::sigh:: If you want a complete in-house-tested Linux OS, and you don't want to be a beta tester, get something like Debian stable.
Yes, it's stale and out of date when compared with other distributions (e.g. Debian unstable), and yes, the default kernel is 2.2.19 (and when the current testing distribution becomes Debian 3.0, they're probably going to include a 2.2 kernel with that too). This is the price you pay for stability.
Here's the difference in philosophy:
MS: Only use the next version of Windows in-house until they're confident that it's nearly ready, then release it to beta testers, then fix bugs they report, then release when they're confident that it's ready.
Linux kernel team: Release N.N.NNpreN versions until it's nearly ready, then release a "final" version; let distributions and "advanced users" (i.e. those who compile their own kernels) decide where they want to be on the spectrum going from cutting-edge features to known reliability.
When do you think the Windows kernel developers last made major (i.e. non-bugfix) changes to, say, the released WinXP kernel, anyway? Probably quite a while back... and yet nobody complains about that, precisely because we never see the latest version until it's considered stable and used in Windows ZQ or something in a couple of years.
I did clerical work in a completely MS-dependent company over the summer. The database front-end I spent 90% of my time in was written in-house anyway (probably attached to an Access DB, but I didn't need to know that), so no advantage in me knowing MS software there.
Interestingly, their multi-line phone system ran on a Unix box (I don't know which Unix, could be Linux, or not), which they could access from a couple of retired, too-slow-for-Office Win95 PCs running Exceed (an X implementation for Windows). The staff there seemed to cope fine with what looked (to me) suspiciously like twm and Tk...
IMO, in an ideal world schools would have at least Macs, some sort of Unix, and Windows (yes, I know this is unrealistic from an admin point of view, I'm talking hypothetically here). It's easy to fall into the trap of "because it's not the same as Windows, it's wrong" if Windows is too ubiqutous. As I remember, when I was at primary and secondary school, we didn't get too confused moving from BBC Micros to Acorn Archimedes to Windows PCs (and yes, my secondary school did just about have all those in active use by pupils, simultaneously!)
If that's what it would take, do we want to "beat Microsoft"?
It sounds like you're saying "to beat a large organization which makes your choices for you," (a common criticism of MS), "we must become them".
In proprietary software, stuff that doesn't "win" disappears when its parent company goes bankrupt, or (if the parent company is big enough) is quietly taken away and shot. (Think MS Bob).
In free/open source software, there doesn't have to be a winner, and stuff that doesn't win survives anyway. You can say that (KDE|Gnome|Windowmaker|Blackbox) is better than (Gnome|Windowmaker|KDE|Ion) all you like, but I can still install and use any or all of them, as long as someone else (like the developer) values them enough to mirror them.
Similarly, Linux is way more popular than *BSD, but you can still install BSD if you want to. If this was proprietary consumer software, one of them would probably have killed off the other by now (probably BSD would have killed Linux, since they had a head start).
One of the things I value about OSS is that there is a choice. There isn't a single vast monolithic product like Windows or Visual Studio; OK, so some bits of a typical Linux desktop (XFree86 for example) don't have a lot of competition, but the more likely you are to make the decision on preference rather than technical grounds (desktop environments, apps, etc.), the more choice you have.
Also, the various open source projects don't have the incentive to kill each other by deliberate incompatibilities that commercial products do. At the moment I use Gnome for my desktop, task bar and console, Gnome Sawfish as my window manager, Konqueror (KDE) as my file manager and KATE (also KDE) to edit text files. The fact that I can do that says a lot about the amount of flexibility you can get from "fragmented" OSS.
No, this isn't the way to make an easy, consistent, don't-need-to-think consumer system. If anyone's going the right way towards doing that, it's distros like Red Hat and Mandrake, and I'm sure I speak for many people here when I say we-do-your-thinking-so-you-don't-have-to distros are the ones I'm least likely to install. Come on, I'm posting on a site subtitled "news for nerds" - I like to think occasionally... (plus, advice to the consumers: the pretty GUI will run out sometime, just like it does in Windows, and to get any further you'll need to use command line stuff or::gasp:: edit config files eventually; might as well get used to it).
OSS's strength is that all the distros are different (diversity!) and more or less compatible. At the same time, you can get the same tools and applications on any of them: I can't necessarily run Red Hat RPMs of, say, Gnome on my Debian system, but I can run the.deb versions, or versions compiled from source. It's still Gnome.
The Start menu is user interface. You need to be able to use it. (On my/usr-using Debian system, each desktop environment's Start-menu-equivalent is organised in a nice neat hierarchy.)
OTOH, if you have a package manager and a sensible $PATH,/usr/bin doesn't *need* to be user-readable (and if you don't find the package manager useful, what are you doing choosing a packaged distro?)
When I run, say, ping from a command line, I don't particularly care whether it's in/bin,/usr/bin,/usr/local/bin or wherever, as long as it runs and none of its required files are "owned" by two packages at once and liable to be deleted or overwritten by mistake (the Debian package installer, for one, specifically prevents you from installing two packages both claiming to own a given file).
