First, what are you doing with these systems? Where I used to work, we had/usr be 1-2G, and/usr/local 3-4G, because everything we did relied on specialized software. 1-2G for/usr was overkill by far. I'd recommend, for a syslog server, 200M / (first partition, or separate out/boot with 50M), 2G/usr, 2G for/opt (does anything but KDE in your setup use this?), and 1G for/tmp (and for the idiots out there, be clear that it doesn't get backed up). You must evaluate your needs for/var and/home yourself, since/home is dependant on no. of users and what sort of crap they'll stick in there, and/var depends on what you'll stick in there. I'd recommend 200-500M per user at most, and for a syslog server, everything else in/var. For a web server, however much space you need plus a gig or two for/var; for a mail server, whatever is left from the others (1G per user should be more than enough).
Your "details" about the setup were useless, though -- the important information is:
/var-based network services provided (mail, web, syslog, DNS), and estimated use
additional software installed into/usr or/usr/local (for most things, this can be one partition, but not necessarily)
/opt-based services, and estimated space used
estimated required RAM, amount of RAM available, and estimated upper limit of use
Bzzzt. "Average" is a generic term, "mean" is a specific term. Just like they teach little kids, modes, means, and medians are all forms of averages. Although in general (non-technical) usage average equates to mean, in technical usage it is non-specific:
"A number that typifies a set of numbers of which it is a function." (from www.dictionary.com
It can be used to refer to mean, mode, median, and other, more esoteric averages.
For immediate compatability, X11 could draw through Quartz like the Xservers for MacOSX do.
OK, let's be clear here. There are no X servers for Quartz. There is an X server for MacOS X Server (which uses Display PostScript, not DisplayPDF), and there is an X server for Darwin (which could be made to work with MacOS X Client, but not through Quartz).
Further, in Linux, it would be drawing Quartz through X11 rather than the other way around.
There was discussion of this on the NeXTSTEP newsgroups just a few weeks ago. So I'd say a lot of people are unhappy with the loss of a resource (deja) to which they've grown to like.
...making bad business decisions to get vengeance... that's why apple is where they are today.
You'll have to forgive me, but you're wrong. Do you have any clue what you're talking about? At this point I think it's safe to say Apple can walk away from ATI any time they feel like it -- 3Dfx and nVidia would love to do OEM packaging of their cards for Apple, and have been publicly pushing Apple to do this for quite some time.
I think Apple is the largest buyer of ATI chips/cards right now, and possibly the only major computer manufacturer that preloads ATI cards. It is most definitely in ATI's best interest to toe any lines Apple draws, because Apple represents the biggest new market for 3dfx and nVidia products.
My point is, 'What's the point?' Please don't post this non-interesting garbage on slashdot!
Well, it is BSD-related. It's also a few weeks late, since it's been out for quite some time. Much nicer would be a link from the/. BSD section to Daemon News, but what do I know?
Re:Was Linux the competitor?
on
Endgame For SCO
·
· Score: 1
(dunno if you're still reading this thread, but...)
What do you mean they don't handle the load well? Are the kernels setup to detect the second proc?
I mean, when the load hits the roof, and memory usage tops out, Linux pukes. Before that, things are hunky-dory, but if too many people start running their statistical programs on full brain volumes (by volume I mean 4d volumes), then the system sometimes crashes. On the Octane, things get veeeerrrry slllooooowww, and then someone finishes and things go back to normal. Of course, it takes some work to hit that level on our dual PIII/600, 256M RAM, 512M swap Linux system (especially compared to the Octane when it was a single R10k/128M), but when it hits the roof...
I'm not knocking Linux, I'm just saying Irix is more solid and better designed in high-load, high-memory situations. Which is reasonable considering some of the places Irix has been used.
But there are now differences between open source and free software (at least according to OSI and FSF, arguably the two big deciders even if neither term is really limited in usage).
The APSL is not free software, but it is open source, for example. There would be more, except the OSI stopped certifying new licenses (AFAICT) a long time ago.
What it comes down to for me is this: Open Source is unconcerned with the users; a good number of annoying licenses (from the user-programmer's PoV) have come into existance because of OSI's policies. Companies have essentially been able to progressively trim down the rights extended to licensees because of their willingness to extend the Open Source blanket to new licenses -- the new licenses are no longer written by hackers to make sure their software isn't misused, but by lawyers intent on making sure as little IP escapes as possible.
