Er, you could go the commercial/OSS route with a Sendmail, Inc sort of thing. I hear great gnashing of teeth from folks with Software.COM, plus you get zero source. Exchange!? Er, not. There was a usenix paper a while ago on Earthlink using sendmail/qpopper to handle 2-3 million. I'm sure that Sendmail can handle a piddling 25k users. IMAP is maturing rapidly and there are commercially supported Source available solutions for that. Most of it depends not on users, but on volume of mail.
Mach is basically a very nice, very fast MICROKERNEL. Don't confuse this with/vmlinu{x,z}.
Also, NeXTStep benefitted tremendously by being written with very object oriented code, and libraries into that code. Some of the MacOS *DESIGN* that lets extensions easily slip into the path of many 'lower level' calls (eg, GUI pop-ups, scanning a disk when it's mounted, etc) were implemented better in NeXT Step. It was noted that Novell was able to port WordPerfect in a ridiculously small amount of time due to the OO structure underlying the OS.
Linux (and the *BSDs) are not written this way. It's been tough enough to get C++ to work correctly under Unix (hence the birth of EGCS).
OTOH, Since NeXTStep was guided by Jobs, there were a number of fatal gaffs made: It used Objective C when C++ was on it's way to becoming standard; it used a PostScript Interface, when X was becoming a standard; it used "NetInfo" rather than the standard NIS (it could be an NIS client, but that wasted the beautiful interface that NI had).
All of these were arguably 'better' but the bottom line was that in my work env of 15 different Unix architectures, I could do everything one way for ALL of those others, but the NeXT's needed special treatment. (The easiest way to piss off NeXT marketing folks was to compare the Display PS to Sunview and ask how to run some GUI program on another display.)
So let us celebrate that the remarkably consistant MacOS interface which was obviously designed rather than hacked together in reaction to market pressures (ala Windows) finally has a decent kernel underneath it. I will be delighted:
if I can telnet to the Mac to fix problems
if I can keep one up for more than a week while doing hard stuff on it
if it finally has preemtive multitasking
if it finally has protected memory
if user-level programs cannot affect priviledged areas and viruses become moot
If it works well with the well documented 44BSD programming interface, then perhaps the Linux side of the world might start writing code that's a tad more Unix portable that it is now (/usr/include/linux/?!).
Slashdot Mirror
Er, you could go the commercial/OSS route with a Sendmail, Inc sort of thing. I hear great gnashing of teeth from folks with Software.COM, plus you get zero source. Exchange!? Er, not. There was a usenix paper a while ago on Earthlink using sendmail/qpopper to handle 2-3 million. I'm sure that Sendmail can handle a piddling 25k users. IMAP is maturing rapidly and there are commercially supported Source available solutions for that. Most of it depends not on users, but on volume of mail.
Also, NeXTStep benefitted tremendously by being written with very object oriented code, and libraries into that code. Some of the MacOS *DESIGN* that lets extensions easily slip into the path of many 'lower level' calls (eg, GUI pop-ups, scanning a disk when it's mounted, etc) were implemented better in NeXT Step. It was noted that Novell was able to port WordPerfect in a ridiculously small amount of time due to the OO structure underlying the OS.
Linux (and the *BSDs) are not written this way. It's been tough enough to get C++ to work correctly under Unix (hence the birth of EGCS).
OTOH, Since NeXTStep was guided by Jobs, there were a number of fatal gaffs made:
It used Objective C when C++ was on it's way to becoming standard; it used a PostScript Interface, when X was becoming a standard; it used "NetInfo" rather than the standard NIS (it could be an NIS client, but that wasted the beautiful interface that NI had).
All of these were arguably 'better' but the bottom line was that in my work env of 15 different Unix architectures, I could do everything one way for ALL of those others, but the NeXT's needed special treatment. (The easiest way to piss off NeXT marketing folks was to compare the Display PS to Sunview and ask how to run some GUI program on another display.)
So let us celebrate that the remarkably consistant MacOS interface which was obviously designed rather than hacked together in reaction to market pressures (ala Windows) finally has a decent kernel underneath it. I will be delighted:
if I can telnet to the Mac to fix problems
if I can keep one up for more than a week while doing hard stuff on it
if it finally has preemtive multitasking
if it finally has protected memory
if user-level programs cannot affect priviledged areas and viruses become moot
If it works well with the well documented 44BSD programming interface, then perhaps the Linux side of the world might start writing code that's a tad more Unix portable that it is now (/usr/include/linux/?!).