On Mac and Windows machines I select something, go somewhere else and hit "paste", then realize that I forgot to "copy" and have to go back and do it all over again. And don't even talk to me about the semantics of focus or window-stacking. And typing 'ls' in a command prompt gives me "invalid command or filename," what's this nonsense?? (I still can't get a command prompt on the Macintosh)
Clearly Macs and Windows machines are difficult and unintuitive to use.
Daniel
(for those who haven't figured it out, the above is..not sarcasm, but certainly not literal truth, since I can and do use Windows and Macintoshes, minus some back-and-forth flailing when I have to cut-and-paste;-) )
Now, anyone who reads my comments knows I'm about as anti-MS as anyone, but...
(just as a side note, it's kinda sad that people feel they have to make this sort of a disclaimer before going up against the Herd..er..Horde..)
Where the hell does this writer come off saying that NT 4.0 lacks a journaling file system?
I'm wondering if anyone can shed some light on this topic? Some people say it does, some people say it doesn't. The "it doesn't" crowd seems to have more empirical evidence, but if it doesn't than what *does* it mean when people say it's journaling? Microsoft doesn't usually flat-out *lie* (they just distort, embellish, stretch, and spin the truth), so I suspect they're talking about some filesystem feature that exists, even if it isn't really journaling.
OTOH, It fragments very badly, but so does ext2.
I'm also not clear on this:) I installed an ext[2]fs defragmenter but haven't used it yet, due to the fact that its documentation essentially says that in most cases you won't see more than a 10% speedup from defragmentation, and given that I don't know whether it'll eat my filesystem alive, I didn't think it was worth it;-)
I'm not sure about this being unequivacally bad. I choose my passwords via an algorithm generally known as the "bang randomly on the keys" approach, and usually *do* write them down for a day or two, until I memorize them. But while this certainly is a theoretical security problem, I don't see it as being important, because:
1) I keep them in my wallet, which is in my possession at all times 2) Even if someone stole the wallet -- no, let's be optimistic and say that someone found it lying on the sidewalk -- there's nothing marking that particular scrap of scribbled-upon paper as different from all the others I lug around; I don't label it "Super-Secret Computer Password" or anything. In addition, if the paper were lost just by itself, there's no way to know *whose* password it is, even assuming the leap in logic required to deduce that it's a password. 3) After two or three days, I tear the paper up into little bitty pieces and throw it out. Now the Evil Person not only has to figure out that this random scribbling is a password and figure out that it's *my* password, they also have to dig through the trash (which is of course mixed with all other trash from the region) and reconstruct the paper from fragments. If someone is determined and capable enough to do this, I'm not sure why they don't just install eavesdropping devices above my computer;-)
Ok, I very nearly said this on LinuxToday, but I've decided to be flamed by a wider audience;-)
Let's assume that the author's "nightmare" scenario comes to pass: IE suddenly becomes the dominant browser and all major websites decide to lock out users of non-IE browsers.
Will gcc's backends all suddenly commit electronic suicide in despair? Perhaps Emacs is going to decide it's Lisped its last? I assume that Linus, Alan, Andrea, and the thousands of other people who've worked on the kernel will pack up and leave. X, of course, will stop working when the Microsoft takeover alters the fabric of reality, as will FTP, SSH, and CVS. And of course every bleeding window manager will suddenly suffer a fatal heart attack (except wm2 which no-one uses anyway) apt will spontaneously combust, ext2fs will cease to function, the RFCs will self-destruct, Python will be a dead parrot, mpg123 will be silenced, GnomeI-See-You won't, and Freeciv as we know it will cease. All SMTP traffic will immediately halt; Exim, Postfix, and Smail will be outlawed; mutt and all lesser email clients[1] will start requiring stamps to be inserted into the floppy drive; and everything will be a general mess. Plus Microsoft will send people over to my house to demagnetize all my disks.
Yes, I can see how this will affect my life in a very significant and direct way.
Ok. Maybe I was still asleep when I heard the report and they said that the World was -expected- to vote against resignation (but hadn't yet). Or maybe (it pains me to say this:) ) NPR was wrong.
NPR reported this morning that while the advising team had voted to resign, the rest of the players voted to continue to the bitter end.
Daniel
Re:Support for multiple POP3 accounts?
on
Mutt Hits 1.0
·
· Score: 2
Is this possible?