Their attempt and failure to trademark "Open Source" made it abundantly clear, as well, that companies didn't have to toe community lines to fit in -- Plan 9, an operating system that I really like, now refers to itself as Open Source even though their license has some of the nastiest clauses I've seen for end-user licenses ("if you bring an intellectual property claim against any Contributor you will lose your license to Plan 9," essentially and paraphrased although IANAL) which would make any business that relies on Plan 9 unable to protect their intellectual property (which, as it happens, includes enforcing the terms of the GPL on other software) since another entity could probably become a contributor to Plan 9 trivially (i.e., find a single bug, fix it, rape other company). A company wouldn't dare call their software Free Software (although possibly freeware or free) unless the license seemed free to the community; but the community doesn't matter to Open Source.
Free Software, OTOH, is all about users. The user's right to modify the code, the user's right to borrow the code, the user's right to learn from the code, and the user's right to be part of a community that shares the code. If nobody makes money from it, well, that's OK because people benefitted. If people do make money from it, well that's great because more people benefitted.
The point to this long rant is that, really, there is a difference between Open Source and Free software, whether the people who started it want to be different or not. That difference is how, at the end of the day, a person like me feels after contributing to both -- with free software, I've added something; with open source, something has been taken from me.
IMHO, =all= the characters in the X-Men comic books, TV animation, movie, etc, human AND mutants, have a mix of "good" and "evil".
Nah. Sabretooth is a pretty evil bastard. A reasonable number of original members in the Brotherhood of Mutants were pretty evil. Not Mystique, though -- their selection of a supermodel for Mystique reduced the character to almost nothing but eye-candy (damnit, I loved her dialogues with the blind old woman in the comics!).
Re:reports of SCO's demise...
on
Endgame For SCO
·
· Score: 1
sigh
You beat me to the punch. Thank you, the commentary in the (slashdot) article (I haven't gotten around to reading the osOpinion article yet) was a little bigoted and short on information. Another post claimed the text was quoted from the original article, which doesn't make it look too good...
Re:Was Linux the competitor?
on
Endgame For SCO
·
· Score: 2
Here in Pgh there's an SGI-only brain imaging lab, but that was right down the hall from the lab I worked in, which was Linux and HP-UX. It was frankly pleasant to work on the SGIs (since we shared resources), but at the same time getting their Octane up to a dual processor system with a measly 256M RAM was apparently very expensive. We had a couple dual processor Linux machines with 256M, but they didn't handle high load as nicely:(
Try IMAP with a reasonable AUTH mechanism (yes, part of the protocol includes the ability to encrypt whole sessions), or APOP. I'm less familiar with APOP, but in general, IMAP were designed to be good protocols -- optional encryption, all information over one channel, etc.
PINE is something of a 'reference implementation' of an IMAP-capable mailreader, I think it was the first. Mutt may 'support' IMAP, but everything I read suggested it supported it somewhat.
However, virtually every non-UNIX GUI mail client nowadays supports IMAP -- Outlook, Mulberry, Eudora, Netscape Messenger (PMMail for Windows and OS/2 does not sigh). Virtually every X-based mail client is 'working on it.'
I saw one in a catalog just a couple yesterday -- the same catalog that sold 'tiny reading glasses in a large pen' as a 'geek toy.' It looks like a glorified version of a phone number manager, rather than something on the level of the Palm Pilot (which, really, is much closer to a computer with limited ports than an addressbook).
It would be nice to believe, but it looks like a cheap knock-off for people who can't tell the difference.
Last time I read anything about it, it didn't support binary inheritance (CORBA can, I believe Bonobo does). Maybe it does now, but I have difficulty believing that a post-design add-on is as good as the real thing.
Why does COM 'suck eggs'?
Hyperbole on my part. CORBA does have its problems (I would like inheritable exceptions like Java), but it is an object model from the ground up.
The most important part of my post, though, was glossed over: Bonobo is not a 'wannabe COM' object model, it is related to COM in the press because it performs a similar function. It clearly starts from a different grounding.
Why not just use COM? Bonobo is ripped off from COM anyway.
No, it's not. COM is fake objects; CORBA is real objects. They refer to it as being 'like COM' to give people an idea of where it stands, but that doesn't mean it sucks eggs like COM.
Bonobo consists of both an object model, and a Gtk+-based implementation of that model, plus the object broker. It's not language- or system-dependant, and you shouldn't even have to care what language a class you inherit your class from is written in.
No, you're wrong about X. John Carmack certainly did port X to MacOS X Server, but a lot of that code -- including the DisplayPostscript that MacOS X Server used -- has gone away (X now runs on Darwin, though).
