It's certainly impressive and all, but -- if you aren't interested in recompiling it, playing with it, keeping track of its progress, possibly helping out with development/debugging -- why are you using -current at all? Was there some crucial feature which you needed and couldn't find in -stable?
In general, most installations don't bang/var and/tmp hard enough to necessitate putting them on seperate drives,
You'd be surprised how often/tmp gets filled up
if you're not careful. Yes, it doesn't particularly matter if you're a laptop user but it does matter if it's a server or even a multi-user workstation. You don't want people to be locked out of their accounts because the lack of a separate/tmp filled up the entire root filesystem.
BTW, it's also good practice to make a small / partition and a larger partition for/usr etc (as the BSD's do, and as the Linuxes usually do not). The reason is that, in case of filesystem damage, you have a better chance of at least being able to boot to single-user mode and fix things from there.
As for/var, of course, you should have an idea of its required size before doing the installation; if it's not a server there's probably no harm in symlinking it to/usr/var or something.
Also, I don't quite understand how lack of atime can effect whether you have new mail. If mail is stored in one big file (which is an evil practice that should be relegated to the dust-bins of history) then mtime is what should be read. If mail is stored as seperate files, then its the mtime of the directory which should be read.
Yes, but once you have read the mail, you don't want the mailbox to be showing "new mail" any more. That's where the atime comes in. If the atime is newer than the mtime, you don't have new mail. Ditto for a maildir type system.
It'll find the config files that need updating, show you a diff between them and choose to keep your old one, use the new one, more merge them together.
That's exactly what I was looking for (and what mergemaster does). I can't find it in my Gentoo in any of/sbin,/usr/sbin,/bin,/usr/bin (I last updated the system around 2 weeks ago). Where is it, which port does it come with?
Gentoo's portage is nice...
on
Gentoo 1.0 Released
·
· Score: 5, Interesting
I mainly use FreeBSD and when I wanted to install linux on my second partition, I picked gentoo because of its "BSD-like" ports system. Personally, it has some nice features and improvements over the FreeBSD system, the flip side is it may take more disk space. For example, it seems to automatically figure out the contents of the package, by first installing in a dummy area and then copying it all to the real destination. (That's basically what you're recommended to do when creating a port in FreeBSD, but it's not automatic.)
The number of ports available doesn't compete with FreeBSD (or, I imagine, the other BSDs) at this point, but that could change quickly.
On the minus side, some of the author's tuning instructions are dubious, or downright silly. He recommends using "noatime" in the filesystem everywhere -- now that may be ok for/home, or
for/usr, say, but for/var, which holds mailboxes, it's not a good idea -- the atime is used to tell whether you have new mail. (In fact, the default partitioning doesn't even create a separate/var or/tmp, and the install guide doesn't suggest you do it. This is not unique to Gentoo, it's a common attitude in the linux world, but it looks like a bad idea to me.)
Also, the global CFLAGS sets -O3, which looks overambitious to me -- the only change -O3 has over -O2 is function inlining, which sometimes helps and sometimes hurts, you definitely don't want to build your system with it. (The FreeBSD project doesn't support anything beyond -O, though I generally have no problem with -O2.)
Anyway, it's easy to fix these small caveats. (Another good thing is Gentoo doesn't clobber your config files when you upgrade; however, something like FreeBSD's "mergemaster" for upgrading/etc would certainly be welcome.) That apart, Gentoo looks like a nice system and I'm happy with it.
This is definately goo, and $55 is a lot cheaper than the $200 to $800 for Microsoft Office. I wonder how they are doing this?
Because they're not giving you Microsoft Office: you still have to buy that from Microsoft, at Microsoft's prices. What Codeweavers are giving you is something which lets you install and run MS Office on Linux using the MS Office CD.
You can patch Mozilla 0.9.9 to get WMP plugins to work. This is (sort of) mentioned in the
article. The patch is
here
and the bug report (resolved) is
here.
And it works fine for me, moreover...
Well, Covalent can do that because the Apache license allows it. (It's basically BSD-like, with some modifications, eg you can't call the modified code Apache.) The GPL doesn't allow you to link to GPL'd code without making the whole thing GPL, which makes it rather to hard to use GPL libraries to write a commercial word processor, for example.
