If your SOC / CSIRT has an oncall policy you should have hired more people. Source: am manager at a CSIRT involved in nation-scale incidents. We have 24/7 coverage and nobody works overtime or is oncall.
I got my first Internet porn from Gopher. I had no access to Usenet then, but one University offered access to Usenet through Gopher and that included alt.* hierarchy and specifically alt.sex.* hierarchy.
If this is so easy, why anyone didn't do that before?
Not my chance, I'm not based in US, and the last two companies I worked for spoiled the chance by making everyone in IT incorporate then signing them up as subcontractor and not a salaried employee. Illegal as a hell, but when it comes out, both sides are guilty,and as a incorporated subcontractor you are not subject of labor laws and you cannot join/start an union.
This is one thing I can't understand as a person with 15 years of experience in the IT industry (various roles): why there aren't any programmers or sysadmins or general IT unions. Everyone thinks that he will outdo the unwashed average masses of the IT workers and hit it big in the next big start-up, or what?
I just quit, after 10+ years in IT. I now do some consulting work, but mainly write technology stuff for a news site, and nobody calls me at three in the morning because a vowel crashed and won't rejoin the rest of the word cluster after the reboot, and the whole site won't load further in the browser. FSM, I really hated the RHCM stuff.
I miss the money, a little, but definitely not the stress, nor the anxiety.
Also, when at a party someone asks "what do you do for a living" answering "I'm a journalist" results in much more positive reactions (also, from the females of the species), than when I answered "I'm a senior sysadmin at [company]". And I can point a finger to the screen and say "I wrote this!", which was not possible in the case of Puppet deployment which took most of the year (along with day to day wrangling with operations stuff and RHCM lossage).
GPS is quite easy to distort. Yesterday I was unable to get a reading (hanheld unit) because of electrically-charged clouds - a storm was coming, and this is not the first time that it fails in such a weather.
I've seen this also when using GPS to track airplane flight I was on. The unit worked well during most of the flight, and then when the plane begun descent and went into a layer of thick, dark clouds, the signal was lost.
But I still don't understand why ban receivers. And in Europe it seems they ban only CD players and cellphones, the usual before flight spiel never mentioned radio receivers.
The ability to run 8-years old binaries provided without source is a very minor advantage. If this is all what BSDs provide, over linux, a pity. And I checked, all systems that I administer don't have any old binaries, nor have any systems in general I know of. Commercial applications are certified to run on specific set of libraries anyway.
I think one could avoid the libc problems using some options of ld, and linux supports versioned symbols in libraries, but checking this would be a long and boring process. I won't check it unless someone pays my consulting fee to do it.
As for the rest: ability to run the very rare and usually not needed 8-years old binary is not an advantage when you have to spend five times more on routine system maintenance. This is what makes userland superior - convenience and ease of usue for the users and sysadmins. And ALL BSDs suck to the great extents in this departament (again, I'm NOT talking about running some GUIs). Read the article I referenced http://quiston.tpsa.com/antirant.txt to see what I'm talking about.
If one admin can support 2-3 BSD machines, and 20-30 linuxes, the choice os obvious, no matter if you could run your grand-grandfather's Difference Engine card stacks on it.
Well IF you OWERWRITTEN the system libc,by hand, with another version, thats asking for trouble. And I can't see a way you should do this. If you want to have another version of lib*, and necessarily put in by hand, yoou put it in/usr/local and dynamic loader gives the program the one it needs. If you upgrade libc via a system upgrade, all apps that reference the given ABI get upgraded too.
Also, your constant references to Debian are quite confusing, since Debian is one of the most conservative Linux distributions. Unless you were running another Debian version than stable in which case you should know what you are doing and you get all troubles you asked for. Debian stable has the all features you describe in BSD, and more, like backported security fixes (while all BSDs I know of offer them as upgrades to program versions). Debian stable is -STABLE, debian testing is beta and debian unstable is alpha.
You lost me completely on the tar example.
As for integration, ports routinely lack the ability to, in case of daemons auomatically include package invocation in/etc/rc* on package installation.
As for your Obviously you can accomplish these sorts of things using Linux, but I have yet to find a linux distribution that had done and packaged up all the work already. i have to disappoint you, debian has all this stuff and fedora core too. Automated is: installing python modules, emacs packages, info files, cron jobs, boot time startups, and more, there is unified configuration system across all packages: if a package needs a particular information, there is a unified way to ask the user for it (or not, if the user wishes, this is a configuration option of configuration system).
Again, lack of a decent installation system in BSD is only the fault of BSD developers (who aren't interested in programming a decent userland).
As for choice, when you install a major distribution (Debian, SuSE, Redhat), you get one syslogger and one default desktop system. You may get and enable the alternatives later, but at the beginning there is one of a kind of each.