Yes. Use fetchmail. Mailreaders aren't in the business of downloading mail; if they are it's a bug:)
Daniel
Re:Maybe somebody could help me with this...
on
Mutt Hits 1.0
·
· Score: 2
Personally, I think you're trying to do the wrong thing. Mail clients shouldn't be in the business of filtering mail; procmail works quite nicely for this and doesn't require you to manually move the mail (even if you just have to type 's' it's a pain when you get >100 messages a day)
(you may have to replace X-MailingList with some other header and .*foo@bar.org with some other regexp)
Daniel
Re:How to set up multiple mail folders in mutt?
on
Mutt Hits 1.0
·
· Score: 2
Mutt stores its mail folders as single mbox files. Any file (/var/spool/mail/bob, ~/mbox, ~/Mail/foo) can be opened this way; as a convenience, Mutt allows you to refer to a file in ~/Mail by '='. So ~/Mail/foo is =foo. Mutt also has a concept of 'mailboxes'; these are mail folders to which mail can be delivered (by the local MTA, procmail, whatever). You can list these with the 'mailboxes' command in ~/.muttrc. To switch which mailbox you're viewing, type 'c'. mutt will default to switching to the next mailbox with new mail, but you can type anything. Tab completion also works, and if you hit 'tab' twice it will move to a file browser. 'tab' in the file browser shows an overview of your defined mailboxes, with information about which ones have new mail.
I suggest you read the Mutt manual before asking any more questions, since it explains everything that I just said.
Mutt also uses SIGWINCH and resize_term(), which are not found in all systems/curses implementations. I don't know if you've heard of it, but there's a useful little utility called autoconf. It's used by mutt; you might want to read up on it.
Really? If I were to go by the article posted here I'd have to conclude that Echelon exists mainly in the paranoid fantasies of gun-toting militiamen out in Montana. In fact, the word 'militia' was where I stopped reading so I can't tell you much about the rest of the coverage.
Actually, I believe that Debian already enables IPv6 in many programs (sometimes with patches). I haven't tried this myself, though, so I don't know how well it works.
How do they come up with this number? The only explanation I can see is that they've built a machine to travel to a parallel universe in which piracy never occurs, and discovered that there are 18,000 more jobs in it. Or for a more mundane explanation: there are three types of lies...
From the Debian Social Contract (emphasis mine): 4. Our Priorities are Our Users and Free Software. We will be guided by the needs of our users and the free-software community. We will place their interests first in our priorities. We will support the needs of our users for operation in many different kinds of computing environment. We won't object to commercial software that is intended to run on Debian systems, and we'll allow others to create value-added distributions containing both Debian and commercial software, without any fee from us. To support these goals, we will provide an integrated system of high-quality, 100% free software, with no legal restrictions that would prevent these kinds of use.
Everything that's officially in Debian is free software, and (with the exception of ssh, grr..) you can usefully run a Debian system without ever touching non-free software. But that doesn't stop people from packaging useful non-free software -- in fact, the Debian ftp archives contain some (in the non-free/ section) as a service to the users.
No it's not. At least, I hope not; there's no good alternative yet (console-apt is not and may never be; I've looked at the code and that's all I'll say on the subject:/ )
Debian has a 2 disk FULL install
Only if you have a good way to get the base system (ie, a local NFS machine or a DOS partition). There was just a discussion on debian-devel about getting base via FTP -- it looks like it may go into potato.
it can install via PPP
But not [easily] on a network where you need dhcp. RedHat handles this nicely. Again, there are noises being made about fixing this in 2.2 but I don't know for sure what's going on. (now, if I can just get debian-boot-request to handle my subscription messages properly..:) ) Also, I'm not sure that ppp is in base2_1.tgz. Is it? I vaguely remember having to sneakernet the ppp deb last summer when I had cause to install it on a new system.
