Also wages were planned by the state, and pretty much determined by how much learning was involved in getting that job. E.g., an electronics engineer or doctor would get a lot better paid than a plumber. This, unfortunately, was not true. Wages for lower-rank scientific workers were well below the level of skilled workers (rmember, it was a "country of the working class". Also, as a plumber, you'd get more chance to collect extra (unofficial) money from your clients. Still, the prestige of a scientific worker was high. It was something to be proud of. Yes, I come from Russia (still living here). I am 34 now, so I remeber 70s and 80s.
The C cross compiler suite, as far as I can tell, is something they use to port their VM between platforms. When you program _for_ the VM, you have to use Limbo. Still, I think the compiler might be one of the most valuable parts of this distribution. It was originally written by Ken Thompson; it is fast; its code is small and readable. If enough people notice, that could be a worty competition to GCC.
I can see a couple of problems with that. First, Scheme requires GC. Kernel memory isn't GC'd. Second, compilers from Scheme to C produce unreadable code (Hobbit is an exception, but it was buggy beyond repair last time I looked). Thus you would need to teach Linus Scheme to get that code into the kernel. Debugging and prototyping -- yes, but read Scheme code in the kernel -- no.
Well maybe I was not clear enough. Of course there were differences. Like no free (public) speech, no free elections, market economy, etc. Important differences.
Still, comparing this to People's Democratic Republic of Korea is an overkill. North Korea is a Stalin-type tyranny, where your private life is constantly under pressure from the State.
In the USSR in the 70s and 80s, nobody really believed the official ideology -- including the authorities themselves. It largely became a ritual, simple rules one should follow so that the state leaves you alone. You didn't criticize Brezhnev at party meetings, much the same way you don't (openly) hack cryptographic software in today's America. There were political prisoners, yes, but you needed to really press for it to become one.
To repeat the important point: the state was easy to ignore. I don't believe this is possible in North Korea, or was possible in Russia in the 30s to 50s.
There were good sides to the regime, too (no wonder Communists are still popular here):
-- guaranteed minimum level of life, much more so than in today's Russia (well, that depended on oil prices, but still...)
-- better education system
-- much lower crime level. That, by the way, is a general tendency: less democracy => less crime.
-- less nationalism. Not that it was perfect. It was harder to enter a university if you were a Jew (I am). But that was nothing compared to today's anti-Caucasus sentiments of many Russians (including officials). I don't know how that translates to Korean situation, though.
I never thought I would defend Brezhnev times before anyone, honest. I would never want to go back. But there are different levels of badness.
Oh yes, I lived in Moscow. I don't think it was _that_ different from Novosibirsk.
> I visited the USSR in the early 80s and I can assure you that your > speculations were quite unfounded. > If you are not convinced, I suggest a trip to North Korea before it > falls apart.
Sorry man. I didn't VISIT the USSR in the early 80s -- i lived there. I still live in Russia. I have lived a year in America, too, so I am able to compare.
The USSR in the 80s, while not a democracy by any means, was _very_ far from the insanity that was and is North Korea. You know, we used to make jokes about Kim Il Sung and the "Free Korea" magazine back in 1983.
Life in Russia in the 70s and 80s was probably closer to life in America than in North Korea. That still does not mean it was "normal" by Western standards, of course.
Still, stack is usually much smaller than the heap. So this
I am not a big fan of Boehm's conservative garbage collector due to the fact it relies on unnecessary sweeps of the program dataspace just in case it *may* contain a bit pattern that might resemble a pointer. Too much unnecessary not used memory is conserved.
I am not a big fan of Boehm's conservative garbage collector due to the fact it relies on unnecessary sweeps of the program dataspace just in case it *may* contain a bit pattern that might resemble a pointer.
There is a way to allocate "typed" memory through Boehm. If you always use the "right" allocation routines, no memory will be scanned which doesn't need to be. And if you use Boehm's collector in code generated from higer-than-C level languages, it's not so hard to make sure you always put type annotations in place.
That said, Boehm's collector is still mark-and-sweep, which is, in the general case, worse than a copying one. But it _is_ extremely convenient for writing GC'd
code in C without too much manual work. And that work you can do later when tuning for speed.