Windows needs C:\Program Files (which is just like/opt) because each program installs itself, and because there aren't well-defined locations for many things (there is no FHS for Windows) so some applications are programmed to look for "./MyDataFile" rather than the Unixish/usr/share/programname/mydatafile.
The reasoning behind/usr and friends isn't based on the same assumptions as Windows; if you have many identical Unix boxes with NFS, you can give each its own/var and/tmp, share/home, possibly share/etc between boxes with the same hardware and configuration, share most of/usr (/usr/bin,/usr/lib, and so on) between boxes with the same hardware, and share/usr/share between everything (at least in theory).
You try doing that with large chunks of C:\Windows on anything except a network of clones. (No, wait, Windows doesn't have a Unix VFS, so you can't attach things to arbitrary points on the filesystem, you have to use drive letters... and you don't even get symlinks to make up for it...) Yes, this is redundant on a home PC; but so are many things. Many of Windows' various NetBIOS and related networking features are redundant on anything except a corporate network, but they're still there.
/usr/local is very useful on a package-managed system - it marks out part of the filesystem which the package management is not to mess with. Anything I do there won't get overwritten next time I upgrade a package, and vice versa. I wish Windows had something similar (oh, wait, Win98 has the Web View "no user-servicable parts inside" pages pasted across precisely the folders you often need to mess with to make things work). The only Linux systems I can think of where/usr/local is useless are (1) ones with broken packages which install into/usr/local anyway (yech), and (2) Linux From Scratch:-) (which is what you use if you're allergic to package managers).
As for/home/myname, it's just a multi-user version of C:\My Documents (you do use that or your own equivalent instead of saving Word documents in C:\Winword or whatever, right?), with the added advantage that it's far quicker to type (although you rarely need to type it anyway thanks to the shell's ~ shortcut) and doesn't have that command-line-tool-breaking space in it... When set up for multiple users, Windows uses C:\Windows\Profiles\myname, which is rather worse than/home/myname, and when set up for one user, it puts a lot of what would be "dotfiles" on Unix straight in the Windows folder (user.dat for example), making just-the-important-stuff backups difficult. Using a subdirectory/home is a nice compromise between cluttering the root directory (/myname) and obscurity (the closest equivalent of Windows profiles directories would probably be/usr/profiles/myname?)
There's still Debian, which has an open-source-only policy (it does package up some proprietary software, like Netscape, for convenience, but that's kept in a different section of the ftp archive and never gets put on the official CDs).
In the excessively unlikely event that Debian (or its owning organisation Software in the Public Interest) dies, there's always Linux From Scratch... (install some other Linux distro to get a platform capable of running gcc, compile a basic compiling environment on it from source, reboot into that, recompile everything from source... no, I haven't tried it, but I'm wondering whether to do so for my old slow computer)
(for those who don't know: there was a BBC radio show called I'm Sorry, I Haven't A Clue. Think of it as Monty Python on the radio and you won't go far wrong. Anyway, they had a game called Mornington Crescent, where you had to traverse London by naming Underground stations in no particular order. The rules, like which station you could get to from which other station, were never documented and had nothing to do with the actual geography or topology of London - instead random rules were occasionally invoked, like "under the Shrewsbury conventions, going from Bakerloo to Covent Garden on the seventeenth Thursday after a new moon results in your opponent being sent to Tower Hamlets, unless you have already been to Embankment, in which case you may only go to Covent Garden via Kings Cross" which of course changed constantly. The winner was the first to reach/say Mornington Crescent.)
Recent Microsoft OS sending random packets after 10 minutes' inactivity? Worrying...
I assume you're not running Seti@home or anything? Or spyware (software which spies on you/your Internet connection; basically any free Windows download manager or accelerator, and anything with built-in ads)?
Try downloading ZoneAlarm (http://www.zonelabs.com) and setting it to be as paranoid as possible. It tells you when stuff tries to access your LAN or the Internet (including which program it was, although some spyware uses Internet Explorer embedding to disguise its Internet use as coming from IE)
If that fails, you could install a packet filter on your Mac (I assume OS X must have some equivalent of Linux's ipchains and iptables?) and see where the packets are going...
Yes,/dev/null... (or "NUL:" if you're on Windows)
At the moment my firewall (iptables) logs and drops unwanted packets. I used it like this for about a couple of hours, looked at the log, then added an extra "don't log, just drop" rule specifically for NetBIOS...
Hardly surprising, since I'm on a LAN Internet connection full of poorly configured Windows Notworking hosts spamming each other (and potentially the entire Internet, since they use 255.255.255.255, although I suspect/hope my college filters NetBIOS out at the router...)
Re:No negative connotation in the US
on
Review: Harry Potter
·
· Score: 2, Informative
The word "philosopher" doesn't have any supernatural connotations here as far as I know (I'm British), but the Philosopher's Stone does (well, if you count alchemy as supernatural). It was (at least mythically) what medieval alchemists tried to find, and depending which version you read, it either turned lead into gold, or gave you eternal life (occasionally both).
They're probably in the same sort of situation I was for a while. I was looking for a cheap, basic web hosting package a while ago. Ironically, the cheapest I could find that did what I wanted ran on WinNT+IIS.