There are lots of good things about Debian, but please get your facts straight:)
you can't possibly tell me E0.16 will be in potato
It already is:)
Commonness of packages
I think you should be aware that (as opposed to RPMs), most Debian packages are distributed by Debian itself. This means that just because you don't see a package listed on Freshmeat or on the homepage doesn't mean you can assume that no package exists, or that the version reported is correct. And unless you need to have the version released last week, unstable is generally fine -- note I said generally, some *specific* maintainers have been slower about keeping up with releases. This is generally with big and complex pieces of software (eg, XFree) in which multiple patches and fixes have to be backed out or put back in inbetween upstream releases, and which require coordination between multiple developers. PHP3, Mysql, and Apache were significantly behind for a while last spring, I'm not sure what the status is now (I haven't tried using them recently)
There are very few programs which I haven't been able to get as Debian packages; for those which I can't, I usually build a package myself. For most software (read: software which uses autoconf and automake) this takes about 5 minutes+time to compile. The only recent examples were squaroid, xarchon, and sawmill (which I'm too lazy to download:P ) -- of these, xarchon and sawmill are going into potato RSN. squaroid I don't know about, but I didn't find it to be that interesting a game, so.. *shrug*
I think that although many people have many opinions -- probably someone will say that "it's too ugly" or "dselect sucks!", there is one and only one problem: the entire lack of a package hierarchy. This problem is peculiar to Debian since no other distribution actually lets the user list every program which could potentially be installed on his or her system, but that's no excuse for not solving it:) The current system provides a few broad groupings of packages (net/, devel/, etc) with no strong policy [that I know of] about where packages go -- libraries can be found in every section (for example, libmagick4g is in graphics/) and no consistency in package organization. This leads to two problems:
1) When setting up a system for the first time, the new user must examine each and every of the 4000 packages and determine which ones he/she wants to install. There's no way to know ahead of time what the next section of packages holds, and so you have to examine it. (in some cases, libs/ and devel/ might be eliminatable)
2) There is no way to get a quick listing of all available software of type X, and no reliable way to do it (besides the needle-in-a-haystack approach). Because of this, installing software on an already-set-up system also requires a huge amount of work.
Both problems are exacerbated by the fact that dselect refuses to merge the different priorities and overarching sections (there's a case to be made for non-free and non-us, but separating Base, Optional, and Extra from one another by default is indefensible), meaning that there are actually *multiple* occurences of each section! So I actually have to search three or four versions of web/ and x11/ to find Web browsers. More annoyingly, these are not just problems for newbies, but also (IMO) hinderances for experienced users -- good organization of information is just a good idea.
I posted a message some time last summer on debian-devel about this but it was mostly ignored, so now I'm working on actually implementing a working system of tagging packages to put them into a logical hierarchy. (yes, another curses APT frontend) I just started, though, so I don't expect it to be able to do anything particularly exciting until after potato is released.
Daniel
PS - historically, the huge number of questions asked in the process of unpacking packages has been a problem. Either in potato or in potato+1 this will be solved in a surpassingly elegant way by debconf.
I have mixed feelings here. On the one hand, it's great that Loki is releasing this, and it looks to be a much better way of installing software than simply having a random perl script to unpack tars (VMware) or something.
On the other hand... I have a great aversion to installing random cruft on my system. I have an even greater aversion to letting a random program install random cruft on my system. Is there a separate file to run to uninstall each program? (looks like it, in fact you have to build a custom program for each app that needs to be uninstalled) So to delete stuff installed with this program I have to hunt down the uninstaller. Probably better to just install in/usr/local/stow and rm -rf the whole lot. And how do I track what I've installed once I install more than two or three? How do I update to the latest patchlevel? Download patches and run them? If I find a file and can't remember what program it belongs to, how do I check?
Loki would be doing a much greater service to themselves and the world by having the installer auto-generate packages from the generic package description. InstallShield-type stuff, where each program in the world ships with its own special installer, is an ugly and inelegant hack. The most redeeming quality this installer could have would be if it took command-line options to specify all the attributes of the install (eg, bindir=, datadir=, etc) and skip the interactive stuff -- this would allow halfway-useful packages to be made fairly easily. Ah well..
You can learn a lot by teaching yourself. A whole lot. However, there is a lot to learn. A whole lot. A whole lot more, in fact, than you can learn by teaching yourself.
Consider this: I can (and have) taught myself programming from assembler to C++; I have taught myself a fair deal about the internals and externals of UNIX systems; I have taught myself to use autoconf and automake. These things can be learned by doing and I believe I have learned them well.
