And this is the link to the homepage about it. Note the e-mail address. Please all do as I did. Go to the heise article, and forward it to FINANCE.Stuttgart@cmg.de with some comments to make them clear that they'd better back off.
Btw as an ex-CMG (holland) employee and shareholder I am very dissapointed about this. The dutch CMG branch always tries to play nice with UNIX/Linux enthousiasts, to get a good image for recruitment.
Imposing tariffs, and demanding that governmental organisations use alternative software, will mean that other companies and OSses get a real and fair chance do develop, and become strong competition for MSFT.
Indeed. They (US corps) can bribe US politicians and judges, and corrupt that system. But luckily the world is larger than the US. They will find out that not every part of the world is so corrupt and loose.
To spare Microsoft because they contribute to the local economy is short sighted. But alas many voters are short sighted too.
Assuming the anti-trust case is just, then Microsoft is found to be guilty of damaging competitors and thus the industry in general.
Actions againt MSFT might cause massive layoffs, but no actions means many, many small layoffs from smaller companies that are destroyed by the predator. Just these are less visible.
B.t.w, using my monitoring script, I noticed that the percentage went down from 5% to 2.5% after 5pm (in Europe, i.e. as the Americans awoke).
This shows that, when more load is coming, the number of requests answered by the NT machines goes down relatively. In other words, they cannot handle the same load as the FreeBSD machines (taking longer to process a request).
Being bored I wrote a script to monitor this ratio continuously. The example below monitors the average over the last 500 request (one request max. per 10s). After some hours of running this I still get (like others) around 5%. I'll track this some days to see if anything changes.
Hmm, if lots of people are going to do this (removing the sleep 10) that might really cause even Hotmail to be slashdotted:).
#!/bin/ksh let i=0 let ms=0 let ap=0 while true; do if [[ ${v[i]} = "ms" ]]; then let ms=ms-1 fi if [[ ${v[i]} = "ap" ]]; then let ap=ap-1 fi server=$(lynx -head -dump http://www.hotmail.com/ |grep Server) if (print $server|grep Microsoft >/dev/null); then v[i]="ms" let ms=ms+1 else v[i]="ap" let ap=ap+1 fi print $ms out of $((ms+ap)), $(print "2k 100 $ms * $ms $ap + / p" | dc) % let i=i+1 if [[ i -eq 500 ]]; then let i=0 fi sleep 10 done
Re:Why doesn't Warner Brothers sue the MPAA?
on
NYT On DeCSS Case
·
· Score: 4
No matter how hard you try, you can't prevent people from making perfect copies of digital data.
In the NYT article, the head of the United States Copyright Office, Ms. Peters, says it is a problem that copyright holder won't have an incentive to produce digital material without assurance of protection.
If I could advice her, I'd tell that those copyright holder then just should refrain from producing digital material.
If the choice would really be between ending fair use and harming freedom of speech, or not producing digital material, the choice is easy: the latter.
No company is forced to produce their material digitally. If the studios are so worried, then they should just stick to VHS tapes.
Of course, presented with such a choice, they would have to reveal their true motives: They would not want to stick to VHS, since what they really want is to enlarge their influence and control.
How can the UNIX offer be 5 times more expensive?!?
I work in a large bank, on a 80GB datawarehouse (mirrored, so 160GB diskspace). An internal competitor uses NT (Compaq) with SQL server for a similar (but smaller) application, we use Solaris/Sparc with Oracle.
Constantly we are being judged on cost/performance but others. Recent comparisons showed that an Intel solution (Compaq) would be 30% cheaper. In return you get CPU's with smaller cache, generally less reliability and it is questionable if our app could run at all on Compaq.
Note that the 30% difference only accounts for the hardware cost Sun/Sparc vs. Compaq/Intel.
Taking into account the OS cost (NT versus Solaris) it is sure that NT would become much more expensive, since Solaris is included with the hardware, and NT licenses for such large applications are extremely expensive. Not to mention the extra system administration costs that NT would cause.
As for Linux/Intel? I would not do it. As mentioned you can gain maybe 30% on HW cost, but for that you can be sure that Linux cannot handle load and scale like Solaris/Sparc can.
Like most european countries. Then you see (and your phone equipment too) the source telephone number.
One could program ones equipment to drop any calls from sources that either don't send CLID or that are on a blacklist (or not on a friends-list).
In fact I never answer the phone when I don't see the source phone number (people can block the CLID if they want to; in that case I don't need to talk to them).
