With all this hype about XML and ubiquiteness of SQL, LDAP directories do not get the attention they deserve. How many of you have installed SQL-based authentication at your site, just to find out how limited a solution it is (maintain more than one database for all kinds of authentications, do you?). Not only does LDAP allow for a flexible hierarchical directory, it's also a standardized Internet protocol whereas SQL isn't. With LDAP, many applications work out of the box because it's a standard. Oh yeah, there's also the OSS server available at openldap.org.
Yes it does. I've tried Evolution before, and it doesn't even do so much to justify its bloat. Go ahead, mod this "flamebait", I don't care.
If you want to have an open product that's better than Outlook, at least make it leaner and more functional. And what's with the identical look and feel? Can't they spend some time in _better_ than Outlook look and feel design?
It highly depends on what you want to scale, and whether Linux can scale it well. For example, until recently Linux was limited to 32-bit filesystems and could only have 2GB size files max. If you look at the requirements of your particular problem, and Linux or a BSD variant matches or exceeds them (NetBSD for example has had 64-bit FS for a while that scaled to terabytes on 32 bit hardware), I don't see why you should spend much more money on a big corp. solution.
"If I want to install a useful system with X and FVWM to do Web browsing, check e-mail and log into remote UNIX boxen, all on a Pentium-90 with 16 MB RAM and a 600 GB hard drive, the ONLY current distribution good for the job is Slackware."
There are a bunch of less-known Linux distributions out there allowing the same. Of course, BSDs will do even better as far as size and memory is concerned. Their C library is considerably smaller, resulting in smaller executables (and possibly faster, which makes BSDs better candidates for older hardware). The package system is great to boot.
What are you going to do now, moderate this message "flamebate"? I am advocating both Linux and BSD. There are some things BSD does better.
It's important we do not participate in any techno wars. The time spent fighting can be used to enhance what we already have instead of competing with waste techonology like Windows. There's no desktop war, and there was none. There are _efforts_ to bring good desktop environments to _Unix_ instead of emulating Windows. Wars is MS' objective to eliminate other OS threats (Helloween documents?).
Imagine 20 people working on your project. How bug-free is this software now? 40 people, 100, 1000, and you now have MS quality product. The problem is the human brain. The more brains you have working on the project, the more difficult it is for every brain in the team to remember, understand, or agree with others' ideas. It's easier for smarter brains, but the formula is still the same. Also, very often one can confuse oneself easily especially when some languages' syntax encourages confusion (declarations and pointer arithmetic in C as an instance). There are cazillions of factors that allow for bugs to easily creep into the code base of unsuspecting programmers. It's writing, and as any other writing needs to be revised for errors.
Happenned to me two days ago. I had to find out just how fucking awful the tech support is. The support person couldn't even create a ticket because their own trouble ticket system was down. What a fucking joke! Their 24.11.49.1 gateway is down (my subnet). I fixed the problem myself. How to fix? Turn DHCP off, since it won't work anyway. Set the netmask to 255.0.0.0 and set your default route to some other gateway e.g. 24.11.50.1. Try it, it worked for me.
I've worked for PPPL a few years ago (hey guys!). They were mainly Sun and some DEC VMS at the time, but one sysadmin was running RedHat, and they were just thinking about setting an Intel Beowulf up. So things have progressed considerably, huh? The desktops were mostly Mac, with some places running NT (computer lab). PPPL is an interesting place to work for since they accumulated quite a bit of equipment over the years (VAX among others).
Seems like piss poor performance for a a web farm like this. Especially if they're serving just the static HTML pages. They probably would have benefited significantly from a more scalable web server like AOL server or the next version of Apache with the state-threaded module and faster disk arrays (write-back battery backed cache and plenty of it) for Apache and MySQL.
What's the big deal? He provides the source to the software allowing an administrator to modify the installation path. In qmail it's conf-qmail. In others, there are make files.
He does like to engage in flame wars and insults, which is not beneficial or constructive for anyone. He also wants his software to run the world IMHO (and OpenBSD is one way to spread his software), so he flames OpenBSD mailing lists for not accepting his software because he doesn't let them modify his software for distribution. He wants to be in control of the installation paths, and not the distributors of his software.
Debian has a nice way to deal with software not fitting their policy. In case of qmail the Debian package is simply a script to download and compile/install qmail. Problem solved.
Of course, if you want relational features, run relational database separately, OR run LDAP on top of your relational DB.
With all this hype about XML and ubiquiteness of SQL, LDAP directories do not get the attention they deserve. How many of you have installed SQL-based authentication at your site, just to find out how limited a solution it is (maintain more than one database for all kinds of authentications, do you?). Not only does LDAP allow for a flexible hierarchical directory, it's also a standardized Internet protocol whereas SQL isn't. With LDAP, many applications work out of the box because it's a standard. Oh yeah, there's also the OSS server available at openldap.org.