Now: how do I teach myself linear algebra? How do I teach myself the fundamentals of quantum mechanics? How do I teach myself about optics? In all these cases, I wouldn't really even know where to begin, and if I did, simply learning one of them would consume so much time that I'd never start on the next. Quantum mechanics you can't even learn by doing, and linear algebra, while useful, is not a structure most people are likely to derive on their own. Fiddling with lenses might get you somewhere in optics, but not everyong is Newton.:-) I could go on: foreign languages, history of obscure places, etc, but I feel that these examples are enough.
This is what teachers, schools, and books are good for: the more theoretical and esoteric bits of learning that a student is not likely to pick up on his or her own. I really hate to use the word 'efficient' in connection with education, but the truth is that you can just learn a lot of things much faster from someone who knows them already than by just flailing around on your own.
I don't mean to disparage the value of hands-on experience with anything from a mathematical equation to a Mozart symphony -- this is indisputably important. It's unfortunate -- no, make that bad -- if you work with computers and can't grope your way to a solution when all else fails. But too many people in this forum and elsewhere have an arrogant "Edukashun? We don't need no steenkin' edukashun!" attitude, thinking that simply because they taught themselves to install Apache that they know everything and there's nothing anyone can teach them. This is generally nonsense; worse is the assumption that there's nothing anyone can teach anyone
, because in a [nominally] democratic society that attitude can end up hurting everyone
.
If you want to believe that what you can learn with your hands is the sum total of human knowledge -- fine. But don't limit other people to this.
One last thing -- I may not have learned to use computers in school (I assume you mean elementary and high school, college lets students choose what they want), but I certainly learned writing, mathematics, history and the basics of various sciences there. These, and the mental skills they encourage (eg, logical reasoning), not only help with computers but are also of paramount importance for any would-be self-schooler. Even if you feel that college is beneath you you will need these skills to effectively bootstrap any curriculum you choose.
Slashdot Mirror
On Mac and Windows machines I select something, go somewhere else and hit "paste", then realize that I forgot to "copy" and have to go back and do it all over again. And don't even talk to me about the semantics of focus or window-stacking. And typing 'ls' in a command prompt gives me "invalid command or filename," what's this nonsense?? (I still can't get a command prompt on the Macintosh)
;-) )
Clearly Macs and Windows machines are difficult and unintuitive to use.
Daniel
(for those who haven't figured it out, the above is..not sarcasm, but certainly not literal truth, since I can and do use Windows and Macintoshes, minus some back-and-forth flailing when I have to cut-and-paste
Now, anyone who reads my comments knows I'm about as anti-MS as anyone, but...
:) I installed an ext[2]fs defragmenter but haven't used it yet, due to the fact that its documentation essentially says that in most cases you won't see more than a 10% speedup from defragmentation, and given that I don't know whether it'll eat my filesystem alive, I didn't think it was worth it ;-)
(just as a side note, it's kinda sad that people feel they have to make this sort of a disclaimer before going up against the Herd..er..Horde..)
Where the hell does this writer come off saying that NT 4.0 lacks a journaling file system?
I'm wondering if anyone can shed some light on this topic? Some people say it does, some people say it doesn't. The "it doesn't" crowd seems to have more empirical evidence, but if it doesn't than what *does* it mean when people say it's journaling? Microsoft doesn't usually flat-out *lie* (they just distort, embellish, stretch, and spin the truth), so I suspect they're talking about some filesystem feature that exists, even if it isn't really journaling.
OTOH, It fragments very badly, but so does ext2.
I'm also not clear on this
Daniel
I'm not sure about this being unequivacally bad. I choose my passwords via an algorithm generally known as the "bang randomly on the keys" approach, and usually *do* write them down for a day or two, until I memorize them. But while this certainly is a theoretical security problem, I don't see it as being important, because:
;-)
1) I keep them in my wallet, which is in my possession at all times
2) Even if someone stole the wallet -- no, let's be optimistic and say that someone found it lying on the sidewalk -- there's nothing marking that particular scrap of scribbled-upon paper as different from all the others I lug around; I don't label it "Super-Secret Computer Password" or anything. In addition, if the paper were lost just by itself, there's no way to know *whose* password it is, even assuming the leap in logic required to deduce that it's a password.
