Corel kind of started this thing with it's 'closed betatest'. They claim they don't need to deliver source code yet, since their dist isn't publicly available yet.
I read in one of the follow-ups that BluePoint refers to Corel wrt not delivering source code (yet?). These large scale so called closed betatest blur the distinction between private use (in-house) which doens't need delivering source code and open distribution which does require it.
Should the GPL maybe be updated to draw a sharp line in this area, e.g. by explicitly requiring that any use, also so called tests, outside a company (i.e. by any non-employee) is considered to be non-private, and must be accompanied by source code?
Still I find that often you can get hired for something you have already done, but while there you can expand your tasks and do new things. First you'll have to proove yourself on the job. I do this all the time. I'm only 33 now but I think this will also work when older (I've seen older people do that).
company in their MIS department. His idea of programming is to copy blocks of code from the sample CDs that come at the back of programming books, without understanding what they do. Coding for him means moving them around, re-arranging stuff, and so on... but without a solid idea what the stuff does!
How very familiar this sounds. I encounter such copied (and thus unmaintainable code) too often. This reminds me of something a friend of mine told me: they should give programmers only editors that lack a copy function.
Since UNIX doesn't die, Netscape/Mozilla can't die
on
Whither Netscape 5.0?
·
· Score: 1
As long as there is no alternative on UNIX (I don't envisage a decent IE5 release for all UNIX variants and Opera isn't gratis) Netscape/Mozilla cannot die.
UNIX is gaining these days, even on the desktop (Linux, FreeBSD, Solaris/Ray-1 etc) and these users need a web browser. If people say Netscape/Mozilla dies, apparently they also say UNIX on the desktop dies. This simply isn't true.
I worked at a firm that gives UNIX courses. Off course we used X-terminals (NC's avant la lettre).
This was perfect, no maintenance whatsoever, they always work, and the students were always amazed when I told them after a few hours that they had been editing and compiling programs with 10 people on a single pentium computer.
Apart from that: I've always worked in UNIX environments. Yes, we have NT computers now, but 95% of the time they're running an X-window server in full screen mode, giving you the idea (and convenience) that you're working on the UNIX servers via an X-terminal. I find that NC's (X-terminals) are perfect also for cubicle land.
My previous job there were 500 developers working on a few UNIX servers, all through X-terminals. There was 1, just 1 system administrator for the whole lot. I guess that a PC network for that many people and a couple of servers needs many many more system administrators.
FreeBSD at least has a couple of aio_* system calls, that do asynchronous IO, conforming to POSIX.2.
Apart from that, isn't it possible to open(2) a file with O_NONBLOCK *and* O_SYNC options? That way you can set the F_SETSIG flag with fcntl(2), and be notified with a SIGIO when the write has been completed to disk.
"But I'm not rejoicing for the ascendancy of Linux or its penguin mascot."
In other words, if it's not BSD, he doesn't want it to succeed. We don't want users to have a choice. No, we just want them to have BSD.
In a way Linux is good for Open Source in general, but the massive and one-sided attention for Linux is also destructive: For the UNIX world one used to see software available for a lot of variants, and usually it would be ported to other (more obscure) variants pretty easily and soon after there was demand.
Nowadays you see a lot of releases specifically for Linux because it is the biggest market (vmware, SO51, shockwave plugin to name a few examples), and also more and more often in binary-only form. This is the effect of the commercialisation of Linux. I have great doubts whether this is good for UNIX in general.
If Linux wasn't that more popular than other Unices (i.e. more balanced market shares) then chances of ports to most of the UNIX variants would be bigger. Thus, somehow Linux limits the choice you have.
Note that I use Linux, not FreeBSD at the moment because I need vmware. I deeply regret that I cannot run FreeBSD and really feel limited in my choice because of this.
Without Linux, maybe vmware as a commercial product (and others) wouldn't have existed one might say. But then open source alternatives would have had a chance. Now such alternatives lack developer interest because commercial products are available.
Which only proves that such matters are subjective. I've tried Debian several times, but personally I hate it. Way too many packages, with a way too nifty package system, dependencies etc. Way too many maintainers that are too dedicated, thus tweaking and patching their "baby" package(s) to death. It's the most horrible dist I've seen. I'd rather use Redhat, even better like Slackware (which I'm currently using) and FreeBSD is the ultimate. Though, since I need vmware alas I can not run it at the moment. Soon I will, when the vmware kernel modules have been ported over; then vmware can run in the Linux emulation.
I'll tell you why it took so long: because some irresponsible distribution vendors (almost all except for Slackware) last year came with glibc2.0 based distributions. glibc2.0 is incompatible with glibc2.1, but since it has the same major release number, they cannot coexist. This is what has created a real mess.
Note that the glibc developers have always said that glibc2.0 was an alpha release. It should never have been used in the real world!
If everyone would have done like Slackware, wait until glibc2.1 is out and stable, then everyone would do the switch now, and it would have taken only a few months instead of more than a year.
Xmms doesn't require glibc2. I compiled it myself with libc5, but did have to create thread-safe Xlibs first (this is easy, just install libc5 linuxthreads and recompile from the Xfree distribution).
On FreeBSD xmms runs even without thread-safe Xlibs, since it runs without threads at all. That's also possible, without any performance loss.
Btw on single-CPU systems using multithreading brings nothing, only bugs and hard-to-debug programs. It is way easier and more reliable to do concurrent tasks within a single process with interrupts etc.
Re:Huray! Now, more people use C++!!
on
GCC 2.95 Released
·
· Score: 1
What the hell are you talking about? Hashes aren't even part of th STL. What lib are you using?
We use STL all over in our software, on various UNIX platforms. Maps and vectors all the way (where in earlier times we had to fiddle with malloc/realloc and free, now everything goes easy and there are no more forgotten free()s thus no memory leaks), never had any problems.
