The ports system rocks! For those of us with fast connections it's far better than RPM. No problems with missing libraries and no hassle.
Even more so for those with slow connections. For years I've used FreeBSD with only a 28k8 modem. The ports system is much much more bandwidth efficient. Example:
RPM: new RPM with some fix. Old RPM based on emacs-20.5, new RPM too. You have to download the whole binary, complete package again. That is outrageious IMO
Port: New fix: Update your port (basically a directory with some files, among which are patch files, in it) usually done by an rsync like mechanism (cvsup) ie the update typically is something like 1kb. Then rebuild the port (the original emacs-20.5.tar.tz source is still lying around in/usr/ports/distfiles from the first time you got this port).
So for refreshes of a port you typically download 1KB, for RPM you have to reget the whole binary again.
One of the strong points of FreeBSD. I've been upgrading my system for over 5 years, without a single reinstall. Still everything is totally clean and organized.
cvsup, together with make world is great for the base system (which is cleanly separated from add-on packages that all reside in/usr/local; alas Linux makes a mess by mixing a core system with all add-on packages).
then the ports system may requiring some uninstalling/reinstalling sometimes (it is a bit more primitive in this respect than most Linux distro's) but that isn't such a big deal. The important thing is that the base system itself is 100% upgradeable and never requires a reinstall. Not even when moving to a new machine or upgrading hardware.
Today I use Linux (mainly because of availability of vmware and good JDK's) but in this respect I long back for FreeBSD.
Re:Obselence -- Something to fear?
on
Too Old To Code?
·
· Score: 1
Therefore it is important to switch jobs *often* to ensure being exposed to new things all the time. This keeps you flexible and used to learning new things fast. Once you've done the same methodology/language for more than 5 years, you're lost and it becomes real hard to adopt to new views/methods (sorry for you:)
I'm 34 now, and since my masters degree 10 years ago I've had 10 jobs. Since I switched so often I decided to become self employed 2 years ago. My last job for example was my first exposure to Oracle, at 33 I started with RDBMS in general.
As you get older, it becomes harder to do something new because companies either want someone young (and cheap) or someone with 100% perfect experience.
But, if you got lot of different experience during your 20s (when that is still possible) there are so many things you can do that in most cases you can match the requested experience for 90%. The remaining 10% are new, which is a new change to get professional experience on a new subject.
You can also get experience on new things by hobbying at home, but I have found that: - companies don't count hobby experience (often not fair, but still) - as you get older, it gets harder and harder to motivate yourself to really invest time in new things for your hobby at home
It's not just the money. Two other reasons I can think of:
- expensive memory makes their CPU's look cheap in comparison - slower memory might let people think they need faster (more expensive) CPU's with more built-in cache
When I started (14) I wanted to know top to bottom how a computer worked. With the commonodore pet, and, a short while later the commodore 64, this was possible: a simple command-line oriented environment without many layers of GUI's, wizards, automated things going on behind your back, i.e. you can really know what your doing.
Fascination started really when I was 15 and got a C compiler for the C64 (dog slow, but still). I'd still recommend it today for a beginner (on a faster platform though). C is small and simple, and a bare gcc environment offers much more oversight than the confusing visual environment.
Why should it be different for kids today?
Yes, using visual c++ you can create complex programs faster (at start) but without knowing what exactly is going on. VB is even worse in this respect. It may help them to create a program, but doesn't necessarily teach them how to program.
Yes, I live in Switzerland and of newly sold TV's at least 50% is widescreen. More and more regular TV programs are transmitted in widescreen (throughout Europe). Within one or two years I think 4:3 TV is history (like black&white).
Poor french. Even in Switzerland with its outrageously expensive Swisscom, you can call for the equiv of $0.05/minute to the US now. Of course by using an alternative carrier via a prefix.
They might try to control ISP's (as Germany has also done) and require them to block certain sites.
But that won't really help either, since international phonecalls have become so cheap (less than $0.05/minute) that those who really need to access some blocked materials can easily dial in into a US ISP.
