Right. What Linus uses is he's own business. It's like the editor wars: my tool is better than yours...
BUT: Linux's main problem has always been the lack of a unified and consistent repository. Unlike the BSD projects, which used CVS from the very start, Linux kernel sources were kept in different SCM systems (at the beginning perhaps even not), so it is now extremely difficult to follow the change history reliably (think SCO law suits etc...).
For someone interested in developer's change history comments and details code changes, Linux kernel sources are a poor resource (despite ChangeLogs). You could always run diff between Linux releases, but you'll still don't have the commit logs. OTOH, the BSD CVS-REPO contains a wealth of useful comments, mistakes to avoid and how they have been resolved etc...
It is now too late for Linux, but nothing would prevent Linus from importing the current sources into CVS (or Subversion), and start from there! This BitKeeper lesson could prove beneficial, if Linus took the right decisions now.
If the fanless obsession were really more widespread, it would put pressure on CPU makers to invent better heat dissipation mechanisms, and to generate less heat in the first place!
I'm using FreeBSD 5.X on an old EPIA 5000 board with VIA VT6102 Rhine II chipset and the vr driver. Yes, I get sometimes vr0: rx packet lost messages on the console. This was only a problem for me while using NFS in UDP mode (I'm running diskless, so NFS reliability is pretty important!). Switching to TCP NFS solved all problems for me. The problem is not so bad as it sounds.
Hush are much too expensive, IMHO. My EPIA board with fanless 12 V DC-DC converter and brick transformer is absolutely silent, and costs less than 400 bucks, including 512 MB RAM. It's not the fastest platform out there, but as a non-gaming desktop box, it's perfectly usable.
A good industrial strength redily available qt is going to move alot of things.
I may burn karma here, but Java's main advantage over PyQt in the GUI arena is Qt's GPL or $$$$ licensing scheme. Most enterprise apps are not intended to be released under the GPL, so it's either Swing for free, or PyQT and a (for small enterprises rather expensive) commercial license from Trolltech.
Sure, for most companies, $$$$ licensing fees are peanuts, but explain to Mgmt to shell out bucks when you could code that same thing for free using Swing/Java? Yes, much better programmer productivity is the key here, but try to sell that to management first!
Yes, you can. Unless the data is encrypted and the reader needs to connect to some online database to get a part of the decryption key for that specific file.
Of course, this is not foolproof. All you need is the decrypt the data stream *once*, then recode the PDF again (cleaning it up from IP addresses and some such). What can't be cleaned up is the initial access to the keyserver that helped decrypt the file in the first place.
Here, life *is* fair:.com was created by the US first, so they own it. They are very generous to share it as a gTLD with the rest of the world. That's all.
This is just another case of first come, first served. We don't collectively complain that some great domain names have been taken. Why should we complain that some TLDs have been there from the start and belong to the US?
That law would only applies, if I commit a felony crime and use a domain registered with fake contact information in some way while committing the crime.
You can get in legal trouble much faster than you think: some unintentional copyright infrigement or someone suing you for libel is not so unlikely. I don't know if it's felony though... But if it is and you happen to be convicted, using a domain with bogus contact info would come back and bite you.
There is a world of difference between, say, the police serving your registrar, webhost or colocation service and some unstable troublemaker typing your domain into a whois interface and getting your name, phone number, address and email address with three seconds of work.
Oh yeah. That troublemaker would need to shell out $$$$ to a p.i. who in turn has good connections to the police.
data which in turn can contain 5. encoded information. Just this slightest possibility will make you 6. a terrorist
OMG! I have a lot of encoded information on my web site: ASCII- and JPEG-encoded! Now I'm in big trouble!
In the Netherlands there are plans to force ISP's to log and keep track of theur users
Software has bugs, servers get "misconfigured", backups fail... esp. regarding those pesky log files. A quick review of BOFH excuses should be required knowledge of every good ISP admin.
Yes, indeed. The whole registry infrastructure could be put up together from open source components that already exist. The servers could be secured and managed just like every other servers. There's nothing at all magical about it.
The real challenge for a registry is not technical. It is a major administrative and legal undertaking. One person was able to manage the whole.za domain from their basement, but.com and.net are a little bit larger and a tad more volatile.
I'm using an old 400Mhz machine here [...] But Firefox is so slow that it simply isn't an option.
Do you have enough RAM? That is generally the most limiting factor. I'm running Firefox on an EPIA 5000 (Eden) board with a fanless VIA C3 processor running at 500 MHz with 256 MB RAM. Firefox is pretty slick and fast on that platform.
I'd buy a SCO Linux license before I'd pay Verisign to register another domain.