The other replies have made the points I planned to make, but there's a larger issue.
If you decide to nuke the nut's city to get the nut, how does that make you different from the nut?
Since George W Bush has repeatedly shown his contempt for the rest of the world, international law, the environment, the future of the planet, why aren't other governments justified in nuking Washington to get that nut who's threatening the rest of the world with nukes?
Simply because America happens to be the self-proclaimed "leader of the free world"?
Real leadership can only come if you build respect. The US has dissipated its goodwill in Europe astonishingly quickly -- all the sympathy after Sept 11 took just a few months to evaporate. If the US is to be different from the USSR and other "evil empires", it has to learn to be responsible.
The US has always
refused to make a "no first use" pledge about nuclear weapons. The Clinton admin was "shocked" by Germany's proposal that NATO make such a pledge.
Soon after Sept 11, senior people in the military were quoted as saying that they wanted the entire Afghanistan/Middle East region to
"glow with radiation."
So, no, I'm not surprised that the US wants to use nukes. Particularly against that axis of evil -- if you can't nuke them, who can you nuke? And if you can't nuke anyone, what are those nukes for?
More like leaving your pharmaceutical factory unlocked and unprotected, for any criminal in the world to use, on the ground that your friends leave the medicines too. That is, it doesn't just hurt you: it hurts others.
I use html code in my email address on my web page, like this:
rsidd@yaho 1.com
Amazingly, not a single spammer has gotten hold of it yet, in over a year; whereas, unobfuscated
addresses used only once, on mailing list archives for example, are picked up immediately.
Its sloppy, yes, if you keep up with the bleeding edge. But, the
bleeding edge is always sloppy. Linux, because it has the number of
developers and a well known mailing list exposes that slop to the
public.
I wonder, how sloppy would freeBSD development be if you synced your
kernel hourly with the dev kernels? How buggy would it be? What does
the interdeveloper mail look like in the freebad world?
I run kernels over a year old. There's no slop. I'm (mostly) happy.
In about 6 months, I'll switch to 2.4.
The point is this: linux 2.4 was supposed to be the "stable" series.
2.3 and 2.5 were supposed to be the "sloppy" series, if you like.
FreeBSD's stable series is currently 4.x. Trust me, you can sync
your source tree to the -stable sources at basically *any* point in
time (I do that routinely) and it will run happily, though issues do
(rarely) occur and it's a good idea to scan the freebsd-stable mailing
list for "HEADS UP" messages before syncing. Nevertheless, -stable
gets features merged from the development branch, -current, regularly;
typically, this means new device drivers get ported, but there have
also been significant changes recently to the pccard code and the
linux emulation layer, among others. Basically, freebsd-stable works
and the system works.
And if you prefer the bleeding edge, there's always -current, which is
what many of the developers run. There you can see the sort of major
changes to the VM and other subsystems which have plagued linux 2.4 --
but nevertheless many people claim -current is quite stable and usable
most of the time. Again, you can (and people do) sync their sources
to this tree any time you want.
I think this system works -- in fact, on the evidence, it works
incredibly well. When FreeBSD 4.0 was released (the first "stable"
release in that branch) it was arguably much better than linux 2.4.x
is even today. Linus's system used to work, but is obviously showing
its problems now, and it's time for a rethink, it seems to me.
To answer your question about interdeveloper mail, read
http://docs.freebsd.org/mail/current/freebsd-curre nt.html
http://docs.freebsd.org/mail/current/freebsd-sta bl e.html
http://docs.freebsd.org/mail/current/freebsd-hac ke rs.html
Disclaimer: I'm not a FreeBSD developer, just a user. I use linux
too, but much less now than formerly.
I couldn't find a date on the article. But
it says only a "dramatic improvement" in the
"long-awaited" GCC 3 will change things...
In fact GCC3 was out some time back and it seems typically (to me) to perform around 20% more slowly than Intel's thing (compared to 40% for older gcc, as the article says). It's not so bad, imo.
For Scott Adams' earlier forays into this stuff,
on
God's Debris
·
· Score: 2, Insightful
see this review of his book The Dilbert Future, and his response. I lost much of my respect for him as a serious commentator after reading the last chapter of The Dilbert Future, and the rest of it after reading the above response.