Little calculation: assume the cache is 256kb, randomly written by one in 5 instructions, at a pace of 200MHZ (how fast is cache these days?). Cache width 4 bytes at a time. I guess these are conservative estimates:
10*10e12 / 200*10e6 * 5 * 256*1024/4 = 16384000000 seconds it would last,
ah, that's 532 years. OK, no problem:)
OTOH if it's only as fast as DRAM, it isn't good enough for cache. invalidating the cache (thus refilling it) after a power-on doesn't really seem a problem to me.
Why not do it both, to ensure success either way: Split in 3 x 3, thus 9 pieces. Three of them may only do the OS, three only apps, three only Internet.
If only it were true what you say, that talent is rewarded instead. Well, in a way it is: not the talent for inventing something briliant, but mostly the talent to play dirty tricks, bribe, monopolize and blackmail.
In that way, unchecked capitalism is evil indeed. It is not the most talented (from a scientific point of view) that often win, also it is often not the most technically sound standard that survives, but it is the cheaters, thieves and frauds.
Copyright law should protect good inventions, so that people get rewarded for their work. Instead it mostly protects vested interests, (near) monopolies and coroporations that can bully others and seemingly do what they want.
The IA64 is not a breakthrough chips. There exist numerous other 64-bit risc chips already, with comparable performance (ultrasparc, soon new version, alpha, power). Also the IA64 is not fully backwards compatible with IA32 (finally), that is current IA32 binaries may run on it but at strongly reduced speed AFAIK (for those, upgrading to IA64 wouldn't make sense).
So I don't see what the fuzz w.r.t. the IA64 is all about. At the time it will actually hit the market, it is already outdated, and for example Linux will have faster (sparc/alpha) platforms it can run on.
this may be the key to its undoing. We may thus look on software piracy as a moral imperative
Amen to that. Normally I'm a law abiding citizen, but this law and what the industry is trying to do with it makes me so mad, that for the first time in my life I feel obliged to 'civil disobedience'. Today I feel guilty when I buy a DVD and/or CD (money is not the problem, I have enough of it) and only feel right if I make illegal copies. Buying them (i.e. not stealing) would make me feel like collaborating with the Germans in WW2.
I hope it creates these feelings in so many people that it becomes like a guerilla war or resistance movement that the industry/government simply cannot win.
The survivor in a proprietary world of closed "standards" indeed is not necessarily the best. In the examples you mention, the loosing of the best has to do with installed base, and incompatability of the better one, which still looses because of the larger installed base. Users are locked into one alternative.
OTOH with open standards, open source and interoperability, things are different. Here there is no locking out of the better alternative, since the two (better and worse) alternative can work together and noone is forced to use the worse alternative for the sake of compatability. Even if one of them implements a different API/standard/protocol, the other one can look at it and copy it.
Alas, w.r.t. Linux there is more and more non-open source (commercial apps). Still, since Linux itself is open, FreeBSD can build a almost 100% perfect Linux emulator (unlike Wine, which cannot work very well since Windows' source is not available).
Note that I do not claim that BSD will win in the end, since I don't claim that it is better. It's internals may be better, but in other respects (support for crappy hardware, easy-to-use packaging) Linux may be better.
But I am convinced that in the end the best one (in perception of end users) will win, and both will evolve to improve further in a darwinian 'struggle for life'.
There is nothing wrong with Xt. It's a nice framework, though a bit cumbersome because of it's OO structure implemented without OO language support.
But Motif OTOH, is a different thing. It uses Xt but in quite an ugly way. OpenLook, which lost from Motif alas, looks better, and also had a much nicer API. Btw there were two implementations from OpenLook: xview with it's own built-in toolkit (which was easier to program), and Olit, OpenLook on Xt, which was also quite nice.
Sun lost the X-window GUI war because Motif looked more like Windows3. Blergh. If they would have won, Unix would have had a single GUI for a long time already.
B.t.w. when Sun realized OpenLook was dead, they also released the sourcecode. But binaries had been available for free for a longer time already.
Yes, they try to sell some more of their X-designer, thinking that giving away Motif will help it to regain some ground.
I think it won't, for reasons others have mentioned. It's nice to be able to run some legacy Motif apps with dynamic libraries ISO the huge statically compiled ones, but that it the only use I think. Mainly Netscape, soon to be replaced by Mozilla which uses Gtk.