I eventually switched from Win98 to Linux, so I was uploading content from a Linux desktop to a Windows server... shouldn't that be the other way round?
[While I was hosted there, my site was "upgraded" to Win2k+IIS because the NT box started returning "403 denied" to every request, and after a week of being unable to fix it they decided to give up, copy everything to their new server, and reformat the old one while they waited for the DNS to refresh...:-( ]
I'm now paying twice as much for hosting with another company on Linux+Apache, which is bad from a financial point of view, but does mean I get a decent bandwidth quota (this is why I moved), SSI, CGI scripts, admins who seem to know what they're doing (always a nice feature), and immunity to Code Red and friends...
According to the FHS,/usr is for the distribution,/usr/local is for the local admin (you), and/opt is... er... pretty ambiguous, actually.
"Distributions may install software in/opt, but should not modify or delete software installed by the local system administrator without the assent of the local system administrator." -- the Filesystem Hierarchy Standard.
Many distros (Debian for one) don't use/opt, ever. Their rationale seems to be that if/usr/local is reserved specifically for the local admin, and the rest of/usr is available for the package maintainers to play with, then there's no ambiguity and everyone is happy.
It sounds to me as if distributions should install into/usr, but if you download (say) KDE or Gnome or StarOffice or whatever separately from your distro, then perhaps that should go in/opt - if, say, Debian doesn't have a package for a hypothetical office suite ObscureOffice, but I download a binary package in a.tar.bz2 or a Windows-style self-extracting executable or something direct from the publisher of ObscureOffice, it might be sensible for that to default to/opt.
the newest pacakges ALL the TIME (i.e, Debian/unstable)
I'd like both, please.
You can't have both, it just doesn't work like that. If you run new software, expect it to fall over occasionally. If you want stable software, you'll have to wait for the people who ran it when it was new software to find and report the bugs...
Come on, this is open source, it gives public beta testing a whole new meaning:-) Think of new Linux software as equivalent to what proprietary software companies' programmers are hacking away at right now, and slightly stale but very solid stuff like Debian stable, Linux kernels before 2.4, etc. as equivalent to what proprietary software companies have just released.
Having said that, Debian unstable is misnamed. stable, testing and unstable really mean "server/production quality", "it hasn't crashed for at least a fortnight * " and "it works for me".:-)
I've been running unstable for a couple of months, and the only problems I've had were a couple of Gnome apps randomly segfaulting on exit sometimes, until they were updated a week or two later (if you've used Windows you'll be used to that sort of thing, except you don't get the free upgrades), and the apt/dpkg dependencies sometimes breaking and preventing me from installing stuff (unstable is always changing so things are bound to be inconsistent sometimes; I fixed this in the end by downloading the Debianized package source for one of the packages involved and compiling my own.deb file for it. No changes needed, just recompiling on my system with my library versions fixed it).
[*: OK, I'm being over-simplistic, but no new bug reports for 2 weeks is one of the criteria for unstable packages going into testing.]
Actually, mp3.com sort of does this (were they a file trading site once? I've only ever known them as a site for unsigned bands). I've found some very good music there by random bands I'd never heard of.
They get round the "what to search for" bit by asking each artist they host for a list of "similar artists", which site users can then search.
On a Linux kernel, to mount a file system, you need root priviledges. No getting round that.
Ordinary users can mount certain filesystems because/bin/mount is suid root, so it runs with root's priviledges. The way it's implemented is that if the owner of/etc/fstab (i.e. root) has specifically marked a device/mount point combination as "user" or "users", ordinary users are allowed to mount it. On a desktop Linux box, removable drives (CD, floppy, etc.) are usually tagged as "user".
This is for two reasons - security (if J.Random User mounts a specially prepared disk with a suid root shell on it, you have trouble) and stopping users from annoying each other (if I decide I want to mount something over/usr/bin, anyone else using a multi-user system will be very annoyed). The "user" marking automatically switches off certain permissions, to avoid the security hole mentioned, while the device can only be mounted where root said it should go (e.g./cdrom) so nobody can mess with the important bits of the filesystem.
AFAIK, in Hurd a user can mount something over/usr/bin if they so wish, and it only affects them - everyone else sees the "real"/usr/bin. This is cool, if potentially rather confusing, because a user can arbitrarily rearrange the filesystem how they want it (as long as they still have access to/bin/mount and/bin/umount:-) without inconveniencing anyone else.
Epic Games' Unreal Tournament and id Software's Quake 3 were both ported to various consoles recently. One of the target consoles had a keyboard and a mouse as an option (so you could get Internet access on it). One of the game companies (I think it was Epic, but I could be wrong) decided to implement PC-style keyboard/mouse controls (the usual setup is W/A/S/D for movement, mouse for aiming and shooting) for those who had the right hardware.