Overall, I'm sorry Boehm's collector is not used more. Heck, it is not even part of any standard Linux distribution. It should have been an option for glib -- that way one could easily write GC'd GTK applications. (I'm not sure taht it should have been the default -- after all, in principle glib could be used for realtime apps. But anything with a GUI is not hard realtime.) By the way, if you do patch glib, the resulting GTK works quite OK.
> > How would I compile, say, Lisp into Java bytecode?
> This is how. Not much on Lisp, but quite a bit for >Scheme. There are also many other languages listed at > >that site that have been made to work inside of Java. I >know that JPython is very popular!
The trouble with JVM and Scheme is that the instruction set does not support tail recursive calls (required by Scheme and very important for any functional language). This makes complete implementation impossible.
.NET has it. What a shame -- Gosling has a Lisp background and ought to have known better.
Let's not forget what Motif was supposed to kill (and killed, eventually) -- Open Look. This was a toolkit which looked much, much better and (at least with the Xview library) was a much greater pleasure to program in. Actually, it had, probably the clearest interface I have ever seen among C or C++ GUI toolkits. I have heard it is difficult to extend (maybe, never tried), but simple tasks like creating forms were really simple and intuitive. This was probably the only toolkit where I could read and understand sources _before_ reading documentation.
What do you think about John Ousterhout's ideas presented in this article?_ Do you feel his critique of the OO pradigm is correct? Don't you feel object-oriented programs are extremely difficult to reuse parts of? (for example, it is a most unpleasant task to deal simultaneously with two libraries assuming different type hierarchies).
I am using RedHat 6.1 with Gnome on an Alpha and I have had no problems with Gnome -- so Iam not quite sure what yuo are talking about. (Well, it's not exactly 6.1, it's RawHide from the time slightly before 6.1. but it's got, as far as I understand, the same Gnome version). OTOH, I could not make my SCSI card initialize properly with AlphaBIOS. With SRM, it works just fine. I like SRM.
and the "voluntary simplicity" movement in software, first identified by name in late 1999, had consumers turning away by the millions from the bloated Office suite and adopting various simplified freeware solutions instead. Hmm... I like the term. Good slogan.
Slashdot Mirror
The migrated code originated from V7, not BSD. (It is under BSD license by now, anyway.)
Also wages were planned by the state, and pretty much determined by how much learning was involved in getting that job. E.g., an electronics engineer or doctor would get a lot better paid than a plumber.
This, unfortunately, was not true. Wages for lower-rank scientific workers were well below the level of skilled workers (rmember, it was a "country of the working class".
Also, as a plumber, you'd get more chance to collect extra (unofficial) money from your clients.
Still, the prestige of a scientific worker was high. It was something to be proud of.
Yes, I come from Russia (still living here). I am 34 now, so I remeber 70s and 80s.
The C cross compiler suite, as far as I can tell, is something they use to port their VM between platforms. When you program _for_ the VM, you have to use Limbo.
Still, I think the compiler might be one of the most valuable parts of this distribution. It was originally written by Ken Thompson; it is fast; its code is small and readable.
If enough people notice, that could be a worty competition to GCC.
I do still use uucp. Postfix handles it without problems.
I can see a couple of problems with that.
First, Scheme requires GC. Kernel memory isn't GC'd.
Second, compilers from Scheme to C produce unreadable code (Hobbit is an exception, but it was buggy beyond repair last time I looked). Thus you would need to teach Linus Scheme to get that code into the kernel.
Debugging and prototyping -- yes, but read Scheme code in the kernel -- no.
Well maybe I was not clear enough. Of course there were differences. Like no free (public) speech, no free elections, market economy, etc. Important differences.
Still, comparing this to People's Democratic Republic of Korea is an overkill. North Korea is a Stalin-type tyranny, where your private life is constantly under pressure from the State.
In the USSR in the 70s and 80s, nobody really believed the official ideology -- including the authorities themselves. It largely became a ritual, simple rules one should follow so that the state leaves you alone. You didn't criticize Brezhnev at party meetings, much the same way you don't (openly) hack cryptographic software in today's America. There were political prisoners, yes, but you needed to really press for it to become one.
To repeat the important point: the state was easy to ignore.
I don't believe this is possible in North Korea, or was possible in Russia in the 30s to 50s.
There were good sides to the regime, too (no wonder Communists are still popular here):
-- guaranteed minimum level of life, much more so than in today's Russia (well, that depended on oil prices, but still...)
