Great Idea!!!
But I think there is something important missing out there: the government vehicles (like Prime Minister's Jaguar) should also have those spies installed.
As I have skimmed through the professor's analysis (exactly 64 pages, not 100) I have noticed one single important point he tries to make: you cannot depend on OSS in case of some damage. The OSS (L)GPL goes against the german law voiding the guarantee of compensating damages. But what the hell guarantee you have using prioprietary software? Has anyone been compensated for loss due to Windows misbehaviour or, say, Oracle DB bug?
The conclusion from this study IMHO is that generally software providers should compensate damages that software bugs cause, it should not only be the problem of the Open Source Community. From that point of view commercial licences are equally flawed.
The database access in Java is really a mess. On one side we have the EJB technology, now grown up and mature, but still having many problems: you cannot use built-in database sequences, instead you have to resort yourself to the well known hack (in the Java Community a hack is called "the pattern") with session bean and entity bean. Reading from database is rather slow - you have to use DAO (write them by hand or have generated like in OptimalJ) or resort to container specific practices - like in JBoss - where you define how many rows should be read at a time.
Even more mess is due to arrival of the new technologies duplicating the functionality of the EJB - like the JDO, which (as I have read in: JBoss chooses the JDO) may be the basis of enterprise Java in tommorow meaning not only the betterment of your applications but also meaning that new sources of incopatibility will emerge.
In this case one fears that his applications will have to be rewritten, the tools will change, the practices you learn now will become obsolete.
Additionally there is still the problem of using the database effectively. The more indirection you have in accessing the DBMS, the more portability you acquire, the less specific you can be. Those questions need answer if the Java Enterprise Technology shall continue to lead.
Even if Debian will not be the desktop distribution of the future, surely there will be many distributions basing on Debian and adding some functionality to it.
One of such excellent distributions is
Knoppix Live on CD distribution, which is one-CD linux able to run on almost any hardware in the universe. It runs even on machines on which RedHat and SuSE cannot (I have seen such!). The heart of Knoppix is its fantastic hardware-auto-configurator.
Debian is the base for Knoppix because it is very stable and has clear config files. I think that the Debian community should put Debian forward as a reference system and a potential basis for other fine-tuned distros.
I can see the problem of Software Patents in the EU from another point of view: if there will be no Software Patents the EU will suffer extreme pression from Bush administration (a.k.a. USA) to introduce such. USA will obiously claim that EU is posing a threat to Intellectual Property, etc. and all such bullshit.
On the other hand if the EU introduces SANE software patent rules (such as quoted on/.), then the Bush administration won't be able to come forward with such absurd claims.
May the Source be with you!
int main() {
int x, y;
if (x == 0)
y = 0;
}
MD5 (first.c) = 38c2cf1f11c9523b429d1234538bb3e1
int main() {
int p, q;
if (p == 0)
q = 0;
}
MD5 (second.c) = e79fe2328e03d64bc7c4b7d36f274761
I think that those two snippets of code are similar. What do YOU think?
In this case I see some likeness to the case of Generic Drugs. Brazil forced lower prices of patented drugs by threatening with ignoring those patents and producing cheap, generic medecines.
They won because a state is still more powerful than any corporation. Imagine what would happen if SCO won the case against Linux, while Brazil would have most of the governmental IT run by Linux. Would the surrender to the power of SCO? I doubt it. So every such case is beneficial to the stability of Open Source community
Did you perform any benchmarks of ReiserFS/Linux vs. UFS/FreeBSD (on the same machine)? My personal experience shows that UFS/FreeBSD is twice faster than ReiserFS/Linux or XFS/Linux. Perhaps the fault is in Linux Kernel? - because obviously ReiserFS is performing very good when confronted to any filesystem on Linux platform. If it were the case - would you consider porting your excellent FS to *BSD?
Slashdot Mirror
Great Idea!!!
But I think there is something important missing out there: the government vehicles (like Prime Minister's Jaguar) should also have those spies installed.