It doesn't matter if the OS is complex and only for relatively few to work with it. To normal users the only thing that matters are the end-user applications.
People learn how to drive a car (end-use app) but not how to fix the engine, in fact with all modern electronics etc that becomes more and more difficult without special equipment etc.
Likewise, a solid and complex-to-use (for the average end-user) OS is fine, as long as it is a good development environment for the specialists, and as long as there are easy and good end-user applications.
This may apply to some, but certainly not to the majority of FreeBSD users. Many of them use WinNT on their desktops for example, which completely counters your opinion that many of them run it only to be excusive.
As for myself, I run it because *for me* it is easier to maintain and keep up-to-date (thanks to cvsup, make world and the ports system) than any Linux variant. Also I want absolute control and don't want any GUI install stuff and such (like RH; in that respect indeed I like Slackware as best from the Linux distro's).
If I really need easy-to-use clicking without knowing what's going on, just to run some toys, wordprocessing and games, I'll just use Windows. For me, Linux is kind of in-between, offering the worst of both worlds.
From the document it becomes clear that they didn't enable softupdates in FFS (BSD). Hence the remark of slow BSD performance with lots of small files. This is a basic thing that anyone does immediately when using FreeBSD. Luckily the licence restrictions on softupdates have been lifted some weeks ago, and from now on it is enabled by default.
Also, in the webserver performance section, they didn't increase the number of possible simultaneous connections. Yes, FreeBSD has lower defaults for such things and one has to modify some kernel-config files to increase that. Apparently they didn't now about this and thus concluded wrongly that BSD scales less.
As ftp.cdrom.com shows (allowing not 500 but 5000 simultaneous connections) scaling further is no problem at all for FreeBSD. Configuring its kernel is just less user-friendly and obvious than Linux. But when performing and publishing benchmarks, one should know what one is doing and know exactly how to tune both systems.
I tried this too (after BSD scored better in bonnie), things like creating 50000 empty files in a directory, removing them, mass copies (using cpio -dump) of large directory trees with lots of small files. This kind of things tests how good the filesystem is at modifying metadata.
FreeBSD is faster with such things than anything else I've seen up till now (including almost any kind of UNIX variant inclusive Linux with all its newer filesystems). This was FFS with softupdates of course. That makes a huge difference indeed.
See my other post: after years of benchmarking filesystems under FreeBSD and Linux (using all kinds of tuning possible) let me assure you that BSD is consistently and clearly faster wrt filesystem IO than any Linux up to date (including ext3fs, reiserfs, kernel 2.4-pre etc).
Both 9MB/s and 4MB/s is very low thow. My current (IDE) system gets bonnie score of 15MB/s in BSD, and 12MB/s in Linux.
I've been testing bonnie and such on FreeBSD and various Linux distro's for years (I'm a benchmark freak) and FreeBSD has always, on every piece of hardware I had (both IDE and SCSI) easily outperformed any Linux (including the latest 2.4 prereleases, reiserfs, ext3fs etc). Of course this is FreeBSD with softupdates compiled in or with asynchronous IO enabled.
One of his main criticisms is that most (large) UNIX apps share little code. Obviously that is not very efficient, but it also has some advantages.
It means that these apps are more self contained and not vulnerable from outside influences (remember the DLL-hell from Win3/Win9x/WinNT ?). If they all use the same shared libraries, new versions (which may be required by some apps) may break lots of other apps.
This can only be avoided by very strong central control. Apple is a good example: Apple has always controlled which apps may be released for it's platform, they must conform to all kinds of standards, both from a technical and from a usability point of view.
This is nice for stability (in theory) and some end-users, but it poses severe restrictions on engineering freedom and using/developing for the platform for fun. Also it is impossible to reach in an Open Source environment IMO.
Re:my fruitloop ex had a mac,..
on
MacOSX and X11
·
· Score: 1
Maybe it would be a good idea if the patent office weren't only paid according to number of patents approved, but would also have to pay (a lot) for each approved patent that is later thrown out in court. That might make them a bit more cautious in approving stupid patents.
At least, they should pay all legal costs involving such lawsuits.
Because it is greek. Also I'm sure for Spanish "tekh" is easier than tecks (they would mix it up with "tesk" all the time). For dutch (which I am) tekh sound easier too.
Btw in many languages (I don't know about english) gnome (including the g pronounced) is a normal word, not hard to pronounce at all. English has the tendency to simply drop "difficult" sounds in words (like the silent g at the beginning).
Commercial software is for 99% american. Public domain/GNU software much less. More "foreign" influence results in words that maybe stranger or more diffucult for americans, but not necessarily for others.