Your experience seems to be gentoo based, and gentoo is no major, nor a serious distribution.
You tried to hand-compile something on linux, while you are comparing this to getting something from ports. You are comparing stuff from the wild, with prepackaged one. I didn't say that there is no packaging on BSDs, What I said is that ports are poor packaging system.
As for your troubles with pm3 installation: this is what you get when you want to use bleeding edge. I personally can't remember when I did compile some piece of software I'm not working on myself. I exclusively use precompiled packages, This way I can comcentrate on doing important stuff, while I don't have to waste my time on uninteresting things, like - for me - pm3 installation. I Would simply apt-get it then use it. The same as for your usage of ports.
As for rest of your comments:
1: This is not very interesting from the real world point of view (i.e. you can't make money out of it) isn't it?
2: All major Linux distributions give me this.
3: I don't see any way of tools integration in BSD, care to elaborate?
3a: this is not an argument
3b: The packaging framework is the core of an OS. I install the core, then I decide which role the system has to fullfill then I install the tools needed. There are 10000+ software packages in Debian, there is no sense in instaling all of those. There is no sense in having a compiler on DNS server. There is no sense in having Samba on database server, and so on and so on.
You probably didn't have securelevels enabled. Am I right?
Well, latest freebsd migration documentation is 9 screens long in my browser, twice as long as the one for openbsd.
Yes, in debian everything is a package. All I can say, is that it is really not my fault that great hackers working on FreeBSD couldn't make a decent packaging system. APT is a extendable system and probably could be adapted with less effort than making a whole new solution.
But again, this supports my original thesis, which is that BSD people aren't really interested in making a decent operating system, because a software management system is a essential part of one. But everyone wants to hack on a kernel it seems. A pity.
And (freebsd's) nineteen steps upgrade algorithm for a modern operating system in 21st century is simply [expletive redacted] ridiculous.
Oh, there is plenty. Package management first comes to mind. Ports are a poor imitation of one.
Taking responsibility for the whole system and not only tiny handful of commands that are enough to make the system a DNS server or a firewall, but nothing more (almost all useful - in the real world sense - programs are in ports which aren't officially supported).
You cannot, you need to reboot to runlevel 0 to do make installworld, and this can be done only from the console. Of course you can disble runlevels but this defeats the whole purpose of it.
And this will almost for sure overwrite some config files.
If the upgrades would be the way I like, I wouldn't have to UNINSTALL applicaction packages before doing an upgrade, as this page suggests. I wouldn't have to add system users by hand.
To upgrade a Debian system to a next release, all I have to do is modify ONE config file, do
apt-get update
,
apt-get dist-upgrade
, then answer a few questions and do a reboot. Description BSD way of system upgrade is FOUR pages long...
This is the userland I'm talking of, not firefox.
BSD way of doing things is fine as long you have one machine to take care of. If you have a server farm to upgrade, good luck.
And yes, I did manage BSD servers for a living. Now I wouldn't touch them with a four meter pole.
is not the kernel, which is nice, but the complete lack of *BSD developers' interest in the outside world. The systems lack a good UI, the distributions lack a real support for more than a very limited core system.
A year ago I wrote a short article on this subject (as a response to some BSD fanatic rant posted to/.). The antirant can be read here.
The BSDs will be a good and interesting Operating Systems, when someone will strip them of the whole Unix legacy and make a GNU/BSD distribution. Especially when Debian/BSD will produce something useful.
The interview is another sign of this problem - Matt Dillon concentrates on kernel issues while forgetting the Real World(TM). I still remember and respect Matt's work on AmigaOS, but his software was much more user-oriented then than it is now.
There is not much of overhead, ever for tiny project.
I use VC even for single file projects (in which case I use RCS) or very small projects I work on.
For my toolbox (mostly EMACS) using RCS is almost transparent - I visit file in read only mode, when I want to edit it, I check out it (C-v v), work on it, then close it (C-x C-q) and describe what I've done. I really wish it would work the same with CVS or subversion.
I do not this for VC itself, but it half for VC in case I screw up something bad so I can gest back to old version,half for backup facilities. is reassuring to know that the file you work on is not the only copy in existence and you may get some recent copy back in case you screw something up. For larger projects I use subversion which is CVS done right. If the projects grows over RCS stage I migrate it to subversion via CVS and cvs2svn.
Since I work on various machines, subversion also allows me to merge changes semi-painlessly and to keep local copies in sync.
And yes, I DO use documnt versioning in OpenOffice.
I do use Firefox on one workstation but this is because SeaMonkey is too heavy. And I use Mutt for mail anyway. Never liked GUI mail clients, and if I have to use one,,it will be Gnus on Emacs on X or Evolution.