As I have skimmed through the professor's analysis (exactly 64 pages, not 100) I have noticed one single important point he tries to make: you cannot depend on OSS in case of some damage. The OSS (L)GPL goes against the german law voiding the guarantee of compensating damages. But what the hell guarantee you have using prioprietary software? Has anyone been compensated for loss due to Windows misbehaviour or, say, Oracle DB bug?
The conclusion from this study IMHO is that generally software providers should compensate damages that software bugs cause, it should not only be the problem of the Open Source Community. From that point of view commercial licences are equally flawed.
The database access in Java is really a mess. On one side we have the EJB technology, now grown up and mature, but still having many problems: you cannot use built-in database sequences, instead you have to resort yourself to the well known hack (in the Java Community a hack is called "the pattern") with session bean and entity bean. Reading from database is rather slow - you have to use DAO (write them by hand or have generated like in OptimalJ) or resort to container specific practices - like in JBoss - where you define how many rows should be read at a time.
Even more mess is due to arrival of the new technologies duplicating the functionality of the EJB - like the JDO, which (as I have read in: JBoss chooses the JDO) may be the basis of enterprise Java in tommorow meaning not only the betterment of your applications but also meaning that new sources of incopatibility will emerge.
In this case one fears that his applications will have to be rewritten, the tools will change, the practices you learn now will become obsolete.
Additionally there is still the problem of using the database effectively. The more indirection you have in accessing the DBMS, the more portability you acquire, the less specific you can be. Those questions need answer if the Java Enterprise Technology shall continue to lead.
Even if Debian will not be the desktop distribution of the future, surely there will be many distributions basing on Debian and adding some functionality to it.
One of such excellent distributions is Knoppix Live on CD distribution, which is one-CD linux able to run on almost any hardware in the universe. It runs even on machines on which RedHat and SuSE cannot (I have seen such!). The heart of Knoppix is its fantastic hardware-auto-configurator.
Debian is the base for Knoppix because it is very stable and has clear config files.
I think that the Debian community should put Debian forward as a reference system and a potential basis for other fine-tuned distros.
I can see the problem of Software Patents in the EU from another point of view: if there will be no Software Patents the EU will suffer extreme pression from Bush administration (a.k.a. USA) to introduce such. USA will obiously claim that EU is posing a threat to Intellectual Property, etc. and all such bullshit. /.), then the Bush administration won't be able to come forward with such absurd claims.
On the other hand if the EU introduces SANE software patent rules (such as quoted on
May the Source be with you!
int main() {
int x, y;
if (x == 0)
y = 0;
}
MD5 (first.c) = 38c2cf1f11c9523b429d1234538bb3e1
int main() {
int p, q;
if (p == 0)
q = 0;
}
MD5 (second.c) = e79fe2328e03d64bc7c4b7d36f274761
I think that those two snippets of code are similar. What do YOU think?
In this case I see some likeness to the case of Generic Drugs. Brazil forced lower prices of patented drugs by threatening with ignoring those patents and producing cheap, generic medecines.
They won because a state is still more powerful than any corporation. Imagine what would happen if SCO won the case against Linux, while Brazil would have most of the governmental IT run by Linux. Would the surrender to the power of SCO? I doubt it. So every such case is beneficial to the stability of Open Source community
You can clearly see on http://www.sco.com/scosource/unixtree/unixhistory0 1.html
an arrow drawn from Linux 2.2.16 kernel to the SCO UnixWare. Unfortunately you don't see any arrow going the other direction :).
Check out http://www.levenez.com/unix/history.html for a non obfuscated version of the graph.
Did you perform any benchmarks of ReiserFS/Linux vs. UFS/FreeBSD (on the same machine)? My personal experience shows that UFS/FreeBSD is twice faster than ReiserFS/Linux or XFS/Linux. Perhaps the fault is in Linux Kernel? - because obviously ReiserFS is performing very good when confronted to any filesystem on Linux platform. If it were the case - would you consider porting your excellent FS to *BSD?