The other (probably id Software, but again, I could be wrong) only included support for the normal console gamepad. I've heard rumours that this was for fairness, because they watched some games of a beta and realised how easy the keyboard and mouse users found it to frag the poor deprived people with gamepads:-)
If you use decent software::cough::open standards::cough:: it really shouldn't determine the client OS. HTTP and TCP/IP are standard and open, so my Linux box with Mozilla can access Win2k/IIS-hosted sites and Windows/IE can access the contents of Linux/Apache servers; but if you insist on the latest version of NetBIOS over NetBEUI and Windows Notworking, you might get problems.
(yes, I know they're not designed for the same things, but more people have heard of HTTP than NFS or whatever the open equivalent of NetBIOS is)
7. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all.
Good, it's not just me with an nVidia card and occasional X lockups then...
Another strategy is to set up getty on your serial port, then when it crashes, plug another computer into your serial port (my old Psion 3c palmtop works wonderfully for this since it has a 56k serial port, and a simple terminal emulator in ROM), log in and either chvt to a normal console (e.g. chvt 1) or just start running ps and kill from the palmtop. Last time this happened I managed to recover without leaving X by killing the X client I was running at the time (FSV).
Critically, because X is nothing to do with the kernel (Microsoft please note, this is a good idea), the instances of getty, bash, ps, grep and kill running on the serial port still all work fine.
Yep, anything other than plain text in an e-mail needs Base64 encoding - instant 33% extra bytes. I sometimes wish the user-friendly hide-everything-from-the-user mail clients just displayed the actual raw message - perhaps then people would realise how huge their attachments were...
The Community Education (adult education, evening classes etc.) bit of my local secondary school (for non-Brits: school for kids aged 11-16) had a.com address - but the area they're useful to extends about 10 miles. They don't even really have a *regional* reach...
As you say, many US commercial sites (those whose products are something physical rather than information) don't really have much usefulness outside the USA. Anyone remember when bbc.co.uk was the (very much international) British Broadcasting Corporation, but bbc.com was something like Baltimore Business Computers Inc.?
The company I worked for over the summer is basically all-MS, but MS aren't making money out of them at the moment. The branch I was in use Win95 + Office 97 workstations and probably NT4 or Win2k on their servers (I was just doing data entry though, so I didn't go near the servers). I'm not sure, but I'd guess that the central office use 95, 98 and MacOS workstations and a mixture of Windows and Unix servers. The whole system is a bit wobbly, but the same's probably true of any large network...
Do not underestimate the naiveness of the average user.
Re:Its the squeeky wheel that gets the most attent
on
Interview With Linus
·
· Score: 1
Here's what I've used, if it's any help:
Mandrake - friendly, commercial, pre-packaged Linux
Good: Easy to install, has more of a Windowsish "just works" effect - automatically installs KDE and looks very friendly. Uses RPMs so you'll find pre-packaged copies of most software.
Bad: When you run out of GUI tools, you get very confused; you'll have to use the console eventually. Everyone and their dog use RPMs in their distro, so you'll find lots of RPMs with dependencies that don't make sense because they're depending on someone else's distro.
Debian - free distro developed by a non-profit organisation
Bad: Going from initial install to fully working X+Gnome or X+KDE system takes a while and involves editing of config files (at least, it did for me). Forces you to read man pages/HOWTOs/documentation. Only Debian use the.deb format, so you won't find many third party.debs (so you'll need to install from source into/usr/local if your chosen software isn't in Debian already). You might find it too hackerish and not Windowsish enough.
Good: Very easy package updating (if you have a fast Internet connection). Only Debian use the.deb format, so all the dependencies and stuff are consistent in pretty much any.deb you find. Forces you to read man pages/HOWTOs/documentation, so you understand stuff rather than clicking on things until it works. You have a choice of stable (out of date but extremely reliable), testing (not quite cutting edge) or unstable (what's just been released). Quake 3 worked better for me on Debian unstable than it did on Windows;-) although that's probably just coincidence.
At the moment I use Debian unstable. Draw your own conclusions...
Last time I used Mandrake, I got everything working fairly quickly by the traditional Windows method of "click on stuff until it works". My system was probably completely insecure, I had no idea how it worked, and I was completely lost where the GUI tools stopped (anyone else notice they stop just short of what you want to do with them?) - but it worked as a desktop OS.
More recently I installed Debian. It took about a week to get X working - but the mass of command-line tools and config files I needed to use from the command line forced me to read man pages, install Lynx and read the LDP Howtos. I understand what's going on a _lot_ better now as a result.
A distro with sensible defaults probably shouldn't need that much tweaking, anyway. The bits of Debian I've reconfigured significantly without using GUI tools:
XFree86 (because I need X to get to the GUI tools:-). This could do with a wizard interface, but I _am_ using proprietary closed-source nVidia drivers which aren't part of the standard X distribution (or indeed included in Debian), so I only have myself to blame here. Having said that, the Debian packages which download and compile the nVidia drivers are great.
Firewalling. This is something you want to get right. OK, your average ZoneAlarm user would be pretty lost setting up an iptables script, but you could probably throw a couple of samples in to the iptables package and have as good a firewall as a Windows one-firewall-fits-all.
Apache. This isn't a desktop app. Enough said.
Exim (the default Debian MTA - think Sendmail, but with a smaller config file). Ditto, and not _that_ critical - I could be using Mozilla Mail or setting my e-mail client to relay straight to my mail server, but I prefer to have mail sent in the background.