3) After two or three days, I tear the paper up into little bitty pieces and throw it out. Now the Evil Person not only has to figure out that this random scribbling is a password and figure out that it's *my* password, they also have to dig through the trash (which is of course mixed with all other trash from the region) and reconstruct the paper from fragments. If someone is determined and capable enough to do this, I'm not sure why they don't just install eavesdropping devices above my computer
Daniel
Errr..except for numbers 4, 6, 7, and possibly 5, I think those have already happened. So you're safe. :-)
Daniel
Ok, I very nearly said this on LinuxToday, but I've decided to be flamed by a wider audience ;-)
Let's assume that the author's "nightmare" scenario comes to pass: IE suddenly becomes the dominant browser and all major websites decide to lock out users of non-IE browsers.
Will gcc's backends all suddenly commit electronic suicide in despair? Perhaps Emacs is going to decide it's Lisped its last? I assume that Linus, Alan, Andrea, and the thousands of other people who've worked on the kernel will pack up and leave. X, of course, will stop working when the Microsoft takeover alters the fabric of reality, as will FTP, SSH, and CVS. And of course every bleeding window manager will suddenly suffer a fatal heart attack (except wm2 which no-one uses anyway) apt will spontaneously combust, ext2fs will cease to function, the RFCs will self-destruct, Python will be a dead parrot, mpg123 will be silenced, GnomeI-See-You won't, and Freeciv as we know it will cease. All SMTP traffic will immediately halt; Exim, Postfix, and Smail will be outlawed; mutt and all lesser email clients[1] will start requiring stamps to be inserted into the floppy drive; and everything will be a general mess. Plus Microsoft will send people over to my house to demagnetize all my disks.
Yes, I can see how this will affect my life in a very significant and direct way.
Daniel
[1] Everything else.
And what if someone came into an Athiests Anonymous meeting (or whatever) and stated that God created man?
As Yoda said..No Different! Only different in your mind! You must unlearn what you have learned!
Daniel
(ech, I think it's too late at night for this)
Probably because the ones who aren't 'wacked-out' are open-minded enough to live and let live..
Daniel
Ok. Maybe I was still asleep when I heard the report and they said that the World was -expected- to vote against resignation (but hadn't yet). Or maybe (it pains me to say this :) ) NPR was wrong.
Daniel
NPR reported this morning that while the advising team had voted to resign, the rest of the players voted to continue to the bitter end.
Daniel
Is this possible?
:)
Yes. Use fetchmail. Mailreaders aren't in the business of downloading mail; if they are it's a bug
Daniel
Personally, I think you're trying to do the wrong thing. Mail clients shouldn't be in the business of filtering mail; procmail works quite nicely for this and doesn't require you to manually move the mail (even if you just have to type 's' it's a pain when you get >100 messages a day)
.*foo@bar.org with some other regexp)
Most lists can be procmailed with something like:
PATH=/usr/local/bin:/usr/bin:/bin
MAILDIR=$HOME/Mail
DEFAULT=/var/spool/mail/bob
LOGFILE=$HOME/Mail/procmail.log
:0:
* X-MailingList:.*foo@bar.org
foo
(you may have to replace X-MailingList with some other header and
Daniel
Mutt stores its mail folders as single mbox files. Any file (/var/spool/mail/bob, ~/mbox, ~/Mail/foo) can be opened this way; as a convenience, Mutt allows you to refer to a file in ~/Mail by '='. So ~/Mail/foo is =foo. Mutt also has a concept of 'mailboxes'; these are mail folders to which mail can be delivered (by the local MTA, procmail, whatever). You can list these with the 'mailboxes' command in ~/.muttrc.
To switch which mailbox you're viewing, type 'c'. mutt will default to switching to the next mailbox with new mail, but you can type anything. Tab completion also works, and if you hit 'tab' twice it will move to a file browser. 'tab' in the file browser shows an overview of your defined mailboxes, with information about which ones have new mail.
I suggest you read the Mutt manual before asking any more questions, since it explains everything that I just said.
Daniel
Mutt also uses SIGWINCH and resize_term(), which are not found in all systems/curses implementations. I don't know if you've heard of it, but there's a useful little utility called autoconf. It's used by mutt; you might want to read up on it.
Daniel
Really? If I were to go by the article posted here I'd have to conclude that Echelon exists mainly in the paranoid fantasies of gun-toting militiamen out in Montana. In fact, the word 'militia' was where I stopped reading so I can't tell you much about the rest of the coverage.
Daniel
Excuse me, but Windows and MacOS *do* constitute a majority of systems.
Unfortunately.