Most UNIX C++ vendors offer good STL implementations. The only one, alas we had to code around bugs is GNU's libstdc++. I know that they've worked a lot on it since the release of EGCS, so I hope that GCC 2.95 brings improvements in libstdc++ so that it becomes useful too.
Slashdot Mirror
Corel kind of started this thing with it's 'closed betatest'. They claim they don't need to deliver source code yet, since their dist isn't publicly available yet.
I read in one of the follow-ups that BluePoint refers to Corel wrt not delivering source code (yet?). These large scale so called closed betatest blur the distinction between private use (in-house) which doens't need delivering source code and open distribution which does require it.
Should the GPL maybe be updated to draw a sharp line in this area, e.g. by explicitly requiring that any use, also so called tests, outside a company (i.e. by any non-employee) is considered to be non-private, and must be accompanied by source code?
Still I find that often you can get hired for something you have already done, but while there you can expand your tasks and do new things. First you'll have to proove yourself on the job. I do this all the time. I'm only 33 now but I think this will also work when older (I've seen older people do that).
company in their MIS department. His idea of programming is to copy blocks of code from the sample CDs that come at
the back of programming books, without understanding what they do. Coding for him means moving them around,
re-arranging stuff, and so on... but without a solid idea what the stuff does!
How very familiar this sounds. I encounter such copied (and thus unmaintainable code) too often. This reminds me of something a friend of mine told me: they should give programmers only editors that lack a copy function.
As long as there is no alternative on UNIX (I don't envisage a decent IE5 release for all UNIX variants and Opera isn't gratis) Netscape/Mozilla cannot die.
UNIX is gaining these days, even on the desktop (Linux, FreeBSD, Solaris/Ray-1 etc) and these users need a web browser. If people say Netscape/Mozilla dies, apparently they also say UNIX on the desktop dies. This simply isn't true.
I worked at a firm that gives UNIX courses. Off course we used X-terminals (NC's avant la lettre).
This was perfect, no maintenance whatsoever, they
always work, and the students were always amazed when I told them after a few hours that they had been editing and compiling programs with 10 people on a single pentium computer.
Apart from that: I've always worked in UNIX environments. Yes, we have NT computers now, but 95% of the time they're running an X-window server in full screen mode, giving you the idea (and convenience) that you're working on the UNIX servers via an X-terminal. I find that NC's (X-terminals) are perfect also for cubicle land.
My previous job there were 500 developers working on a few UNIX servers, all through X-terminals. There was 1, just 1 system administrator for the whole lot. I guess that a PC network for that many people and a couple of servers needs many many more system administrators.
FreeBSD at least has a couple of aio_* system calls, that do asynchronous IO, conforming to POSIX.2.
Apart from that, isn't it possible to open(2) a file with O_NONBLOCK *and* O_SYNC options? That way you can set the F_SETSIG flag with fcntl(2), and be notified with a SIGIO when the write has been completed to disk.
In other words, if it's not BSD, he doesn't want it to succeed. We don't want users to have a choice. No, we just want them to have BSD.
In a way Linux is good for Open Source in general, but the massive and one-sided attention for Linux is also destructive: For the UNIX world one used to see software available for a lot of variants, and usually it would be ported to other (more obscure) variants pretty easily and soon after there was demand.
Nowadays you see a lot of releases specifically for Linux because it is the biggest market (vmware, SO51, shockwave plugin to name a few examples), and also more and more often in binary-only form. This is the effect of the commercialisation of Linux. I have great doubts whether this is good for UNIX in general.
If Linux wasn't that more popular than other Unices (i.e. more balanced market shares) then chances of ports to most of the UNIX variants would be bigger. Thus, somehow Linux limits the choice you have.
Note that I use Linux, not FreeBSD at the moment because I need vmware. I deeply regret that I cannot run FreeBSD and really feel limited in my choice because of this.
Without Linux, maybe vmware as a commercial product (and others) wouldn't have existed one might say. But then open source alternatives would have had a chance. Now such alternatives lack developer interest because commercial products are available.
Which only proves that such matters are subjective. I've tried Debian several times, but personally I hate it. Way too many packages, with a way too nifty package system, dependencies etc. Way too many maintainers that are too dedicated, thus tweaking and patching their "baby" package(s) to death. It's the most horrible dist I've seen. I'd rather use Redhat, even better like Slackware (which I'm currently using) and FreeBSD is the ultimate. Though, since I need vmware alas I can not run it at the moment. Soon I will, when the vmware kernel modules have been ported over; then vmware can run in the Linux emulation.
Note that the glibc developers have always said that glibc2.0 was an alpha release. It should never have been used in the real world!
If everyone would have done like Slackware, wait until glibc2.1 is out and stable, then everyone would do the switch now, and it would have taken only a few months instead of more than a year.
On FreeBSD xmms runs even without thread-safe Xlibs, since it runs without threads at all. That's also possible, without any performance loss.
Btw on single-CPU systems using multithreading brings nothing, only bugs and hard-to-debug programs. It is way easier and more reliable to do concurrent tasks within a single process with interrupts etc.
What the hell are you talking about? Hashes aren't even part of th STL. What lib are you using?
We use STL all over in our software, on various UNIX platforms. Maps and vectors all the way (where in earlier times we had to fiddle with malloc/realloc and free, now everything goes easy and there are no more forgotten free()s thus no memory leaks), never had any problems.
Most UNIX C++ vendors offer good STL implementations. The only one, alas we had to code around bugs is GNU's libstdc++. I know that they've worked a lot on it since the release of EGCS, so I hope that GCC 2.95 brings improvements in libstdc++ so that it becomes useful too.