The real issue the other poster was driving home, though, was that porting to MacOS X in no implies "porting to UNIX," regardless of whether a X server exists for MacOS X.
Calling OS/2 OO is like calling Linux OO. The WPS is OO, not OS/2. You do something without going via the WPS you're screwed, all your OO stuff is ratshit.
Fair enough, not everything's OO. It does have a language-neutral OO API inherent to the system, though, which is more than Windows can say. Or Linux for that matter.
OS/2 IS DOS - a 32bit DOS. that's what it was designed to be.
No, it's not a 32bit DOS. It supports, from OS/2 1.0, a wealth of features not available in DOS until Windows95. Preemptive multitasking was there, multithreading, SMP support (OK, so that came in 2.x I think), plenty to differentiate it.
the pretty interface didn't show up until after they realised they were dead in the water against Microsoft.
Ummm... no. As a matter of fact, the pretty interface showed up while Microsoft and IBM were working together on it. That IBM used that in their marketing of Warp as 'a better Windows than Windows' helped them split.
What a load of shit. it's not released yet. it's not being used in the Real World{tm}. How could you or anyone possibly make such a comparison?
MacOSX Server has been released for a year, like I said, and it's solid. MacOS X isn't available yet, but I doubt they could ruin the stability enough to compare with NT in the interim.
'ease of use' is inversly proportional to the user's IQ.
Ooooh, more crap. Short-sighted, pointy-clicky "ease of use" is painful to intelligent people. Something like the text editor joe is a good example real ease of use: a steady learning curve that gets a person moving in five minutes, and lets them do almost anything in five hours. Good user interface is part of the design philosophy, not a stupid GUI that the programmers ask other people to develop.
Give Microsoft credit for attempting to creating an object-oriented operation system, while trying to maintain compatibility with the past. If you look at the internals, there is a considerable amount of power in their object methods.
No. I'll give them that credit when they deserve it; as it is, OS/2 provides a real object-oriented API, and maintains rather equivalent compatibility with the past (I am comparing OS/2 and Windows98 here, so 'compatibility' is used in reference to DOS and Windows 3.1). OS/2 also doesn't have the DOS background that Windows98 does, had preemptive multitasking at its inception, and is mostly dead because of a feud between IBM and MS.
Far more than Apple, I might point out, who has taken 16 years to give us preemptive multitasking (technically, they still haven't, of course).
First of all, MacOS X Server has been shipping for, I believe, a year now. If you regard Jobs as being Apple, then Jobs has been shipping preemptive multitasking since before OS/2 had a GUI (which is before Windows had preemptive multitasking too), and he's been working on bringing that back to Apple since '96.
Further, Apple hasn't delivered yet because they're more ambitious. They've gone through several attempts that simply didn't live up to their standards. If you want to look at places that innovate with backwards compatibility, look at Apple -- architecture change, no problem. Complete replacement of everything from the kernel to the userland, your old programs will still work. MacOS X is really innovative; it's more modern than UNIX, more stable than Windows NT or OS/2, and easier to use than anything.
Slashdot Mirror
First, what are you doing with these systems? Where I used to work, we had /usr be 1-2G, and /usr/local 3-4G, because everything we did relied on specialized software. 1-2G for /usr was overkill by far. I'd recommend, for a syslog server, 200M / (first partition, or separate out /boot with 50M), 2G /usr, 2G for /opt (does anything but KDE in your setup use this?), and 1G for /tmp (and for the idiots out there, be clear that it doesn't get backed up). You must evaluate your needs for /var and /home yourself, since /home is dependant on no. of users and what sort of crap they'll stick in there, and /var depends on what you'll stick in there. I'd recommend 200-500M per user at most, and for a syslog server, everything else in /var. For a web server, however much space you need plus a gig or two for /var; for a mail server, whatever is left from the others (1G per user should be more than enough).
Your "details" about the setup were useless, though -- the important information is:
Bzzzt. "Average" is a generic term, "mean" is a specific term. Just like they teach little kids, modes, means, and medians are all forms of averages. Although in general (non-technical) usage average equates to mean, in technical usage it is non-specific:
"A number that typifies a set of numbers of which it is a function." (from www.dictionary.comIt can be used to refer to mean, mode, median, and other, more esoteric averages.
OK, let's be clear here. There are no X servers for Quartz. There is an X server for MacOS X Server (which uses Display PostScript, not DisplayPDF), and there is an X server for Darwin (which could be made to work with MacOS X Client, but not through Quartz).