The kernel. If I wanted a true Windowsish desktop OS I'd still be using Debian's everything-as-a-module kernel (and when was the last time you changed your Windows kernel? Switching from 3.1 to 95 or 95/98/ME to NT/2k/XP, I suspect). make menuconfig isn't bad anyway, actually.
Precisely. Open protocols give you a choice. If you're reading reasonably well written HTML transferred from a working HTTP server, it shouldn't matter whether you're using Mozilla or IE or Opera or Lynx or... I wish there was a standard, stable word processing format, because complete compatibility between AbiWord/KWord/OpenOffice/etc. would be cool. Especially if MS eventually included filters for the format in Word:-) (not going to happen though)
As for the "winner takes all" model, something Microsoft didn't realise is that open source means a project can build on another project's work without the second project having died. Because *BSD is free and open source, Linux developers can and do use BSD code even though BSD is still available. In the Microsoft world, on the other hand, to do that sort of thing you have to either buy the company whose code you want, buy licensing for that code, or kill the company and buy their code cheaper.
Slashdot Mirror
::sigh:: If you want a complete in-house-tested Linux OS, and you don't want to be a beta tester, get something like Debian stable.
Yes, it's stale and out of date when compared with other distributions (e.g. Debian unstable), and yes, the default kernel is 2.2.19 (and when the current testing distribution becomes Debian 3.0, they're probably going to include a 2.2 kernel with that too). This is the price you pay for stability.
Here's the difference in philosophy:
MS: Only use the next version of Windows in-house until they're confident that it's nearly ready, then release it to beta testers, then fix bugs they report, then release when they're confident that it's ready.
Linux kernel team: Release N.N.NNpreN versions until it's nearly ready, then release a "final" version; let distributions and "advanced users" (i.e. those who compile their own kernels) decide where they want to be on the spectrum going from cutting-edge features to known reliability.
When do you think the Windows kernel developers last made major (i.e. non-bugfix) changes to, say, the released WinXP kernel, anyway? Probably quite a while back... and yet nobody complains about that, precisely because we never see the latest version until it's considered stable and used in Windows ZQ or something in a couple of years.
I did clerical work in a completely MS-dependent company over the summer. The database front-end I spent 90% of my time in was written in-house anyway (probably attached to an Access DB, but I didn't need to know that), so no advantage in me knowing MS software there.
Interestingly, their multi-line phone system ran on a Unix box (I don't know which Unix, could be Linux, or not), which they could access from a couple of retired, too-slow-for-Office Win95 PCs running Exceed (an X implementation for Windows). The staff there seemed to cope fine with what looked (to me) suspiciously like twm and Tk...
IMO, in an ideal world schools would have at least Macs, some sort of Unix, and Windows (yes, I know this is unrealistic from an admin point of view, I'm talking hypothetically here). It's easy to fall into the trap of "because it's not the same as Windows, it's wrong" if Windows is too ubiqutous. As I remember, when I was at primary and secondary school, we didn't get too confused moving from BBC Micros to Acorn Archimedes to Windows PCs (and yes, my secondary school did just about have all those in active use by pupils, simultaneously!)
If that's what it would take, do we want to "beat Microsoft"?
::gasp:: edit config files eventually; might as well get used to it).
.deb versions, or versions compiled from source. It's still Gnome.
It sounds like you're saying "to beat a large organization which makes your choices for you," (a common criticism of MS), "we must become them".
In proprietary software, stuff that doesn't "win" disappears when its parent company goes bankrupt, or (if the parent company is big enough) is quietly taken away and shot. (Think MS Bob).
In free/open source software, there doesn't have to be a winner, and stuff that doesn't win survives anyway. You can say that (KDE|Gnome|Windowmaker|Blackbox) is better than (Gnome|Windowmaker|KDE|Ion) all you like, but I can still install and use any or all of them, as long as someone else (like the developer) values them enough to mirror them.
Similarly, Linux is way more popular than *BSD, but you can still install BSD if you want to. If this was proprietary consumer software, one of them would probably have killed off the other by now (probably BSD would have killed Linux, since they had a head start).
One of the things I value about OSS is that there is a choice. There isn't a single vast monolithic product like Windows or Visual Studio; OK, so some bits of a typical Linux desktop (XFree86 for example) don't have a lot of competition, but the more likely you are to make the decision on preference rather than technical grounds (desktop environments, apps, etc.), the more choice you have.
Also, the various open source projects don't have the incentive to kill each other by deliberate incompatibilities that commercial products do. At the moment I use Gnome for my desktop, task bar and console, Gnome Sawfish as my window manager, Konqueror (KDE) as my file manager and KATE (also KDE) to edit text files. The fact that I can do that says a lot about the amount of flexibility you can get from "fragmented" OSS.