How is this offtopic?
What a bunch of slanderous HORSESHIT.
BIIIG deal. Any sysadmin worth his solt can do it.
Yes it does. I've tried Evolution before, and it doesn't even do so much to justify its bloat. Go ahead, mod this "flamebait", I don't care.
If you want to have an open product that's better than Outlook, at least make it leaner and more functional. And what's with the identical look and feel? Can't they spend some time in _better_ than Outlook look and feel design?
And just exactly what did RH do to make it better to run Oracle? Increased shmem limits? That's it?
It highly depends on what you want to scale, and whether Linux can scale it well. For example, until recently Linux was limited to 32-bit filesystems and could only have 2GB size files max. If you look at the requirements of your particular problem, and Linux or a BSD variant matches or exceeds them (NetBSD for example has had 64-bit FS for a while that scaled to terabytes on 32 bit hardware), I don't see why you should spend much more money on a big corp. solution.
You know, you CAN use 2.4.x kernels with 6.2. Which is exactly what I am doing without _any_ problems.
"If I want to install a useful system with X and FVWM to do Web browsing, check e-mail and log into remote UNIX boxen, all on a Pentium-90 with 16 MB RAM and a 600 GB hard drive, the ONLY current distribution good for the job is Slackware."
There are a bunch of less-known Linux distributions out there allowing the same. Of course, BSDs will do even better as far as size and memory is concerned. Their C library is considerably smaller, resulting in smaller executables (and possibly faster, which makes BSDs better candidates for older hardware). The package system is great to boot.
What are you going to do now, moderate this message "flamebate"? I am advocating both Linux and BSD. There are some things BSD does better.
Do not trust a business a community website like slashdot.
It's important we do not participate in any techno wars. The time spent fighting can be used to enhance what we already have instead of competing with waste techonology like Windows. There's no desktop war, and there was none. There are _efforts_ to bring good desktop environments to _Unix_ instead of emulating Windows. Wars is MS' objective to eliminate other OS threats (Helloween documents?).
Attempt to make this drivel effective in the light of the terrorist events.
The time factor can be reduced when the software bloat is kept to the minimum and modularity to the maximum.
Imagine 20 people working on your project. How bug-free is this software now? 40 people, 100, 1000, and you now have MS quality product. The problem is the human brain. The more brains you have working on the project, the more difficult it is for every brain in the team to remember, understand, or agree with others' ideas. It's easier for smarter brains, but the formula is still the same. Also, very often one can confuse oneself easily especially when some languages' syntax encourages confusion (declarations and pointer arithmetic in C as an instance). There are cazillions of factors that allow for bugs to easily creep into the code base of unsuspecting programmers. It's writing, and as any other writing needs to be revised for errors.
Abuse the game from crack.com had it's levels written in Lisp. And it was an awesome game too.
A good analogy could be sendmail vs. qmail as piston vs. McMaster rotary engines.
Happenned to me two days ago. I had to find out just how fucking awful the tech support is. The support person couldn't even create a ticket because their own trouble ticket system was down. What a fucking joke! Their 24.11.49.1 gateway is down (my subnet). I fixed the problem myself. How to fix? Turn DHCP off, since it won't work anyway. Set the netmask to 255.0.0.0 and set your default route to some other gateway e.g. 24.11.50.1. Try it, it worked for me.
I've worked for PPPL a few years ago (hey guys!). They were mainly Sun and some DEC VMS at the time, but one sysadmin was running RedHat, and they were just thinking about setting an Intel Beowulf up. So things have progressed considerably, huh? The desktops were mostly Mac, with some places running NT (computer lab). PPPL is an interesting place to work for since they accumulated quite a bit of equipment over the years (VAX among others).
I second that. Try ion http://www.students.tut.fi/~tuomov/ion/. You might never go back to desktop hogs again.
Not everything was inherited from BSD. TCP/IP implementation in Linux is not BSD based.
Seems like piss poor performance for a a web farm like this. Especially if they're serving just the static HTML pages. They probably would have benefited significantly from a more scalable web server like AOL server or the next version of Apache with the state-threaded module and faster disk arrays (write-back battery backed cache and plenty of it) for Apache and MySQL.
What happenned to project Monterrey?
What's the big deal? He provides the source to the software allowing an administrator to modify the installation path. In qmail it's conf-qmail. In others, there are make files.
He does like to engage in flame wars and insults, which is not beneficial or constructive for anyone. He also wants his software to run the world IMHO (and OpenBSD is one way to spread his software), so he flames OpenBSD mailing lists for not accepting his software because he doesn't let them modify his software for distribution. He wants to be in control of the installation paths, and not the distributors of his software.
Debian has a nice way to deal with software not fitting their policy. In case of qmail the Debian package is simply a script to download and compile/install qmail. Problem solved.
Yes, Python is by far a better language. Perl is much easier to write in than to understand written code. Maintainability just sucks.