He should stick to cartoons about management, he's certainly good at that.
I often encode with oggenc at 64 kb/s (avg under 60 kb/s). It sounds very close to cd quality to me, and a compression of 1:25 or 1:30. For better quality, I choose 128 kb/s (avg 110 kb/s, compression around 1:13) and I honestly can't tell the difference between that and the original CD. This is on a medium-end music system (Bose "bookshelf" speakers, Kenwood amplifier). MP3s sound strangely hollow and tinny even at 192 kb/s; admittedly I haven't tried encoding any myself in a couple of years, but whatever I've downloaded is pretty much unlistenable on my music system, unless you like your music to sound like it's synthesised by a computer.
And of course, a comment on a humour site, in a story about another humour site, is modded down as "offtopic", perhaps because someone's sensibilities were hurt.
This is a test of the theory that any post which says "Slashdot sucks" will get modded up.
It says it's ignoring them, but the top few "hits" typically do include the exact page. I just tried, for instance, "All your base are belong to us". It claims to ignore "are" and "to" but the top few hits contain the exact phrase. (The same happens with your example "Hail to the chief", though it says it's ignoring "to the".)
The mass media is for, well, the masses --
pretty much by definition. Specialised or
non-mainstream opinions will not get popular no matter how high-quality the publication. Something like salon is by its very nature non-mainstream.
It is possible for web-based startups to succeed over traditional mass media. In India, for example, the extremely popular rediff is not controlled by the traditional newspaper houses (it came from an advertising agency, I think) and more recently, an upstart called tehelka has been making big news, mainly on corruption exposes.
What's needed is intelligent marketing, good content for a broad-spectrum audience with the occasional "scoop" to retain audience interest, and high-quality presentation. The same things as for traditional print media or TV. Nothing new here.
Well ok, since black ink is really black particles in a suspension.
A very small electric field will do very little to separate them (they attract each other too strongly). A very large field will separate them totally (their mutual attraction isn't good enough to overcome it). An intermediately strong field will partially separate them. Clear?
OK let's be clearer here: you can have addition with pigments, and subtraction with filters too.
When you shine light through a red filter, it absorbs everything except red.
When you shine light through a green filter, it absorbs everything except green.
If you put the filters on top of each other, it absorbs everything (or would if the filters were perfect). This is subtractive.
Similarly, if you mix red paint and green paint, the result will absorb both red and green. But if you have separated dots of both, it will emit both red and green. It's additive. There's no way a red dot can subtract light from a spatially separated green dot.
Try thinking about the physics, rather than the terminology. Whether the source of light is in front or behind is irrelevant, what matters is only what's reaching your eyes. (Incidentally, this whole RGB stuff is an approximation made possible by biology, and not perfect: eg, the blue green line of mercury light can't be reproduced by RGB.)
Slashdot Mirror
It's certainly impressive and all, but -- if you aren't interested in recompiling it, playing with it, keeping track of its progress, possibly helping out with development/debugging -- why are you using -current at all? Was there some crucial feature which you needed and couldn't find in -stable?
You'd be surprised how often /tmp gets filled up
if you're not careful. Yes, it doesn't particularly matter if you're a laptop user but it does matter if it's a server or even a multi-user workstation. You don't want people to be locked out of their accounts because the lack of a separate /tmp filled up the entire root filesystem.
BTW, it's also good practice to make a small / partition and a larger partition for /usr etc (as the BSD's do, and as the Linuxes usually do not). The reason is that, in case of filesystem damage, you have a better chance of at least being able to boot to single-user mode and fix things from there.
As for /var, of course, you should have an idea of its required size before doing the installation; if it's not a server there's probably no harm in symlinking it to /usr/var or something.
Also, I don't quite understand how lack of atime can effect whether you have new mail. If mail is stored in one big file (which is an evil practice that should be relegated to the dust-bins of history) then mtime is what should be read. If mail is stored as seperate files, then its the mtime of the directory which should be read.
Yes, but once you have read the mail, you don't want the mailbox to be showing "new mail" any more. That's where the atime comes in. If the atime is newer than the mtime, you don't have new mail. Ditto for a maildir type system.