-- better education system
-- much lower crime level. That, by the way, is a general tendency: less democracy => less crime.
-- less nationalism. Not that it was perfect. It was harder to enter a university if you were a Jew (I am). But that was nothing compared to today's anti-Caucasus sentiments of many Russians (including officials). I don't know how that translates to Korean situation, though.
I never thought I would defend Brezhnev times before anyone, honest. I would never want to go back. But there are different levels of badness.
Oh yes, I lived in Moscow. I don't think it was _that_ different from Novosibirsk.
> I visited the USSR in the early 80s and I can assure you that your
> speculations were quite unfounded.
> If you are not convinced, I suggest a trip to North Korea before it
> falls apart.
Sorry man. I didn't VISIT the USSR in the early 80s -- i lived there. I still live in Russia.
I have lived a year in America, too, so I am able to compare.
The USSR in the 80s, while not a democracy by any means,
was _very_ far from the insanity that was and is North Korea. You know, we used to make jokes about Kim Il Sung and the "Free Korea" magazine
back in 1983.
Life in Russia in the 70s and 80s was probably closer to life in
America than in North Korea. That still does not mean it was "normal" by Western
standards, of course.
This all sounds very Plan 9-ish. (Not that you can read files as directories in Plan 9.)
Still, stack is usually much smaller than the heap. So this
I am not a big fan of Boehm's conservative garbage collector due to the fact it relies on unnecessary sweeps of the program dataspace just in case it *may* contain a bit pattern that might resemble a pointer. Too much unnecessary not used memory is conserved.
overestimates the risk.
There is a way to allocate "typed" memory through Boehm. If you always use the "right" allocation routines, no memory will be scanned which doesn't need to be. And if you use Boehm's collector in code generated from higer-than-C level languages, it's not so hard to make sure you always put type annotations in place.
That said, Boehm's collector is still mark-and-sweep, which is, in the general case, worse than a copying one. But it _is_ extremely convenient for writing GC'd code in C without too much manual work. And that work you can do later when tuning for speed.
Overall, I'm sorry Boehm's collector is not used more. Heck, it is not even part of any standard Linux distribution. It should have been an option for glib -- that way one could easily write GC'd GTK applications. (I'm not sure taht it should have been the default -- after all, in principle glib could be used for realtime apps. But anything with a GUI is not hard realtime.) By the way, if you do patch glib, the resulting GTK works quite OK.
> > How would I compile, say, Lisp into Java bytecode?
> This is how. Not much on Lisp, but quite a bit for >Scheme. There are also many other languages listed at > >that site that have been made to work inside of Java. I >know that JPython is very popular!
The trouble with JVM and Scheme is that the instruction set does not support tail recursive calls (required by Scheme and very important for any functional language). This makes complete implementation impossible.
.NET has it. What a shame -- Gosling has a Lisp background and ought to have known better.
Let's not forget what Motif was supposed to kill (and killed, eventually) -- Open Look. This was a toolkit which looked much, much better and (at least with the Xview library) was a much greater pleasure to program in. Actually, it had, probably the clearest interface I have ever seen among C or C++ GUI toolkits. I have heard it is difficult to extend (maybe, never tried), but simple tasks like creating forms were really simple and intuitive. This was probably the only toolkit where I could read and understand sources _before_ reading documentation.
What do you think about John Ousterhout's ideas presented in this article?_ Do you feel his critique of the OO pradigm is correct? Don't you feel object-oriented programs are extremely difficult to reuse parts of? (for example, it is a most unpleasant task to deal simultaneously with two libraries assuming different type hierarchies).
I am using RedHat 6.1 with Gnome on an Alpha
and I have had no problems with Gnome -- so
Iam not quite sure what yuo are talking about.
(Well, it's not exactly 6.1, it's RawHide from
the time slightly before 6.1. but it's got, as far as I
understand, the same Gnome version). OTOH, I could not
make my SCSI card initialize properly with AlphaBIOS. With
SRM, it works just fine. I like SRM.
and the "voluntary simplicity" movement in software, first identified by name in late 1999, had consumers
turning away by the millions from the bloated Office suite and adopting various simplified freeware solutions instead.
Hmm... I like the term. Good slogan.