Sorry to disappoint you, but if you own any.net or.com domain, you are already paying Verisign for it. Even if you registered with another registrar. All registrars (worldwide) pay a fee to the registry operator, which is for.com and.net Verisign. Verisign have a (apparently crappy) registrar business too, but that is not to be confused with the registry operation business.
DeNIC? Under which jurisdiction? If they operate under German laws, some US domains will have to be purged from the registry. If ICANN, a US company, is really going down this route, they'll be submerged with law suits seeking damages. Is this really something they would want to expose themselves to?
Perhaps that's because current competitors and bidders like, say, DeNIC and others are not really desirable from a technical, legal and political point of view?
Verisign is certainly not a good custodian for.net and.com (due to that Sitefinder debacle), but are other registry operators, at least the ones who are currently seeking to take the job, any better?
Liberty is either respected or it is not. The realpolitik guys will say "only when it's expedient". The libertarians will say, "all the time regardless". The Democrats and Republicans will say, "as long as you play nice", and that's the government we've got.
It's actually quite simple: as long as liberty (of speech) is deemed important by society, we'll have liberty. As soon as other "values" override this in society's eyes (like, say, security), liberty won't prevail, at least not on the short run.
Life is always a balancing act between liberty and security. Currently, the US are swinging away from liberty, but this is just a temporary set back. The time will come again, when individual civil liberties will be valued again. In the mean time, everyone has to ask him or herself, how to individually react to repression: to cave in, or to stand up. And, IMHO, we should respect their decision, no matter how much we agree or disagree.
This is not even an issue for the BSDs, where GNOME and KDE are all (meta-)ports. People who need them, can simply install them; those who don't, won't. Why is that a distro issue?
Nobody was hurt, I was protected (insured) so I did not lose anything
You may not have lost anything, but the community did. Do you seriously think that the insurance policies wouldn't get more expensive? Let that happen on a wide scale, and the whole insurance business would be broke.
You may live in a part of the world where only true and hardened criminals bear arms. You're certainly lucky for this. Unfortunately, we can't affort this kind of luxury in the US.
Slashdot Mirror
No one is sending astronauts to their death, this article is looking for a flame war.
A flame war? With all this H2 in the tank, that would be catastrophic!
Who bloody cares?
Right. What Linus uses is he's own business. It's like the editor wars: my tool is better than yours...
BUT: Linux's main problem has always been the lack of a unified and consistent repository. Unlike the BSD projects, which used CVS from the very start, Linux kernel sources were kept in different SCM systems (at the beginning perhaps even not), so it is now extremely difficult to follow the change history reliably (think SCO law suits etc...).
For someone interested in developer's change history comments and details code changes, Linux kernel sources are a poor resource (despite ChangeLogs). You could always run diff between Linux releases, but you'll still don't have the commit logs. OTOH, the BSD CVS-REPO contains a wealth of useful comments, mistakes to avoid and how they have been resolved etc...
It is now too late for Linux, but nothing would prevent Linus from importing the current sources into CVS (or Subversion), and start from there! This BitKeeper lesson could prove beneficial, if Linus took the right decisions now.
Perforce is also used by the FreeBSD project for highly experimental stuff that still has to go into the main CVS repo.
If the fanless obsession were really more widespread, it would put pressure on CPU makers to invent better heat dissipation mechanisms, and to generate less heat in the first place!
I'm using FreeBSD 5.X on an old EPIA 5000 board with VIA VT6102 Rhine II chipset and the vr driver. Yes, I get sometimes vr0: rx packet lost messages on the console. This was only a problem for me while using NFS in UDP mode (I'm running diskless, so NFS reliability is pretty important!). Switching to TCP NFS solved all problems for me. The problem is not so bad as it sounds.
Hush are much too expensive, IMHO. My EPIA board with fanless 12 V DC-DC converter and brick transformer is absolutely silent, and costs less than 400 bucks, including 512 MB RAM. It's not the fastest platform out there, but as a non-gaming desktop box, it's perfectly usable.
A good industrial strength redily available qt is going to move alot of things.
I may burn karma here, but Java's main advantage over PyQt in the GUI arena is Qt's GPL or $$$$ licensing scheme. Most enterprise apps are not intended to be released under the GPL, so it's either Swing for free, or PyQT and a (for small enterprises rather expensive) commercial license from Trolltech.
Sure, for most companies, $$$$ licensing fees are peanuts, but explain to Mgmt to shell out bucks when you could code that same thing for free using Swing/Java? Yes, much better programmer productivity is the key here, but try to sell that to management first!
PDF = Privacy Depleted Format
Yes, you can. Unless the data is encrypted and the reader needs to connect to some online database to get a part of the decryption key for that specific file.