No, this isn't the way to make an easy, consistent, don't-need-to-think consumer system. If anyone's going the right way towards doing that, it's distros like Red Hat and Mandrake, and I'm sure I speak for many people here when I say we-do-your-thinking-so-you-don't-have-to distros are the ones I'm least likely to install. Come on, I'm posting on a site subtitled "news for nerds" - I like to think occasionally... (plus, advice to the consumers: the pretty GUI will run out sometime, just like it does in Windows, and to get any further you'll need to use command line stuff or
OSS's strength is that all the distros are different (diversity!) and more or less compatible. At the same time, you can get the same tools and applications on any of them: I can't necessarily run Red Hat RPMs of, say, Gnome on my Debian system, but I can run the
The Start menu is user interface. You need to be able to use it. (On my /usr-using Debian system, each desktop environment's Start-menu-equivalent is organised in a nice neat hierarchy.)
/usr/bin doesn't *need* to be user-readable (and if you don't find the package manager useful, what are you doing choosing a packaged distro?)
/bin, /usr/bin, /usr/local/bin or wherever, as long as it runs and none of its required files are "owned" by two packages at once and liable to be deleted or overwritten by mistake (the Debian package installer, for one, specifically prevents you from installing two packages both claiming to own a given file).
/opt) because each program installs itself, and because there aren't well-defined locations for many things (there is no FHS for Windows) so some applications are programmed to look for "./MyDataFile" rather than the Unixish /usr/share/programname/mydatafile.
/usr and friends isn't based on the same assumptions as Windows; if you have many identical Unix boxes with NFS, you can give each its own /var and /tmp, share /home, possibly share /etc between boxes with the same hardware and configuration, share most of /usr (/usr/bin, /usr/lib, and so on) between boxes with the same hardware, and share /usr/share between everything (at least in theory).
/usr/local is useless are (1) ones with broken packages which install into /usr/local anyway (yech), and (2) Linux From Scratch :-) (which is what you use if you're allergic to package managers).
/home/myname, it's just a multi-user version of C:\My Documents (you do use that or your own equivalent instead of saving Word documents in C:\Winword or whatever, right?), with the added advantage that it's far quicker to type (although you rarely need to type it anyway thanks to the shell's ~ shortcut) and doesn't have that command-line-tool-breaking space in it... When set up for multiple users, Windows uses C:\Windows\Profiles\myname, which is rather worse than /home/myname, and when set up for one user, it puts a lot of what would be "dotfiles" on Unix straight in the Windows folder (user.dat for example), making just-the-important-stuff backups difficult. Using a subdirectory /home is a nice compromise between cluttering the root directory (/myname) and obscurity (the closest equivalent of Windows profiles directories would probably be /usr/profiles/myname?)
OTOH, if you have a package manager and a sensible $PATH,
When I run, say, ping from a command line, I don't particularly care whether it's in
Windows needs C:\Program Files (which is just like
The reasoning behind
You try doing that with large chunks of C:\Windows on anything except a network of clones. (No, wait, Windows doesn't have a Unix VFS, so you can't attach things to arbitrary points on the filesystem, you have to use drive letters... and you don't even get symlinks to make up for it...) Yes, this is redundant on a home PC; but so are many things. Many of Windows' various NetBIOS and related networking features are redundant on anything except a corporate network, but they're still there.
/usr/local is very useful on a package-managed system - it marks out part of the filesystem which the package management is not to mess with. Anything I do there won't get overwritten next time I upgrade a package, and vice versa. I wish Windows had something similar (oh, wait, Win98 has the Web View "no user-servicable parts inside" pages pasted across precisely the folders you often need to mess with to make things work). The only Linux systems I can think of where
As for
There's still Debian, which has an open-source-only policy (it does package up some proprietary software, like Netscape, for convenience, but that's kept in a different section of the ftp archive and never gets put on the official CDs).
In the excessively unlikely event that Debian (or its owning organisation Software in the Public Interest) dies, there's always Linux From Scratch... (install some other Linux distro to get a platform capable of running gcc, compile a basic compiling environment on it from source, reboot into that, recompile everything from source... no, I haven't tried it, but I'm wondering whether to do so for my old slow computer)
Sounds kind of like Mornington Crescent...
(for those who don't know: there was a BBC radio show called I'm Sorry, I Haven't A Clue. Think of it as Monty Python on the radio and you won't go far wrong. Anyway, they had a game called Mornington Crescent, where you had to traverse London by naming Underground stations in no particular order. The rules, like which station you could get to from which other station, were never documented and had nothing to do with the actual geography or topology of London - instead random rules were occasionally invoked, like "under the Shrewsbury conventions, going from Bakerloo to Covent Garden on the seventeenth Thursday after a new moon results in your opponent being sent to Tower Hamlets, unless you have already been to Embankment, in which case you may only go to Covent Garden via Kings Cross" which of course changed constantly. The winner was the first to reach/say Mornington Crescent.)
Recent Microsoft OS sending random packets after 10 minutes' inactivity? Worrying...
I assume you're not running Seti@home or anything? Or spyware (software which spies on you/your Internet connection; basically any free Windows download manager or accelerator, and anything with built-in ads)?