Slashdot Mirror
The best european developers are not that keen to move to States, no matter type of visa.
So the proper course of action is to start development company in Europe to price undercut the US developers. Seems like a great business opportunity.
If your SOC / CSIRT has an oncall policy you should have hired more people. Source: am manager at a CSIRT involved in nation-scale incidents. We have 24/7 coverage and nobody works overtime or is oncall.
I got my first Internet porn from Gopher. I had no access to Usenet then, but one University offered access to Usenet through Gopher and that included alt.* hierarchy and specifically alt.sex.* hierarchy.
They're fighting Russian hackers and Russian, American and Israeli spies (APT-s). Read their reports.
If this is so easy, why anyone didn't do that before?
Not my chance, I'm not based in US, and the last two companies I worked for spoiled the chance by making everyone in IT incorporate then signing them up as subcontractor and not a salaried employee. Illegal as a hell, but when it comes out, both sides are guilty,and as a incorporated subcontractor you are not subject of labor laws and you cannot join/start an union.
This is one thing I can't understand as a person with 15 years of experience in the IT industry (various roles): why there aren't any programmers or sysadmins or general IT unions. Everyone thinks that he will outdo the unwashed average masses of the IT workers and hit it big in the next big start-up, or what?
I just quit, after 10+ years in IT. I now do some consulting work, but mainly write technology stuff for a news site, and nobody calls me at three in the morning because a vowel crashed and won't rejoin the rest of the word cluster after the reboot, and the whole site won't load further in the browser. FSM, I really hated the RHCM stuff.
I miss the money, a little, but definitely not the stress, nor the anxiety.
Also, when at a party someone asks "what do you do for a living" answering "I'm a journalist" results in much more positive reactions (also, from the females of the species), than when I answered "I'm a senior sysadmin at [company]". And I can point a finger to the screen and say "I wrote this!", which was not possible in the case of Puppet deployment which took most of the year (along with day to day wrangling with operations stuff and RHCM lossage).
So, there is life after "Microserfs", after all.
GPS is quite easy to distort. Yesterday I was unable to get a reading (hanheld unit) because of electrically-charged clouds - a storm was coming, and this is not the first time that it fails in such a weather.
I've seen this also when using GPS to track airplane flight I was on. The unit worked well during most of the flight, and then when the plane begun descent and went into a layer of thick, dark clouds, the signal was lost.
But I still don't understand why ban receivers. And in Europe it seems they ban only CD players and cellphones, the usual before flight spiel never mentioned radio receivers.
I think one could avoid the libc problems using some options of ld, and linux supports versioned symbols in libraries, but checking this would be a long and boring process. I won't check it unless someone pays my consulting fee to do it.
As for the rest: ability to run the very rare and usually not needed 8-years old binary is not an advantage when you have to spend five times more on routine system maintenance. This is what makes userland superior - convenience and ease of usue for the users and sysadmins. And ALL BSDs suck to the great extents in this departament (again, I'm NOT talking about running some GUIs).
Read the article I referenced http://quiston.tpsa.com/antirant.txt to see what I'm talking about.
If one admin can support 2-3 BSD machines, and 20-30 linuxes, the choice os obvious, no matter if you could run your grand-grandfather's Difference Engine card stacks on it.
Also, your constant references to Debian are quite confusing, since Debian is one of the most conservative Linux distributions. Unless you were running another Debian version than stable in which case you should know what you are doing and you get all troubles you asked for. Debian stable has the all features you describe in BSD, and more, like backported security fixes (while all BSDs I know of offer them as upgrades to program versions). Debian stable is -STABLE, debian testing is beta and debian unstable is alpha.
You lost me completely on the tar example.
As for integration, ports routinely lack the ability to, in case of daemons auomatically include package invocation in
As for your
Obviously you can accomplish these sorts of things using Linux, but I have yet to find a linux distribution that had done and packaged up all the work already. i have to disappoint you, debian has all this stuff and fedora core too. Automated is: installing python modules, emacs packages, info files, cron jobs, boot time startups, and more, there is unified configuration system across all packages: if a package needs a particular information, there is a unified way to ask the user for it (or not, if the user wishes, this is a configuration option of configuration system).
Again, lack of a decent installation system in BSD is only the fault of BSD developers (who aren't interested in programming a decent userland).
As for choice, when you install a major distribution (Debian, SuSE, Redhat), you get one syslogger and one default desktop system. You may get and enable the alternatives later, but at the beginning there is one of a kind of each.
Your experience seems to be gentoo based, and gentoo is no major, nor a serious distribution.