Daniel
Actually, I believe that Debian already enables IPv6 in many programs (sometimes with patches). I haven't tried this myself, though, so I don't know how well it works.
Daniel
As I said: there are three types of lies.. :)
Daniel
How do they come up with this number? The only explanation I can see is that they've built a machine to travel to a parallel universe in which piracy never occurs, and discovered that there are 18,000 more jobs in it.
Or for a more mundane explanation: there are three types of lies...
Daniel
From the Debian Social Contract (emphasis mine):
4. Our Priorities are Our Users and Free Software.
We will be guided by the needs of our users and the free-software
community. We will place their interests first in our priorities. We
will support the needs of our users for operation in many different
kinds of computing environment. We won't object to commercial software
that is intended to run on Debian systems, and we'll allow others to
create value-added distributions containing both Debian and commercial
software, without any fee from us. To support these goals, we will
provide an integrated system of high-quality, 100% free software, with
no legal restrictions that would prevent these kinds of use.
Everything that's officially in Debian is free software, and (with the exception of ssh, grr..) you can usefully run a Debian system without ever touching non-free software. But that doesn't stop people from packaging useful non-free software -- in fact, the Debian ftp archives contain some (in the non-free/ section) as a service to the users.
Daniel
Debian 2.2 is going to replace dselect
:/ )
:) )
:)
No it's not. At least, I hope not; there's no good alternative yet (console-apt is not and may never be; I've looked at the code and that's all I'll say on the subject
Debian has a 2 disk FULL install
Only if you have a good way to get the base system (ie, a local NFS machine or a DOS partition). There was just a discussion on debian-devel about getting base via FTP -- it looks like it may go into potato.
it can install via PPP
But not [easily] on a network where you need dhcp. RedHat handles this nicely. Again, there are noises being made about fixing this in 2.2 but I don't know for sure what's going on. (now, if I can just get debian-boot-request to handle my subscription messages properly..
Also, I'm not sure that ppp is in base2_1.tgz. Is it? I vaguely remember having to sneakernet the ppp deb last summer when I had cause to install it on a new system.
There are lots of good things about Debian, but please get your facts straight
Daniel
you can't possibly tell me E0.16 will be in potato
:)
:P ) -- of these, xarchon and sawmill are going into potato RSN. squaroid I don't know about, but I didn't find it to be that interesting a game, so.. *shrug*
It already is
Commonness of packages
I think you should be aware that (as opposed to RPMs), most Debian packages are distributed by Debian itself. This means that just because you don't see a package listed on Freshmeat or on the homepage doesn't mean you can assume that no package exists, or that the version reported is correct. And unless you need to have the version released last week, unstable is generally fine -- note I said generally, some *specific* maintainers have been slower about keeping up with releases. This is generally with big and complex pieces of software (eg, XFree) in which multiple patches and fixes have to be backed out or put back in inbetween upstream releases, and which require coordination between multiple developers. PHP3, Mysql, and Apache were significantly behind for a while last spring, I'm not sure what the status is now (I haven't tried using them recently)
There are very few programs which I haven't been able to get as Debian packages; for those which I can't, I usually build a package myself. For most software (read: software which uses autoconf and automake) this takes about 5 minutes+time to compile. The only recent examples were squaroid, xarchon, and sawmill (which I'm too lazy to download
Daniel
I think that although many people have many opinions -- probably someone will say that "it's too ugly" or "dselect sucks!", there is one and only one problem: the entire lack of a package hierarchy. This problem is peculiar to Debian since no other distribution actually lets the user list every program which could potentially be installed on his or her system, but that's no excuse for not solving it :)
The current system provides a few broad groupings of packages (net/, devel/, etc) with no strong policy [that I know of] about where packages go -- libraries can be found in every section (for example, libmagick4g is in graphics/) and no consistency in package organization. This leads to two problems:
1) When setting up a system for the first time, the new user must examine each and every of the 4000 packages and determine which ones he/she wants to install. There's no way to know ahead of time what the next section of packages holds, and so you have to examine it. (in some cases, libs/ and devel/ might be eliminatable)
2) There is no way to get a quick listing of all available software of type X, and no reliable way to do it (besides the needle-in-a-haystack approach). Because of this, installing software on an already-set-up system also requires a huge amount of work.