I must say, I have always been extremely irritated by Motif. Apart from it's ugly look & feel, ugly API (compared to the sample Athena widget set), it seemed outrageous to me that one could get a whole UNIX(like) operating system for free (Linux/*BSD) but had to pay for just a GUI toolkit, even for the runtime alone, so much money that whole 'gratis' advantage of Linux/*BSD would be annihilated by it.
Even MSFT acknowledged that a serious OS needs serious command line and batching facilities.
No, Linux's/UNIX's command line will never die. True, the 'normal' user won't see or use it, but the poweruser/system admin needs it to automate tasks, to work faster and more effective.
Yes it takes more time to learn, but once you've learned it, it is simply faster and more convenient than a GUI.
Sadly, most peoples idea of so called archaic CLI interfaces have been formed by the rotten DOS/Windows CLI, which doesn't compare to those of UNIX, VMS, OS/390 etc.
I guess most Windows PC support helpdesks, you need 1 person in about 30-40 PC's (including say 1 file/printserver per group of 30-40 users).
OTOH, when I worked at Lucent where everyone in the building had a Sun Sparcstation, 1,5 person was doing the whole system administration, installing (nicely through kickstart from the network) and network for >500 people. I have to say they had everything very nicely organized and were 100% automated. I tried to break in my own system (try to get root) in which I usually succeed, but here I didn't:(
In general I'd say, for PC-LAN / windows, 1 person per 30 workstations.
For UNIX: 1 person per 250 UNIX workstations.
For servers, it depends quite on the size and the complexity of the applications of course.
No please not for the KDE/GNOME project! I don't want to be forced to install multitudes of libraries, Corba and/or run some desktop environment just to watch multimedia stuff.
It may use any toolkit (such as gtk) but should not force you to run a particular desktop/windowmanager.
btw I use twm (since 12 years) and will never abandon it.
This is ridiculous. I am allowed to download what I want. How about text-browsers such as Lynx?
In fact I hope this whole advertisement-based "economy" falls apart, since it is annoying, and don't you think that the ads make something free. Who pays for the ads? The companies selling products. Of course these costs are calculated through in the price of the products. In the end the consumer always pays. The bad thing is, advertisement creates the illusion that something is free, yet you're paying for it in another way (as a consumer), and you cannot even choose.
If you'd pay for the content directly, then you can choose whether you think it is worth some of your money or not, and products can get a little cheaper since less advertisement costs have to be made (some will remain since consumers have to be informed somehow).
OTOH indeed they should get rid of it. Since they have better things to do than bother with such nice add-ons. Let them first release a stable Mozilla. Since the source is available, people can always add such niceities later.
Re:Competing with pretty good products
on
Why Not MySQL?
·
· Score: 1
I don't see your problem. It is open source, that is you can get the source of the latest product in order to fix bugs that bother you, or enhance things (no I don't give in to some self-appointed organization that likes to redefine the meaning of normal words by their ideological context).
Yes, the license isn't as free as you might like, but anyway the most important application of open source you have: if you encounter a limitation in the product (bug, missing feature) you can fix it yourself and are not SOL.
Okay, you cannot submit the diffs and expect them to be included in the main-line, so maybe you have to reaply the fixes for each coming release.
OTOH, if you submit some bug for, say, the Linux kernel, what guarantee do you have that it will be included in the official tree within a certain amount of time?
This has been the top RFE (request for enhancement) for months. In february I already posted on the RFE list: Why does the top RFE not even have a feedback / evaluation answer from SUN?
What is the use of the whole RFE mechanism if the top RFE simply gets ignored?
I find it very frustrating that SUN doesn't even give any feedback, but simply ignores the by far most requested RFE for Java.
Very unfortunate for me, as while I am attached to FreeBSD, at the moment I develop Java full-time (though I'd prefer C++) and I need a decent Java implementation on my main platform.
If I have to use the Linux version (together with the Linux tools, linux Netscape, Oracle) then I'm almost only using FreeBSD to emulate Linux. Then I might as well simply use Linux, sigh...