Try downloading ZoneAlarm (http://www.zonelabs.com) and setting it to be as paranoid as possible. It tells you when stuff tries to access your LAN or the Internet (including which program it was, although some spyware uses Internet Explorer embedding to disguise its Internet use as coming from IE)
If that fails, you could install a packet filter on your Mac (I assume OS X must have some equivalent of Linux's ipchains and iptables?) and see where the packets are going...
Yes, /dev/null... (or "NUL:" if you're on Windows)
At the moment my firewall (iptables) logs and drops unwanted packets. I used it like this for about a couple of hours, looked at the log, then added an extra "don't log, just drop" rule specifically for NetBIOS...
Hardly surprising, since I'm on a LAN Internet connection full of poorly configured Windows Notworking hosts spamming each other (and potentially the entire Internet, since they use 255.255.255.255, although I suspect/hope my college filters NetBIOS out at the router...)
The word "philosopher" doesn't have any supernatural connotations here as far as I know (I'm British), but the Philosopher's Stone does (well, if you count alchemy as supernatural). It was (at least mythically) what medieval alchemists tried to find, and depending which version you read, it either turned lead into gold, or gave you eternal life (occasionally both).
They're probably in the same sort of situation I was for a while. I was looking for a cheap, basic web hosting package a while ago. Ironically, the cheapest I could find that did what I wanted ran on WinNT+IIS.
:-( ]
I eventually switched from Win98 to Linux, so I was uploading content from a Linux desktop to a Windows server... shouldn't that be the other way round?
[While I was hosted there, my site was "upgraded" to Win2k+IIS because the NT box started returning "403 denied" to every request, and after a week of being unable to fix it they decided to give up, copy everything to their new server, and reformat the old one while they waited for the DNS to refresh...
I'm now paying twice as much for hosting with another company on Linux+Apache, which is bad from a financial point of view, but does mean I get a decent bandwidth quota (this is why I moved), SSI, CGI scripts, admins who seem to know what they're doing (always a nice feature), and immunity to Code Red and friends...
According to the FHS, /usr is for the distribution, /usr/local is for the local admin (you), and /opt is... er... pretty ambiguous, actually.
Many distros (Debian for one) don't use /opt, ever. Their rationale seems to be that if /usr/local is reserved specifically for the local admin, and the rest of /usr is available for the package maintainers to play with, then there's no ambiguity and everyone is happy.
It sounds to me as if distributions should install into /usr, but if you download (say) KDE or Gnome or StarOffice or whatever separately from your distro, then perhaps that should go in /opt - if, say, Debian doesn't have a package for a hypothetical office suite ObscureOffice, but I download a binary package in a .tar.bz2 or a Windows-style self-extracting executable or something direct from the publisher of ObscureOffice, it might be sensible for that to default to /opt.
I'd like both, please.
You can't have both, it just doesn't work like that. If you run new software, expect it to fall over occasionally. If you want stable software, you'll have to wait for the people who ran it when it was new software to find and report the bugs...
Come on, this is open source, it gives public beta testing a whole new meaning :-) Think of new Linux software as equivalent to what proprietary software companies' programmers are hacking away at right now, and slightly stale but very solid stuff like Debian stable, Linux kernels before 2.4, etc. as equivalent to what proprietary software companies have just released.
Having said that, Debian unstable is misnamed. stable, testing and unstable really mean "server/production quality", "it hasn't crashed for at least a fortnight * " and "it works for me". :-)
I've been running unstable for a couple of months, and the only problems I've had were a couple of Gnome apps randomly segfaulting on exit sometimes, until they were updated a week or two later (if you've used Windows you'll be used to that sort of thing, except you don't get the free upgrades), and the apt/dpkg dependencies sometimes breaking and preventing me from installing stuff (unstable is always changing so things are bound to be inconsistent sometimes; I fixed this in the end by downloading the Debianized package source for one of the packages involved and compiling my own .deb file for it. No changes needed, just recompiling on my system with my library versions fixed it).
[*: OK, I'm being over-simplistic, but no new bug reports for 2 weeks is one of the criteria for unstable packages going into testing.]
Actually, mp3.com sort of does this (were they a file trading site once? I've only ever known them as a site for unsigned bands). I've found some very good music there by random bands I'd never heard of.
They get round the "what to search for" bit by asking each artist they host for a list of "similar artists", which site users can then search.
On a Linux kernel, to mount a file system, you need root priviledges. No getting round that.
Ordinary users can mount certain filesystems because /bin/mount is suid root, so it runs with root's priviledges. The way it's implemented is that if the owner of /etc/fstab (i.e. root) has specifically marked a device/mount point combination as "user" or "users", ordinary users are allowed to mount it. On a desktop Linux box, removable drives (CD, floppy, etc.) are usually tagged as "user".
This is for two reasons - security (if J.Random User mounts a specially prepared disk with a suid root shell on it, you have trouble) and stopping users from annoying each other (if I decide I want to mount something over /usr/bin, anyone else using a multi-user system will be very annoyed). The "user" marking automatically switches off certain permissions, to avoid the security hole mentioned, while the device can only be mounted where root said it should go (e.g. /cdrom) so nobody can mess with the important bits of the filesystem.