Of course, this is not foolproof. All you need is the decrypt the data stream *once*, then recode the PDF again (cleaning it up from IP addresses and some such). What can't be cleaned up is the initial access to the keyserver that helped decrypt the file in the first place.
Life isn't fair.
Here, life *is* fair: .com was created by the US first, so they own it. They are very generous to share it as a gTLD with the rest of the world. That's all.
This is just another case of first come, first served. We don't collectively complain that some great domain names have been taken. Why should we complain that some TLDs have been there from the start and belong to the US?
That law would only applies, if I commit a felony crime and use a domain registered with fake contact information in some way while committing the crime.
You can get in legal trouble much faster than you think: some unintentional copyright infrigement or someone suing you for libel is not so unlikely. I don't know if it's felony though... But if it is and you happen to be convicted, using a domain with bogus contact info would come back and bite you.
There is a world of difference between, say, the police serving your registrar, webhost or colocation service and some unstable troublemaker typing your domain into a whois interface and getting your name, phone number, address and email address with three seconds of work.
Oh yeah. That troublemaker would need to shell out $$$$ to a p.i. who in turn has good connections to the police.
data which in turn can contain 5. encoded information. Just this slightest possibility will make you 6. a terrorist
OMG! I have a lot of encoded information on my web site: ASCII- and JPEG-encoded! Now I'm in big trouble!
In the Netherlands there are plans to force ISP's to log and keep track of theur users
Software has bugs, servers get "misconfigured", backups fail... esp. regarding those pesky log files. A quick review of BOFH excuses should be required knowledge of every good ISP admin.
GoDaddy already sent out a mass mailing to all its customers about this.
Running the DNS isn't rocket science
Yes, indeed. The whole registry infrastructure could be put up together from open source components that already exist. The servers could be secured and managed just like every other servers. There's nothing at all magical about it.
The real challenge for a registry is not technical. It is a major administrative and legal undertaking. One person was able to manage the whole .za domain from their basement, but .com and .net are a little bit larger and a tad more volatile.
I'm using an old 400Mhz machine here [...] But Firefox is so slow that it simply isn't an option.
Do you have enough RAM? That is generally the most limiting factor. I'm running Firefox on an EPIA 5000 (Eden) board with a fanless VIA C3 processor running at 500 MHz with 256 MB RAM. Firefox is pretty slick and fast on that platform.
I'd buy a SCO Linux license before I'd pay Verisign to register another domain.
Sorry to disappoint you, but if you own any .net or .com domain, you are already paying Verisign for it. Even if you registered with another registrar. All registrars (worldwide) pay a fee to the registry operator, which is for .com and .net Verisign. Verisign have a (apparently crappy) registrar business too, but that is not to be confused with the registry operation business.
DeNIC? Under which jurisdiction? If they operate under German laws, some US domains will have to be purged from the registry. If ICANN, a US company, is really going down this route, they'll be submerged with law suits seeking damages. Is this really something they would want to expose themselves to?
Perhaps that's because current competitors and bidders like, say, DeNIC and others are not really desirable from a technical, legal and political point of view?
Verisign is certainly not a good custodian for .net and .com (due to that Sitefinder debacle), but are other registry operators, at least the ones who are currently seeking to take the job, any better?
Liberty is either respected or it is not. The realpolitik guys will say "only when it's expedient". The libertarians will say, "all the time regardless". The Democrats and Republicans will say, "as long as you play nice", and that's the government we've got.
It's actually quite simple: as long as liberty (of speech) is deemed important by society, we'll have liberty. As soon as other "values" override this in society's eyes (like, say, security), liberty won't prevail, at least not on the short run.
Life is always a balancing act between liberty and security. Currently, the US are swinging away from liberty, but this is just a temporary set back. The time will come again, when individual civil liberties will be valued again. In the mean time, everyone has to ask him or herself, how to individually react to repression: to cave in, or to stand up. And, IMHO, we should respect their decision, no matter how much we agree or disagree.
The expiration date should be public
What about saying that you have received a gag order? That's a pretty important information too, no matter how long the original gag order lasts.
But, sir, I do save all my logs to /dev/null. You can take that device with you and analyze its contents.
This is not even an issue for the BSDs, where GNOME and KDE are all (meta-)ports. People who need them, can simply install them; those who don't, won't. Why is that a distro issue?
I now use MBNA for all of my online shopping
This is an american bank, right? Does anybody know of similar services in Europe?
Nobody was hurt, I was protected (insured) so I did not lose anything
You may not have lost anything, but the community did. Do you seriously think that the insurance policies wouldn't get more expensive? Let that happen on a wide scale, and the whole insurance business would be broke.
You may live in a part of the world where only true and hardened criminals bear arms. You're certainly lucky for this. Unfortunately, we can't affort this kind of luxury in the US.