Sorry, found the gentoolkit thing in app-admin -- I don't have it installed. I'll give it a try soon.
That's exactly what I was looking for (and what mergemaster does). I can't find it in my Gentoo in any of
The number of ports available doesn't compete with FreeBSD (or, I imagine, the other BSDs) at this point, but that could change quickly.
On the minus side, some of the author's tuning instructions are dubious, or downright silly. He recommends using "noatime" in the filesystem everywhere -- now that may be ok for /home, or
for /usr, say, but for /var, which holds mailboxes, it's not a good idea -- the atime is used to tell whether you have new mail. (In fact, the default partitioning doesn't even create a separate /var or /tmp, and the install guide doesn't suggest you do it. This is not unique to Gentoo, it's a common attitude in the linux world, but it looks like a bad idea to me.)
Also, the global CFLAGS sets -O3, which looks overambitious to me -- the only change -O3 has over -O2 is function inlining, which sometimes helps and sometimes hurts, you definitely don't want to build your system with it. (The FreeBSD project doesn't support anything beyond -O, though I generally have no problem with -O2.)
Anyway, it's easy to fix these small caveats. (Another good thing is Gentoo doesn't clobber your config files when you upgrade; however, something like FreeBSD's "mergemaster" for upgrading /etc would certainly be welcome.) That apart, Gentoo looks like a nice system and I'm happy with it.
Because they're not giving you Microsoft Office: you still have to buy that from Microsoft, at Microsoft's prices. What Codeweavers are giving you is something which lets you install and run MS Office on Linux using the MS Office CD.
Still cheaper than a Windows licence though...
Deja vu?
You can patch Mozilla 0.9.9 to get WMP plugins to work. This is (sort of) mentioned in the article. The patch is here and the bug report (resolved) is here. And it works fine for me, moreover...
Well, Covalent can do that because the Apache license allows it. (It's basically BSD-like, with some modifications, eg you can't call the modified code Apache.) The GPL doesn't allow you to link to GPL'd code without making the whole thing GPL, which makes it rather to hard to use GPL libraries to write a commercial word processor, for example.
If you decide to nuke the nut's city to get the nut, how does that make you different from the nut?
Since George W Bush has repeatedly shown his contempt for the rest of the world, international law, the environment, the future of the planet, why aren't other governments justified in nuking Washington to get that nut who's threatening the rest of the world with nukes?
Simply because America happens to be the self-proclaimed "leader of the free world"?
Real leadership can only come if you build respect. The US has dissipated its goodwill in Europe astonishingly quickly -- all the sympathy after Sept 11 took just a few months to evaporate. If the US is to be different from the USSR and other "evil empires", it has to learn to be responsible.
Soon after Sept 11, senior people in the military were quoted as saying that they wanted the entire Afghanistan/Middle East region to "glow with radiation."
So, no, I'm not surprised that the US wants to use nukes. Particularly against that axis of evil -- if you can't nuke them, who can you nuke? And if you can't nuke anyone, what are those nukes for?
More like leaving your pharmaceutical factory unlocked and unprotected, for any criminal in the world to use, on the ground that your friends leave the medicines too. That is, it doesn't just hurt you: it hurts others.
I use html code in my email address on my web page, like this:
rsidd@yaho 1.com
Amazingly, not a single spammer has gotten hold of it yet, in over a year; whereas, unobfuscated
addresses used only once, on mailing list archives for example, are picked up immediately.
Obviously these spambots aren't so intelligent.
I wonder, how sloppy would freeBSD development be if you synced your kernel hourly with the dev kernels? How buggy would it be? What does the interdeveloper mail look like in the freebad world?
I run kernels over a year old. There's no slop. I'm (mostly) happy. In about 6 months, I'll switch to 2.4.
The point is this: linux 2.4 was supposed to be the "stable" series. 2.3 and 2.5 were supposed to be the "sloppy" series, if you like.