AFAIK, in Hurd a user can mount something over /usr/bin if they so wish, and it only affects them - everyone else sees the "real" /usr/bin. This is cool, if potentially rather confusing, because a user can arbitrarily rearrange the filesystem how they want it (as long as they still have access to /bin/mount and /bin/umount :-) without inconveniencing anyone else.
Epic Games' Unreal Tournament and id Software's Quake 3 were both ported to various consoles recently. One of the target consoles had a keyboard and a mouse as an option (so you could get Internet access on it). One of the game companies (I think it was Epic, but I could be wrong) decided to implement PC-style keyboard/mouse controls (the usual setup is W/A/S/D for movement, mouse for aiming and shooting) for those who had the right hardware.
The other (probably id Software, but again, I could be wrong) only included support for the normal console gamepad. I've heard rumours that this was for fairness, because they watched some games of a beta and realised how easy the keyboard and mouse users found it to frag the poor deprived people with gamepads :-)
If you use decent software ::cough::open standards::cough:: it really shouldn't determine the client OS. HTTP and TCP/IP are standard and open, so my Linux box with Mozilla can access Win2k/IIS-hosted sites and Windows/IE can access the contents of Linux/Apache servers; but if you insist on the latest version of NetBIOS over NetBEUI and Windows Notworking, you might get problems.
(yes, I know they're not designed for the same things, but more people have heard of HTTP than NFS or whatever the open equivalent of NetBIOS is)
Good, it's not just me with an nVidia card and occasional X lockups then...
Another strategy is to set up getty on your serial port, then when it crashes, plug another computer into your serial port (my old Psion 3c palmtop works wonderfully for this since it has a 56k serial port, and a simple terminal emulator in ROM), log in and either chvt to a normal console (e.g. chvt 1) or just start running ps and kill from the palmtop. Last time this happened I managed to recover without leaving X by killing the X client I was running at the time (FSV).
Critically, because X is nothing to do with the kernel (Microsoft please note, this is a good idea), the instances of getty, bash, ps, grep and kill running on the serial port still all work fine.
Yep, anything other than plain text in an e-mail needs Base64 encoding - instant 33% extra bytes. I sometimes wish the user-friendly hide-everything-from-the-user mail clients just displayed the actual raw message - perhaps then people would realise how huge their attachments were...
The Community Education (adult education, evening classes etc.) bit of my local secondary school (for non-Brits: school for kids aged 11-16) had a .com address - but the area they're useful to extends about 10 miles. They don't even really have a *regional* reach...
As you say, many US commercial sites (those whose products are something physical rather than information) don't really have much usefulness outside the USA. Anyone remember when bbc.co.uk was the (very much international) British Broadcasting Corporation, but bbc.com was something like Baltimore Business Computers Inc.?
The company I worked for over the summer is basically all-MS, but MS aren't making money out of them at the moment. The branch I was in use Win95 + Office 97 workstations and probably NT4 or Win2k on their servers (I was just doing data entry though, so I didn't go near the servers). I'm not sure, but I'd guess that the central office use 95, 98 and MacOS workstations and a mixture of Windows and Unix servers. The whole system is a bit wobbly, but the same's probably true of any large network...
Do not underestimate the naiveness of the average user.
Here's what I've used, if it's any help:
Mandrake - friendly, commercial, pre-packaged Linux- Good: Easy to install, has more of a Windowsish "just works" effect - automatically installs KDE and looks very friendly. Uses RPMs so you'll find pre-packaged copies of most software.
- Bad: When you run out of GUI tools, you get very confused; you'll have to use the console eventually. Everyone and their dog use RPMs in their distro, so you'll find lots of RPMs with dependencies that don't make sense because they're depending on someone else's distro.
Debian - free distro developed by a non-profit organisationAt the moment I use Debian unstable. Draw your own conclusions...
Last time I used Mandrake, I got everything working fairly quickly by the traditional Windows method of "click on stuff until it works". My system was probably completely insecure, I had no idea how it worked, and I was completely lost where the GUI tools stopped (anyone else notice they stop just short of what you want to do with them?) - but it worked as a desktop OS.
More recently I installed Debian. It took about a week to get X working - but the mass of command-line tools and config files I needed to use from the command line forced me to read man pages, install Lynx and read the LDP Howtos. I understand what's going on a _lot_ better now as a result.
A distro with sensible defaults probably shouldn't need that much tweaking, anyway. The bits of Debian I've reconfigured significantly without using GUI tools:
Precisely. Open protocols give you a choice. If you're reading reasonably well written HTML transferred from a working HTTP server, it shouldn't matter whether you're using Mozilla or IE or Opera or Lynx or... I wish there was a standard, stable word processing format, because complete compatibility between AbiWord/KWord/OpenOffice/etc. would be cool. Especially if MS eventually included filters for the format in Word :-) (not going to happen though)
As for the "winner takes all" model, something Microsoft didn't realise is that open source means a project can build on another project's work without the second project having died. Because *BSD is free and open source, Linux developers can and do use BSD code even though BSD is still available. In the Microsoft world, on the other hand, to do that sort of thing you have to either buy the company whose code you want, buy licensing for that code, or kill the company and buy their code cheaper.