Further, in Linux, it would be drawing Quartz through X11 rather than the other way around.
You mean like GNUstep? Cocoa is Yellowbox renamed in OPENSTEP renamed is being cloned by GNUstep.
You forgot to get extra monitors for the extra video cards, you forgot to get AirPort (card and hub).
Your history is so far out of whack that I can't really address it... suffice to say:
- BSD and AT&T UNIX shared stuff from Version 6 through Version 8, IIRC, but SysV was developed afterwards by AT&T in conjunction with Sun
- SCO now handles SysV, and the Open Group handles UNIX
- BSD and SysV are the two major strands of UNIX, not SysV and "SrV"
- Linux is not UNIX and neither are {Net|Open|Free}BSD, but they might as well be
This might help you out a bit...There was discussion of this on the NeXTSTEP newsgroups just a few weeks ago. So I'd say a lot of people are unhappy with the loss of a resource (deja) to which they've grown to like.
Fluids are fluids, but some fluids support the weight of vehicles better :)
You'll have to forgive me, but you're wrong. Do you have any clue what you're talking about? At this point I think it's safe to say Apple can walk away from ATI any time they feel like it -- 3Dfx and nVidia would love to do OEM packaging of their cards for Apple, and have been publicly pushing Apple to do this for quite some time.
I think Apple is the largest buyer of ATI chips/cards right now, and possibly the only major computer manufacturer that preloads ATI cards. It is most definitely in ATI's best interest to toe any lines Apple draws, because Apple represents the biggest new market for 3dfx and nVidia products.
Well, it is BSD-related. It's also a few weeks late, since it's been out for quite some time. Much nicer would be a link from the /. BSD section to Daemon News, but what do I know?
(dunno if you're still reading this thread, but...)
I mean, when the load hits the roof, and memory usage tops out, Linux pukes. Before that, things are hunky-dory, but if too many people start running their statistical programs on full brain volumes (by volume I mean 4d volumes), then the system sometimes crashes. On the Octane, things get veeeerrrry slllooooowww, and then someone finishes and things go back to normal. Of course, it takes some work to hit that level on our dual PIII/600, 256M RAM, 512M swap Linux system (especially compared to the Octane when it was a single R10k/128M), but when it hits the roof...
I'm not knocking Linux, I'm just saying Irix is more solid and better designed in high-load, high-memory situations. Which is reasonable considering some of the places Irix has been used.
But there are now differences between open source and free software (at least according to OSI and FSF, arguably the two big deciders even if neither term is really limited in usage).
The APSL is not free software, but it is open source, for example. There would be more, except the OSI stopped certifying new licenses (AFAICT) a long time ago.
What it comes down to for me is this: Open Source is unconcerned with the users; a good number of annoying licenses (from the user-programmer's PoV) have come into existance because of OSI's policies. Companies have essentially been able to progressively trim down the rights extended to licensees because of their willingness to extend the Open Source blanket to new licenses -- the new licenses are no longer written by hackers to make sure their software isn't misused, but by lawyers intent on making sure as little IP escapes as possible.
Their attempt and failure to trademark "Open Source" made it abundantly clear, as well, that companies didn't have to toe community lines to fit in -- Plan 9, an operating system that I really like, now refers to itself as Open Source even though their license has some of the nastiest clauses I've seen for end-user licenses ("if you bring an intellectual property claim against any Contributor you will lose your license to Plan 9," essentially and paraphrased although IANAL) which would make any business that relies on Plan 9 unable to protect their intellectual property (which, as it happens, includes enforcing the terms of the GPL on other software) since another entity could probably become a contributor to Plan 9 trivially (i.e., find a single bug, fix it, rape other company). A company wouldn't dare call their software Free Software (although possibly freeware or free) unless the license seemed free to the community; but the community doesn't matter to Open Source.
Free Software, OTOH, is all about users. The user's right to modify the code, the user's right to borrow the code, the user's right to learn from the code, and the user's right to be part of a community that shares the code. If nobody makes money from it, well, that's OK because people benefitted. If people do make money from it, well that's great because more people benefitted.
The point to this long rant is that, really, there is a difference between Open Source and Free software, whether the people who started it want to be different or not. That difference is how, at the end of the day, a person like me feels after contributing to both -- with free software, I've added something; with open source, something has been taken from me.