You tried to hand-compile something on linux, while you are comparing this to getting something from ports. You are comparing stuff from the wild, with prepackaged one. I didn't say that there is no packaging on BSDs, What I said is that ports are poor packaging system.
As for your troubles with pm3 installation: this is what you get when you want to use bleeding edge. I personally can't remember when I did compile some piece of software I'm not working on myself. I exclusively use precompiled packages, This way I can comcentrate on doing important stuff, while I don't have to waste my time on uninteresting things, like - for me - pm3 installation. I Would simply apt-get it then use it. The same as for your usage of ports.
As for rest of your comments:
1: This is not very interesting from the real world point of view (i.e. you can't make money out of it) isn't it?
2: All major Linux distributions give me this.
3: I don't see any way of tools integration in BSD, care to elaborate?
3a: this is not an argument
3b: The packaging framework is the core of an OS. I install the core, then I decide which role the system has to fullfill then I install the tools needed. There are 10000+ software packages in Debian, there is no sense in instaling all of those. There is no sense in having a compiler on DNS server. There is no sense in having Samba on database server, and so on and so on.
Well, latest freebsd migration documentation is 9 screens long in my browser, twice as long as the one for openbsd.
Yes, in debian everything is a package. All I can say, is that it is really not my fault that great hackers working on FreeBSD couldn't make a decent packaging system. APT is a extendable system and probably could be adapted with less effort than making a whole new solution.
But again, this supports my original thesis, which is that BSD people aren't really interested in making a decent operating system, because a software management system is a essential part of one. But everyone wants to hack on a kernel it seems. A pity.
And (freebsd's) nineteen steps upgrade algorithm for a modern operating system in 21st century is simply [expletive redacted] ridiculous.
Sending money to other countries and not by visa, is expensive. The rates start from about $10 per transfer.
If you aren't in supported country you cannot make a paypal account at all.
Oh, there is plenty. Package management first comes to mind. Ports are a poor imitation of one.
Taking responsibility for the whole system and not only tiny handful of commands that are enough to make the system a DNS server or a firewall, but nothing more (almost all useful - in the real world sense - programs are in ports which aren't officially supported).
And this will almost for sure overwrite some config files.
If the upgrades would be the way I like, I wouldn't have to UNINSTALL applicaction packages before doing an upgrade, as this page suggests. I wouldn't have to add system users by hand.
To upgrade a Debian system to a next release, all I have to do is modify ONE config file, do
, , then answer a few questions and do a reboot. Description BSD way of system upgrade is FOUR pages long...This is the userland I'm talking of, not firefox.
BSD way of doing things is fine as long you have one machine to take care of. If you have a server farm to upgrade, good luck.
And yes, I did manage BSD servers for a living. Now I wouldn't touch them with a four meter pole.
as for the rest, I'm not talking about GUIs. I'm talking about support for the whole system.
is not the kernel, which is nice, but the complete lack of *BSD developers' interest in the outside world. The systems lack a good UI, the distributions lack a real support for more than a very limited core system.
A year ago I wrote a short article on this subject (as a response to some BSD fanatic rant posted to /.). The antirant can be read here.
The BSDs will be a good and interesting Operating Systems, when someone will strip them of the whole Unix legacy and make a GNU/BSD distribution. Especially when Debian/BSD will produce something useful.
The interview is another sign of this problem - Matt Dillon concentrates on kernel issues while forgetting the Real World(TM). I still remember and respect Matt's work on AmigaOS, but his software was much more user-oriented then than it is now.
I use VC even for single file projects (in which case I use RCS) or very small projects I work on.
For my toolbox (mostly EMACS) using RCS is almost transparent - I visit file in read only mode, when I want to edit it, I check out it (C-v v), work on it, then close it (C-x C-q) and describe what I've done. I really wish it would work the same with CVS or subversion.
I do not this for VC itself, but it half for VC in case I screw up something bad so I can gest back to old version,half for backup facilities. is reassuring to know that the file you work on is not the only copy in existence and you may get some recent copy back in case you screw something up. For larger projects I use subversion which is CVS done right. If the projects grows over RCS stage I migrate it to subversion via CVS and cvs2svn.
Since I work on various machines, subversion also allows me to merge changes semi-painlessly and to keep local copies in sync.
And yes, I DO use documnt versioning in OpenOffice.
I do use Firefox on one workstation but this is because SeaMonkey is too heavy. And I use Mutt for mail anyway. Never liked GUI mail clients, and if I have to use one, ,it will be Gnus on Emacs on X or Evolution.
Do some Bluebox/Bluescreen stuff a'la Sky Captain & the world of Tomorrow. Also, some raygun.
Of course you can sell GPL software. But with sources, or providing source on request.
I would.
use w3cache.icm.edu.pl