Both problems are exacerbated by the fact that dselect refuses to merge the different priorities and overarching sections (there's a case to be made for non-free and non-us, but separating Base, Optional, and Extra from one another by default is indefensible), meaning that there are actually *multiple* occurences of each section! So I actually have to search three or four versions of web/ and x11/ to find Web browsers.
More annoyingly, these are not just problems for newbies, but also (IMO) hinderances for experienced users -- good organization of information is just a good idea.
I posted a message some time last summer on debian-devel about this but it was mostly ignored, so now I'm working on actually implementing a working system of tagging packages to put them into a logical hierarchy. (yes, another curses APT frontend) I just started, though, so I don't expect it to be able to do anything particularly exciting until after potato is released.
Daniel
PS - historically, the huge number of questions asked in the process of unpacking packages has been a problem. Either in potato or in potato+1 this will be solved in a surpassingly elegant way by debconf.
I think that concern is a red herring.
Daniel
I have mixed feelings here. On the one hand, it's great that Loki is releasing this, and it looks to be a much better way of installing software than simply having a random perl script to unpack tars (VMware) or something.
/usr/local/stow and rm -rf the whole lot. And how do I track what I've installed once I install more than two or three? How do I update to the latest patchlevel? Download patches and run them? If I find a file and can't remember what program it belongs to, how do I check?
On the other hand...
I have a great aversion to installing random cruft on my system. I have an even greater aversion to letting a random program install random cruft on my system. Is there a separate file to run to uninstall each program? (looks like it, in fact you have to build a custom program for each app that needs to be uninstalled) So to delete stuff installed with this program I have to hunt down the uninstaller. Probably better to just install in
Loki would be doing a much greater service to themselves and the world by having the installer auto-generate packages from the generic package description. InstallShield-type stuff, where each program in the world ships with its own special installer, is an ugly and inelegant hack. The most redeeming quality this installer could have would be if it took command-line options to specify all the attributes of the install (eg, bindir=, datadir=, etc) and skip the interactive stuff -- this would allow halfway-useful packages to be made fairly easily. Ah well..
Daniel
You can learn a lot by teaching yourself. A whole lot. However, there is a lot to learn. A whole lot. A whole lot more, in fact, than you can learn by teaching yourself.
:-) I could go on: foreign languages, history of obscure places, etc, but I feel that these examples are enough.
Consider this: I can (and have) taught myself programming from assembler to C++; I have taught myself a fair deal about the internals and externals of UNIX systems; I have taught myself to use autoconf and automake. These things can be learned by doing and I believe I have learned them well.
Now: how do I teach myself linear algebra? How do I teach myself the fundamentals of quantum mechanics? How do I teach myself about optics? In all these cases, I wouldn't really even know where to begin, and if I did, simply learning one of them would consume so much time that I'd never start on the next. Quantum mechanics you can't even learn by doing, and linear algebra, while useful, is not a structure most people are likely to derive on their own. Fiddling with lenses might get you somewhere in optics, but not everyong is Newton.
This is what teachers, schools, and books are good for: the more theoretical and esoteric bits of learning that a student is not likely to pick up on his or her own. I really hate to use the word 'efficient' in connection with education, but the truth is that you can just learn a lot of things much faster from someone who knows them already than by just flailing around on your own.
I don't mean to disparage the value of hands-on experience with anything from a mathematical equation to a Mozart symphony -- this is indisputably important. It's unfortunate -- no, make that bad -- if you work with computers and can't grope your way to a solution when all else fails. But too many people in this forum and elsewhere have an arrogant "Edukashun? We don't need no steenkin' edukashun!" attitude, thinking that simply because they taught themselves to install Apache that they know everything and there's nothing anyone can teach them. This is generally nonsense; worse is the assumption that there's nothing anyone can teach
anyone
, because in a [nominally] democratic society that attitude can end up hurting
everyone
.
If you want to believe that what you can learn with your hands is the sum total of human knowledge -- fine. But don't limit other people to this.
One last thing -- I may not have learned to use computers in school (I assume you mean elementary and high school, college lets students choose what they want), but I certainly learned writing, mathematics, history and the basics of various sciences there. These, and the mental skills they encourage (eg, logical reasoning), not only help with computers but are also of paramount importance for any would-be self-schooler. Even if you feel that college is beneath you you will need these skills to effectively bootstrap any curriculum you choose.
Daniel