You've not been following SUN's past. Sun created things that nowadays are UNIX standards, such as NFS, NIS (yellow pages), Openwindows though that failed and Motif won alas, just to name a few.
They have always worked with open API's. Unlike most other companies that kept the specifications closed. They still do.
They want control over their own products, and especially don't want to hand it over to ISO/ANSI etc since when that happens, you can wait for 10 years for the next version of the spec (i.e. it is dead. just look at how C++ evolves since AT&T handed it over to ISO).
They don't want control to monopolize the industry, but because they feel that is the best way to create open standards, while preserving the possibility for further and speedy evolution and improvements.
IBM's JVM is not better than Sun's. It may have been faster at times (not compared to the new 1.3 with hotspot) but contained bugs, deviations from the Java standard.
As for Sun not submitting Java to ISO: very good. Look at how C++ and other standards are dragging on. Once one of these international bureaucracies gets their hands on it, you can forget further evolution in a reasonable pace (and Java still needs evolution). Sun, with it's JCP, found the ideal compromise between making an open standard without fully loosing control, thus still being able to press for speedy improvements.
In fact that is one of the big advantages often touted towards MSFT: they can have fast evolution, simply by doing things themselves without having to go through lengthy ISO processes etc. The drawback is that it's no longer adhering to standards. JCP is the right compromise.
Maybe in 10 years time, when Java is 'finished' it is time to hand it over to ISO.
In the meantime, SUN is still investing a lot in Java and it's evolution. Why should big and rich companies like IBM not pay for the brand in order to support this? At least they have a real say in the JCP, not something that could be said of MSFT API's and products (for which they also would have to pay licensing fees).
I'm also very suspicous on IBM's commitment to Open Source. Why al this attention for Linux? What is behind it? Maybe they fear other open source products (like *BSD) that might really threaten their own products (AIX, Monterey) and try to kill those off by single sided attention for the weaker contestant.
Slashdot Mirror
Who says the CMG SAMBA was pre-existing? I doubt that very much.
What do you expect: should OSS projects in the future register their names in every country of the world? Who should pay for that?
Or should OSS projects, because of lack of funds, accept that they have to change their name every once in a while...
And this is the link to the homepage about it. Note the e-mail address. Please all do as I did. Go to the heise article, and forward it to FINANCE.Stuttgart@cmg.de with some comments to make them clear that they'd better back off.
Btw as an ex-CMG (holland) employee and shareholder I am very dissapointed about this. The dutch CMG branch always tries to play nice with UNIX/Linux enthousiasts, to get a good image for recruitment.
Imposing tariffs, and demanding that governmental organisations use alternative software, will mean that other companies and OSses get a real and fair chance do develop, and become strong competition for MSFT.
Indeed. They (US corps) can bribe US politicians and judges, and corrupt that system. But luckily the world is larger than the US. They will find out that not every part of the world is so corrupt and loose.
To spare Microsoft because they contribute to the local economy is short sighted. But alas many voters are short sighted too.
Assuming the anti-trust case is just, then Microsoft is found to be guilty of damaging competitors and thus the industry in general.
Actions againt MSFT might cause massive layoffs, but no actions means many, many small layoffs from smaller companies that are destroyed by the predator. Just these are less visible.
B.t.w, using my monitoring script, I noticed that the percentage went down from 5% to 2.5% after 5pm (in Europe, i.e. as the Americans awoke).
This shows that, when more load is coming, the number of requests answered by the NT machines goes down relatively. In other words, they cannot handle the same load as the FreeBSD machines (taking longer to process a request).
Being bored I wrote a script to monitor this ratio continuously. The example below monitors the average over the last 500 request (one request max. per 10s). After some hours of running this I still get (like others) around 5%. I'll track this some days to see if anything changes.
:).
Hmm, if lots of people are going to do this (removing the sleep 10) that might really cause even Hotmail to be slashdotted
#!/bin/ksh
let i=0
let ms=0
let ap=0
while true;
do
if [[ ${v[i]} = "ms" ]]; then
let ms=ms-1
fi
if [[ ${v[i]} = "ap" ]]; then
let ap=ap-1
fi
server=$(lynx -head -dump http://www.hotmail.com/ |grep Server)
if (print $server|grep Microsoft >/dev/null); then
v[i]="ms"
let ms=ms+1
else
v[i]="ap"
let ap=ap+1
fi
print $ms out of $((ms+ap)), $(print "2k 100 $ms * $ms $ap + / p" | dc) %
let i=i+1
if [[ i -eq 500 ]]; then
let i=0
fi
sleep 10
done
No matter how hard you try, you can't prevent people from making perfect copies of digital data.