Slashdot Mirror
Even more so for those with slow connections. For years I've used FreeBSD with only a 28k8 modem. The ports system is much much more bandwidth efficient. Example:
RPM: new RPM with some fix. Old RPM based on emacs-20.5, new RPM too. You have to download the whole binary, complete package again. That is outrageious IMO
Port: New fix: Update your port (basically a directory with some files, among which are patch files, in it) usually done by an rsync like mechanism (cvsup) ie the update typically is something like 1kb. Then rebuild the port (the original emacs-20.5.tar.tz source is still lying around in
So for refreshes of a port you typically download 1KB, for RPM you have to reget the whole binary again.
One of the strong points of FreeBSD. I've been upgrading my system for over 5 years, without a single reinstall. Still everything is totally clean and organized.
/usr/local; alas Linux makes a mess by mixing a core system with all add-on packages).
cvsup, together with make world is great for the base system (which is cleanly separated from add-on packages that all reside in
then the ports system may requiring some uninstalling/reinstalling sometimes (it is a bit more primitive in this respect than most Linux distro's) but that isn't such a big deal. The important thing is that the base system itself is 100% upgradeable and never requires a reinstall. Not even when moving to a new machine or upgrading hardware.
Today I use Linux (mainly because of availability of vmware and good JDK's) but in this respect I long back for FreeBSD.
Therefore it is important to switch jobs *often* to ensure being exposed to new things all the time. This keeps you flexible and used to learning new things fast. Once you've done the same methodology/language for more than 5 years, you're lost and it becomes real hard to adopt to new views/methods (sorry for you :)
I'm 34 now, and since my masters degree 10 years ago I've had 10 jobs. Since I switched so often I decided to become self employed 2 years ago. My last job for example was my first exposure to Oracle, at 33 I started with RDBMS in general.
As you get older, it becomes harder to do something new because companies either want someone young (and cheap) or someone with 100% perfect experience.
But, if you got lot of different experience during your 20s (when that is still possible) there are so many things you can do that in most cases you can match the requested experience for 90%. The remaining 10% are new, which is a new change to get professional experience on a new subject.
You can also get experience on new things by hobbying at home, but I have found that:
- companies don't count hobby experience (often not fair, but still)
- as you get older, it gets harder and harder to motivate yourself to really invest time in new things for your hobby at home
It's not just the money. Two other reasons I can think of:
- expensive memory makes their CPU's look cheap in comparison
- slower memory might let people think they need faster (more expensive) CPU's with more built-in cache
When I started (14) I wanted to know top to bottom how a computer worked. With the commonodore pet, and, a short while later the commodore 64, this was possible: a simple command-line oriented environment without many layers of GUI's, wizards, automated things going on behind your back, i.e. you can really know what your doing.
Fascination started really when I was 15 and got a C compiler for the C64 (dog slow, but still). I'd still recommend it today for a beginner (on a faster platform though). C is small and simple, and a bare gcc environment offers much more oversight than the confusing visual environment.
Why should it be different for kids today?
Yes, using visual c++ you can create complex programs faster (at start) but without knowing what exactly is going on. VB is even worse in this respect. It may help them to create a program, but doesn't necessarily teach them how to program.
Yes, I live in Switzerland and of newly sold TV's at least 50% is widescreen. More and more regular TV programs are transmitted in widescreen (throughout Europe). Within one or two years I think 4:3 TV is history (like black&white).
Poor french. Even in Switzerland with its outrageously expensive Swisscom, you can call for the equiv of $0.05/minute to the US now. Of course by using an alternative carrier via a prefix.
They might try to control ISP's (as Germany has also done) and require them to block certain sites.
But that won't really help either, since international phonecalls have become so cheap (less than $0.05/minute) that those who really need to access some blocked materials can easily dial in into a US ISP.
Little calculation: assume the cache is 256kb, randomly written by one in 5 instructions, at a pace of 200MHZ (how fast is cache these days?).
/4 =
:)
Cache width 4 bytes at a time.
I guess these are conservative estimates:
10*10e12 / 200*10e6 * 5 * 256*1024
16384000000 seconds it would last,
ah, that's 532 years. OK, no problem
OTOH if it's only as fast as DRAM, it isn't good enough for cache. invalidating the cache (thus refilling it) after a power-on doesn't really seem a problem to me.