FreeBSD's stable series is currently 4.x. Trust me, you can sync your source tree to the -stable sources at basically *any* point in time (I do that routinely) and it will run happily, though issues do (rarely) occur and it's a good idea to scan the freebsd-stable mailing list for "HEADS UP" messages before syncing. Nevertheless, -stable gets features merged from the development branch, -current, regularly; typically, this means new device drivers get ported, but there have also been significant changes recently to the pccard code and the linux emulation layer, among others. Basically, freebsd-stable works and the system works.
And if you prefer the bleeding edge, there's always -current, which is what many of the developers run. There you can see the sort of major changes to the VM and other subsystems which have plagued linux 2.4 -- but nevertheless many people claim -current is quite stable and usable most of the time. Again, you can (and people do) sync their sources to this tree any time you want.
I think this system works -- in fact, on the evidence, it works incredibly well. When FreeBSD 4.0 was released (the first "stable" release in that branch) it was arguably much better than linux 2.4.x is even today. Linus's system used to work, but is obviously showing its problems now, and it's time for a rethink, it seems to me.
To answer your question about interdeveloper mail, reade nt.html
a bl e.html
c ke rs.html
http://docs.freebsd.org/mail/current/freebsd-curr
http://docs.freebsd.org/mail/current/freebsd-st
http://docs.freebsd.org/mail/current/freebsd-ha
Disclaimer: I'm not a FreeBSD developer, just a user. I use linux too, but much less now than formerly.
In fact GCC3 was out some time back and it seems typically (to me) to perform around 20% more slowly than Intel's thing (compared to 40% for older gcc, as the article says). It's not so bad, imo.
He should stick to cartoons about management, he's certainly good at that.
I often encode with oggenc at 64 kb/s (avg under 60 kb/s). It sounds very close to cd quality to me, and a compression of 1:25 or 1:30. For better quality, I choose 128 kb/s (avg 110 kb/s, compression around 1:13) and I honestly can't tell the difference between that and the original CD. This is on a medium-end music system (Bose "bookshelf" speakers, Kenwood amplifier). MP3s sound strangely hollow and tinny even at 192 kb/s; admittedly I haven't tried encoding any myself in a couple of years, but whatever I've downloaded is pretty much unlistenable on my music system, unless you like your music to sound like it's synthesised by a computer.
This is a test of the theory that any post which says "Slashdot sucks" will get modded up.
They used to be an amusing diversion to me. Now I respect them more than most of the mainstream press.
Good idea.<P>
But, well, the fusion guy's problem was solved.
What about the AI character's problem?
Windows is incompatible with what I use everywhere else; I can't get work done on it, unless you count web browsing.
MacOS 9 isn't good enough either. I'm not sure about OS X, but on an iBook I'd probably end up installing linux anyway.
It says it's ignoring them, but the top few "hits" typically do include the exact page. I just tried, for instance, "All your base are belong to us". It claims to ignore "are" and "to" but the top few hits contain the exact phrase. (The same happens with your example "Hail to the chief", though it says it's ignoring "to the".)
It is possible for web-based startups to succeed over traditional mass media. In India, for example, the extremely popular rediff is not controlled by the traditional newspaper houses (it came from an advertising agency, I think) and more recently, an upstart called tehelka has been making big news, mainly on corruption exposes.
What's needed is intelligent marketing, good content for a broad-spectrum audience with the occasional "scoop" to retain audience interest, and high-quality presentation. The same things as for traditional print media or TV. Nothing new here.
A very small electric field will do very little to separate them (they attract each other too strongly). A very large field will separate them totally (their mutual attraction isn't good enough to overcome it). An intermediately strong field will partially separate them. Clear?
When you shine light through a red filter, it absorbs everything except red.
When you shine light through a green filter, it absorbs everything except green.
If you put the filters on top of each other, it absorbs everything (or would if the filters were perfect). This is subtractive.
Similarly, if you mix red paint and green paint, the result will absorb both red and green. But if you have separated dots of both, it will emit both red and green. It's additive. There's no way a red dot can subtract light from a spatially separated green dot.
Try thinking about the physics, rather than the terminology. Whether the source of light is in front or behind is irrelevant, what matters is only what's reaching your eyes. (Incidentally, this whole RGB stuff is an approximation made possible by biology, and not perfect: eg, the blue green line of mercury light can't be reproduced by RGB.)