In the NYT article, the head of the United States Copyright Office, Ms. Peters, says it is a problem that copyright holder won't have an incentive to produce digital material without assurance of protection.
If I could advice her, I'd tell that those copyright holder then just should refrain from producing digital material.
If the choice would really be between ending fair use and harming freedom of speech, or not producing digital material, the choice is easy: the latter.
No company is forced to produce their material digitally. If the studios are so worried, then they should just stick to VHS tapes.
Of course, presented with such a choice, they would have to reveal their true motives: They would not want to stick to VHS, since what they really want is to enlarge their influence and control.
How can the UNIX offer be 5 times more expensive?!?
I work in a large bank, on a 80GB datawarehouse (mirrored, so 160GB diskspace). An internal competitor uses NT (Compaq) with SQL server for a similar (but smaller) application, we use Solaris/Sparc with Oracle.
Constantly we are being judged on cost/performance but others. Recent comparisons showed that an Intel solution (Compaq) would be 30% cheaper. In return you get CPU's with smaller cache, generally less reliability and it is questionable if our app could run at all on Compaq.
Note that the 30% difference only accounts for the hardware cost Sun/Sparc vs. Compaq/Intel.
Taking into account the OS cost (NT versus Solaris) it is sure that NT would become much more expensive, since Solaris is included with the hardware, and NT licenses for such large applications are extremely expensive. Not to mention the extra system administration costs that NT would cause.
As for Linux/Intel? I would not do it. As mentioned you can gain maybe 30% on HW cost, but for that you can be sure that Linux cannot handle load and scale like Solaris/Sparc can.
Like most european countries. Then you see (and your phone equipment too) the source telephone number.
One could program ones equipment to drop any calls from sources that either don't send CLID or that are on a blacklist (or not on a friends-list).
In fact I never answer the phone when I don't see the source phone number (people can block the CLID if they want to; in that case I don't need to talk to them).
In line with this analogy:
It doesn't matter if the OS is complex and only for relatively few to work with it. To normal users the only thing that matters are the end-user applications.
People learn how to drive a car (end-use app) but not how to fix the engine, in fact with all modern electronics etc that becomes more and more difficult without special equipment etc.
Likewise, a solid and complex-to-use (for the average end-user) OS is fine, as long as it is a good development environment for the specialists, and as long as there are easy and good end-user applications.
Many Linux dists ship libc*.so with debugging symbols in them. That doesn't affect the in-core memory requirements.
It is irritating however, when you have to upgrade to the latest libc RPM you have to download huge files.
This may apply to some, but certainly not to the majority of FreeBSD users. Many of them use WinNT on their desktops for example, which completely counters your opinion that many of them run it only to be excusive.
As for myself, I run it because *for me* it is easier to maintain and keep up-to-date (thanks to cvsup, make world and the ports system) than any Linux variant. Also I want absolute control and don't want any GUI install stuff and such (like RH; in that respect indeed I like Slackware as best from the Linux distro's).
If I really need easy-to-use clicking without knowing what's going on, just to run some toys, wordprocessing and games, I'll just use Windows. For me, Linux is kind of in-between, offering the worst of both worlds.
From the document it becomes clear that they didn't enable softupdates in FFS (BSD). Hence the remark of slow BSD performance with lots of small files. This is a basic thing that anyone does immediately when using FreeBSD. Luckily the licence restrictions on softupdates have been lifted some weeks ago, and from now on it is enabled by default.
Also, in the webserver performance section, they didn't increase the number of possible simultaneous connections. Yes, FreeBSD has lower defaults for such things and one has to modify some kernel-config files to increase that. Apparently they didn't now about this and thus concluded wrongly that BSD scales less.
As ftp.cdrom.com shows (allowing not 500 but 5000 simultaneous connections) scaling further is no problem at all for FreeBSD. Configuring its kernel is just less user-friendly and obvious than Linux. But when performing and publishing benchmarks, one should know what one is doing and know exactly how to tune both systems.
I tried this too (after BSD scored better in bonnie), things like creating 50000 empty files in a directory, removing them, mass copies (using cpio -dump) of large directory trees with lots of small files. This kind of things tests how good the filesystem is at modifying metadata.
FreeBSD is faster with such things than anything else I've seen up till now (including almost any kind of UNIX variant inclusive Linux with all its newer filesystems). This was FFS with softupdates of course. That makes a huge difference indeed.