Why not do it both, to ensure success either way:
Split in 3 x 3, thus 9 pieces. Three of them may only do the OS, three only apps, three only Internet.
If only it were true what you say, that talent is rewarded instead. Well, in a way it is: not the talent for inventing something briliant, but mostly the talent to play dirty tricks, bribe, monopolize and blackmail.
In that way, unchecked capitalism is evil indeed. It is not the most talented (from a scientific point of view) that often win, also it is often not the most technically sound standard that survives, but it is the cheaters, thieves and frauds.
Copyright law should protect good inventions, so that people get rewarded for their work. Instead it mostly protects vested interests, (near) monopolies and coroporations that can bully others and seemingly do what they want.
But then they would not have read the letter, and would have considered no legal response had come from Andover.
Breakthrough seems exaggerated.
The IA64 is not a breakthrough chips. There exist numerous other 64-bit risc chips already, with comparable performance (ultrasparc, soon new version, alpha, power). Also the IA64 is not fully backwards compatible with IA32 (finally), that is current IA32 binaries may run on it but at strongly reduced speed AFAIK (for those, upgrading to IA64 wouldn't make sense).
So I don't see what the fuzz w.r.t. the IA64 is all about. At the time it will actually hit the market, it is already outdated, and for example Linux will have faster (sparc/alpha) platforms it can run on.
Amen to that. Normally I'm a law abiding citizen, but this law and what the industry is trying to do with it makes me so mad, that for the first time in my life I feel obliged to 'civil disobedience'. Today I feel guilty when I buy a DVD and/or CD (money is not the problem, I have enough of it) and only feel right if I make illegal copies. Buying them (i.e. not stealing) would make me feel like collaborating with the Germans in WW2.
I hope it creates these feelings in so many people that it becomes like a guerilla war or resistance movement that the industry/government simply cannot win.
The survivor in a proprietary world of closed "standards" indeed is not necessarily the best. In the examples you mention, the loosing of the best has to do with installed base, and incompatability of the better one, which still looses because of the larger installed base. Users are locked into one alternative.
OTOH with open standards, open source and interoperability, things are different. Here there is no locking out of the better alternative, since the two (better and worse) alternative can work together and noone is forced to use the worse alternative for the sake of compatability. Even if one of them implements a different API/standard/protocol, the other one can look at it and copy it.
Alas, w.r.t. Linux there is more and more non-open source (commercial apps). Still, since Linux itself is open, FreeBSD can build a almost 100% perfect Linux emulator (unlike Wine, which cannot work very well since Windows' source is not available).
Note that I do not claim that BSD will win in the end, since I don't claim that it is better. It's internals may be better, but in other respects (support for crappy hardware, easy-to-use packaging) Linux may be better.
But I am convinced that in the end the best one (in perception of end users) will win, and both will evolve to improve further in a darwinian 'struggle for life'.
There is nothing wrong with Xt. It's a nice framework, though a bit cumbersome because of it's OO structure implemented without OO language support.
But Motif OTOH, is a different thing. It uses Xt but in quite an ugly way.
OpenLook, which lost from Motif alas, looks better, and also had a much nicer API. Btw there were two implementations from OpenLook: xview with it's own built-in toolkit (which was easier to program), and Olit, OpenLook on Xt, which was also quite nice.
Sun lost the X-window GUI war because Motif looked more like Windows3. Blergh. If they would have won, Unix would have had a single GUI for a long time already.
B.t.w. when Sun realized OpenLook was dead, they also released the sourcecode. But binaries had been available for free for a longer time already.
Only if you use more than one application using Motif 2.1 simultaneously you'll save some memory. Otherwise, the static version will be faster.
Yes, they try to sell some more of their X-designer, thinking that giving away Motif will help it to regain some ground.
I think it won't, for reasons others have mentioned. It's nice to be able to run some legacy Motif apps with dynamic libraries ISO the huge statically compiled ones, but that it the only use I think. Mainly Netscape, soon to be replaced by Mozilla which uses Gtk.
