. It appears that, so far, nobody has been able to find an example or even construct a hypothetical instance with an example of (hypothetical) infingement of just "the software".
As I read the British recommendation, the RSA algorithm would represent a "technological advance" and therefore be patentable. It's not like people built or wrote about mechanical public key cryptography devices using modular exponentiation before and all the RSA people did was implement it in software.
A patent can be granted in the US if the process that is happening in software is an old process, but not in software. And a patent can be granted for "frivoulous" innovation because the line for innovation has not been clearly drawn.
This is how urban myths get started. I have been following software patents in the US for more than a decade. I am not a lawyer, but I have read through much of Title 17 (all the relevant parts, I believe), and many court decisions. If you cannot provide even a hypothetical example to illustrate this supposed distiction, can you identify a law or court decision (in the US) that makes the distinction you are supporting?
Please do not simply quote another unsubstantiated slashdot message. Again, that is how urban myths get started.
Could you provide an exmple (hyptothetical if necessary, but specific) of a situation where the "software" would be patented but the "technological innovation" in it would not be, and then give an example of infringement?
From the report:
19. The Government's conclusion is thus to reaffirm the principle that patents are for technological innovations. Software should not be patentable where there is no technological innovation, and technological innovations should not cease to be patentable merely because the innovation lies in software.
The "St. Louis Business Journal" Article that Slashdot points to has a link at the botton labelled "Click for permission to reprint", and the page that it takes you to claims to be able to charge $5 for the right to link, as if copyright could restrict that. As I see it, this makes the "St. Louis Business Journal" a scam intended to take advantage of people who don't understand copyright and fair use, and, therefore, not a trustable source of news.
It's only social pressure, but I would like to recommend that slashdot label unconfirmed "St. Louis Business Journal" stories as rumour, since their misleading statements about copyright show that publication by the "St. Louis Business Journal" does not make it particularly likely that a statement is true. Also, keeping the "St. Louis Business Journal" in quotation marks helps make it clear that that site is not acknowledged as the journal of record for business in St. Louis.
I am not a lawyer, so don't rely on this as legal advice, but I believe that in the US, amusingly named the Supreme Court case Hustler v. Moral Majority (1988?) would apply.
In that case, the Moral Majority had sent a copyrighted letter to some of its members complaining about Hustler Magazine. Hustler then reproduced the letter in its entirety and distributed it to a much wider audience to solicit financial contributions in support of freedom of speech. The Moral Majority sued for copyright infringement, and the Supreme Court sided with Hustler, ruling that this kind of reproduction was fair use, even though the document was reproduced verbatim in its entirety and used for financial gain.
This type of free speech is very important in maintaining accountability. There are so many times when I have seen subtle threats or other underhanded tactics perpetrated by email and am glad that the victim of those attacks posts this "private" email publicly. The possibility that such email abuse may be made public by the victim provides an important check. This is especially important where the perpetrator knows that the victim is unlikely to have the resources to spend ~$100k per instance to spend on the next best recourse, litigation.
For instance, if we did not have this right, you can bet that a lot of spam would be written to fit whatever the legal criteria of "private" was to prevent victims from forwarding the evidence.
I remember seeing a documentary in the last year about Battlestar Galictica that mentioned that the fellow who played Commander Apollo had started shooting a sequel to it but was stopped by some obnoxious copyright claims by one of the original producers. The documentary did show some of a trailer for the new movie and that piece of the trailer was literally Star Wars level. It was incredible.
The BSD distributions no longer promote GPL-incompatible copying conditions by default. The new BSD copying conditions are the result of cosultation to achieve GPL compatiblity in the eyes of the FSF. The Univesity of California also agreed to the change for the BSD software that they own the copyrights to.
Whether or not OpenSSH is legally required to change their name given the prior permissions on earlier versions of SSH is a minor issue, since only the Linux/BSD distribution makers need to track the change and the updated software will continue to ship in all of the distributions.
What is of more concern to me is whether the shell command "ssh" will have to change, since that would actually make a lot shell scripts and other programs incompatble.
Ricochet modems also have a packet-based peer-to-peer "STAR" mode which Linux supports through the strip.o kernel module, basically acting like a low bandwidth but longer range wireless ethernet, at, I believe, the same legally limited 1 watt of power used by the Metricom pole top repeaters. Stanford University has a network of these things called MosquitoNet.
At ~10X the range, and therefore ~100X the coverage area of 802.11b wireless ethernet, the 128kbps $99 metricom units could easily be used by nerds or local ISP's to blanket most metropolitan areas with their own wireless internet service.
By the way, since metricom modem cards are made by separate companies like Novatel and Sierra Wireless (don't know about the external modems) and the ISP's are also independent companies, I think Metricom-based networks would find a way to continue if, heaven forbid, Metricom were to go under. I certainly hope the Metricom people make a fortune. They have made a great product, which I use every day.
Someone please moderate the parent posting (not mine) up. I am glad to see somone posting useful references on this stuff.
Your project, slashcode, sourceforge, parts of gnu mailman, etc. all point to the need to a standard file layout for web server programs, so that people could install these big services on the web servers as easily as they today unpack tar files from / that put things in /usr/bin, /usr/lib, /var/man, etc.
OpenDivx is not Open Source, at least as opensource.org defines the term. For example, you are forbidden by the DivX copyright terms to use the software in a way that does not implement MPEG-4, and you are forbidden from using it where "Encoded Content is a primary or substantial product."
I'm not saying that releasing OpenDivX with a promotional copyright is itself harmful or immoral, but their misapplication of the "Open Source" term to a scheme that shares few of the public benefits that attract developers is misleading.
In practice, free software has solved the "software recycling problem" in a way that object oriented programming proponents promised and generally failed to do, because it is much easier to reuse free software in ways that were not originally anticipated, and it is economically unoptimal and often impossible to anticipate all of those ways.
For example, people adapt hardware drivers (kernel, X windows, etc.) to new hardware and there are frequently features that require internal changes that could not have been easily anticipated in the original programming interface.
This is one reason why mutually compatible copying conditions are so important to free software.
I also think that embarrassment over this why one well known OO proponent denounced free software.
This is not to say that inheritance, data abstraction, and calls of the form object->function(object,...) are useless. They are useful at times, and sometimes even in the peculiar combination called "object oriented programming", but their benefits have been oversold and the demand for free software is a testament to the fact that OO failed to remove most of the barriers to reuse that make the proprietary model so much less efficient.
If you are claiming that specific patent, trademarks, copyright, or nondisclosure agreements would have to be executed, please provide identify them. Otherwise, what legal restriction are you referring to when you say that one would have to "license the technology?"