Nah. Sabretooth is a pretty evil bastard. A reasonable number of original members in the Brotherhood of Mutants were pretty evil. Not Mystique, though -- their selection of a supermodel for Mystique reduced the character to almost nothing but eye-candy (damnit, I loved her dialogues with the blind old woman in the comics!).
sigh
You beat me to the punch. Thank you, the commentary in the (slashdot) article (I haven't gotten around to reading the osOpinion article yet) was a little bigoted and short on information. Another post claimed the text was quoted from the original article, which doesn't make it look too good...
Here in Pgh there's an SGI-only brain imaging lab, but that was right down the hall from the lab I worked in, which was Linux and HP-UX. It was frankly pleasant to work on the SGIs (since we shared resources), but at the same time getting their Octane up to a dual processor system with a measly 256M RAM was apparently very expensive. We had a couple dual processor Linux machines with 256M, but they didn't handle high load as nicely :(
Don't get a video card, get a PC Weasel 2000.
It works like a (text-only) video card, but outputs to a serial line, so you can use a regular PC like a real server.
Of course, my vote would probably go for FreeBSD if you insist on a PC for NFS-serving, but other things (Irix and/or Solaris) might be better.
Try IMAP with a reasonable AUTH mechanism (yes, part of the protocol includes the ability to encrypt whole sessions), or APOP. I'm less familiar with APOP, but in general, IMAP were designed to be good protocols -- optional encryption, all information over one channel, etc.
PINE is something of a 'reference implementation' of an IMAP-capable mailreader, I think it was the first. Mutt may 'support' IMAP, but everything I read suggested it supported it somewhat.
However, virtually every non-UNIX GUI mail client nowadays supports IMAP -- Outlook, Mulberry, Eudora, Netscape Messenger (PMMail for Windows and OS/2 does not sigh). Virtually every X-based mail client is 'working on it.'
I saw one in a catalog just a couple yesterday -- the same catalog that sold 'tiny reading glasses in a large pen' as a 'geek toy.' It looks like a glorified version of a phone number manager, rather than something on the level of the Palm Pilot (which, really, is much closer to a computer with limited ports than an addressbook).
It would be nice to believe, but it looks like a cheap knock-off for people who can't tell the difference.
Note that, as a matter of fact, a sharp-looking graphite (or tuxedo) iBook is also available. Take a look at Apple's iBook site.
The most important part of my post, though, was glossed over: Bonobo is not a 'wannabe COM' object model, it is related to COM in the press because it performs a similar function. It clearly starts from a different grounding.
Bonobo consists of both an object model, and a Gtk+-based implementation of that model, plus the object broker. It's not language- or system-dependant, and you shouldn't even have to care what language a class you inherit your class from is written in.
No, you're wrong about X. John Carmack certainly did port X to MacOS X Server, but a lot of that code -- including the DisplayPostscript that MacOS X Server used -- has gone away (X now runs on Darwin, though).
The real issue the other poster was driving home, though, was that porting to MacOS X in no implies "porting to UNIX," regardless of whether a X server exists for MacOS X.
No, it's not a 32bit DOS. It supports, from OS/2 1.0, a wealth of features not available in DOS until Windows95. Preemptive multitasking was there, multithreading, SMP support (OK, so that came in 2.x I think), plenty to differentiate it.
Ummm... no. As a matter of fact, the pretty interface showed up while Microsoft and IBM were working together on it. That IBM used that in their marketing of Warp as 'a better Windows than Windows' helped them split.
MacOSX Server has been released for a year, like I said, and it's solid. MacOS X isn't available yet, but I doubt they could ruin the stability enough to compare with NT in the interim.
Ooooh, more crap. Short-sighted, pointy-clicky "ease of use" is painful to intelligent people. Something like the text editor joe is a good example real ease of use: a steady learning curve that gets a person moving in five minutes, and lets them do almost anything in five hours. Good user interface is part of the design philosophy, not a stupid GUI that the programmers ask other people to develop.
First of all, MacOS X Server has been shipping for, I believe, a year now. If you regard Jobs as being Apple, then Jobs has been shipping preemptive multitasking since before OS/2 had a GUI (which is before Windows had preemptive multitasking too), and he's been working on bringing that back to Apple since '96.
Further, Apple hasn't delivered yet because they're more ambitious. They've gone through several attempts that simply didn't live up to their standards. If you want to look at places that innovate with backwards compatibility, look at Apple -- architecture change, no problem. Complete replacement of everything from the kernel to the userland, your old programs will still work. MacOS X is really innovative; it's more modern than UNIX, more stable than Windows NT or OS/2, and easier to use than anything.