See my other post: after years of benchmarking filesystems under FreeBSD and Linux (using all kinds of tuning possible) let me assure you that BSD is consistently and clearly faster wrt filesystem IO than any Linux up to date (including ext3fs, reiserfs, kernel 2.4-pre etc).
Both 9MB/s and 4MB/s is very low thow. My current (IDE) system gets bonnie score of 15MB/s in BSD, and 12MB/s in Linux.
I've been testing bonnie and such on FreeBSD and various Linux distro's for years (I'm a benchmark freak) and FreeBSD has always, on every piece of hardware I had (both IDE and SCSI) easily outperformed any Linux (including the latest 2.4 prereleases, reiserfs, ext3fs etc). Of course this is FreeBSD with softupdates compiled in or with asynchronous IO enabled.
Usually the difference was much more than 10%.
One of his main criticisms is that most (large) UNIX apps share little code. Obviously that is not very efficient, but it also has some advantages.
It means that these apps are more self contained and not vulnerable from outside influences (remember the DLL-hell from Win3/Win9x/WinNT ?). If they all use the same shared libraries, new versions (which may be required by some apps) may break lots of other apps.
This can only be avoided by very strong central control. Apple is a good example: Apple has always controlled which apps may be released for it's platform, they must conform to all kinds of standards, both from a technical and from a usability point of view.
This is nice for stability (in theory) and some end-users, but it poses severe restrictions on engineering freedom and using/developing for the platform for fun. Also it is impossible to reach in an Open Source environment IMO.
Buy a good color-postscript laserprinter.
Since that means that this increasing amount of web space can't be seen by an increasing amount of users.
This will help webpage authors to avoid IE's proprietary stuff. If they don't they will get an increasing amount of complaints.
Maybe it would be a good idea if the patent office weren't only paid according to number of patents approved, but would also have to pay (a lot) for each approved patent that is later thrown out in court. That might make them a bit more cautious in approving stupid patents.
At least, they should pay all legal costs involving such lawsuits.
Because it is greek. Also I'm sure for Spanish "tekh" is easier than tecks (they would mix it up with "tesk" all the time).
For dutch (which I am) tekh sound easier too.
Btw in many languages (I don't know about english) gnome (including the g pronounced) is a normal word, not hard to pronounce at all. English has the tendency to simply drop "difficult" sounds in words (like the silent g at the beginning).
Commercial software is for 99% american. Public domain/GNU software much less. More "foreign" influence results in words that maybe stranger or more diffucult for americans, but not necessarily for others.
If everyone just used postscript, all rendering would be the same and there would be no such problems.
You've not been following SUN's past. Sun created things that nowadays are UNIX standards, such as NFS, NIS (yellow pages), Openwindows though that failed and Motif won alas, just to name a few.
They have always worked with open API's. Unlike most other companies that kept the specifications closed. They still do.
They want control over their own products, and especially don't want to hand it over to ISO/ANSI etc since when that happens, you can wait for 10 years for the next version of the spec (i.e. it is dead. just look at how C++ evolves since AT&T handed it over to ISO).
They don't want control to monopolize the industry, but because they feel that is the best way to create open standards, while preserving the possibility for further and speedy evolution and improvements.
IBM's JVM is not better than Sun's. It may have been faster at times (not compared to the new 1.3 with hotspot) but contained bugs, deviations from the Java standard.
As for Sun not submitting Java to ISO: very good. Look at how C++ and other standards are dragging on. Once one of these international bureaucracies gets their hands on it, you can forget further evolution in a reasonable pace (and Java still needs evolution). Sun, with it's JCP, found the ideal compromise between making an open standard without fully loosing control, thus still being able to press for speedy improvements.
In fact that is one of the big advantages often touted towards MSFT: they can have fast evolution, simply by doing things themselves without having to go through lengthy ISO processes etc. The drawback is that it's no longer adhering to standards. JCP is the right compromise.
Maybe in 10 years time, when Java is 'finished' it is time to hand it over to ISO.
In the meantime, SUN is still investing a lot in Java and it's evolution. Why should big and rich companies like IBM not pay for the brand in order to support this? At least they have a real say in the JCP, not something that could be said of MSFT API's and products (for which they also would have to pay licensing fees).
I'm also very suspicous on IBM's commitment to Open Source. Why al this attention for Linux? What is behind it? Maybe they fear other open source products (like *BSD) that might really threaten their own products (AIX, Monterey) and try to kill those off by single sided attention for the weaker contestant.