I must say, I have always been extremely irritated by Motif. Apart from it's ugly look & feel, ugly API (compared to the sample Athena widget set), it seemed outrageous to me that one could get a whole UNIX(like) operating system for free (Linux/*BSD) but had to pay for just a GUI toolkit, even for the runtime alone, so much money that whole 'gratis' advantage of Linux/*BSD would be annihilated by it.
Even MSFT acknowledged that a serious OS needs serious command line and batching facilities.
No, Linux's/UNIX's command line will never die. True, the 'normal' user won't see or use it, but the poweruser/system admin needs it to automate tasks, to work faster and more effective.
Yes it takes more time to learn, but once you've learned it, it is simply faster and more convenient than a GUI.
Sadly, most peoples idea of so called archaic CLI interfaces have been formed by the rotten DOS/Windows CLI, which doesn't compare to those of UNIX, VMS, OS/390 etc.
To give some examples:
:(
I guess most Windows PC support helpdesks, you need 1 person in about 30-40 PC's (including say 1 file/printserver per group of 30-40 users).
OTOH, when I worked at Lucent where everyone in the building had a Sun Sparcstation, 1,5 person was doing the whole system administration, installing (nicely through kickstart from the network) and network for >500 people.
I have to say they had everything very nicely organized and were 100% automated. I tried to break in my own system (try to get root) in which I usually succeed, but here I didn't
In general I'd say, for PC-LAN / windows, 1 person per 30 workstations.
For UNIX: 1 person per 250 UNIX workstations.
For servers, it depends quite on the size and the complexity of the applications of course.
No please not for the KDE/GNOME project! I don't want to be forced to install multitudes of libraries, Corba and/or run some desktop environment just to watch multimedia stuff.
It may use any toolkit (such as gtk) but should not force you to run a particular desktop/windowmanager.
btw I use twm (since 12 years) and will never abandon it.
This is ridiculous. I am allowed to download what I want. How about text-browsers such as Lynx?
In fact I hope this whole advertisement-based "economy" falls apart, since it is annoying, and don't you think that the ads make something free. Who pays for the ads? The companies selling products. Of course these costs are calculated through in the price of the products. In the end the consumer always pays.
The bad thing is, advertisement creates the illusion that something is free, yet you're paying for it in another way (as a consumer), and you cannot even choose.
If you'd pay for the content directly, then you can choose whether you think it is worth some of your money or not, and products can get a little cheaper since less advertisement costs have to be made (some will remain since consumers have to be informed somehow).
OTOH indeed they should get rid of it. Since they have better things to do than bother with such nice add-ons. Let them first release a stable Mozilla. Since the source is available, people can always add such niceities later.
I don't see your problem. It is open source, that is you can get the source of the latest product in order to fix bugs that bother you, or enhance things (no I don't give in to some self-appointed organization that likes to redefine the meaning of normal words by their ideological context).
Yes, the license isn't as free as you might like, but anyway the most important application of open source you have: if you encounter a limitation in the product (bug, missing feature) you can fix it yourself and are not SOL.
Okay, you cannot submit the diffs and expect them to be included in the main-line, so maybe you have to reaply the fixes for each coming release.
OTOH, if you submit some bug for, say, the Linux kernel, what guarantee do you have that it will be included in the official tree within a certain amount of time?
What's the problem?
Just get the patch (few minutes at most), apply the patch, do a
'make bzlilo; make modules; make modules_install; reboot'
and you have your new kernel. It's not as if it would cost you an hour or more.
This has been the top RFE (request for enhancement) for months. In february I already posted on the RFE list:
Why does the top RFE not even have a feedback / evaluation answer from SUN?
What is the use of the whole RFE mechanism if the top RFE simply gets ignored?
I find it very frustrating that SUN doesn't even give any feedback, but simply ignores the by far most requested RFE for Java.
Very unfortunate for me, as while I am attached to FreeBSD, at the moment I develop Java full-time (though I'd prefer C++) and I need a decent Java implementation on my main platform.
If I have to use the Linux version (together with the Linux tools, linux Netscape, Oracle) then I'm almost only using FreeBSD to emulate Linux. Then I might as well simply use Linux, sigh...