Why, oh why, would anyone use Word to publish something like that?
First, because many/most users do not know any other
editor than Word (in fact, for many/most Word is the only piece of software they know - you would be surprised how many users never have used the file manager, or even know it exists).
Second, because most Word users don't know that Word
can export into other formats than.doc
I just installed four systems on my laptop: Debian, Redhat, Gentoo, and FreeBSD.
Debian: installed like a charm, network ok, X configuration ok, no problems at all.
Redhat: (a) installer died when probing the X configuration (on two years old hardware). Any
newbie would have thrown it out of the window at that point...
(b) I did not have too much diskspace for each system (about one Gb), and it was very difficult to trim down the bloat. I wasn't really
successful (everything depends on python and gnome
and hundreds of obscure libraries), did not have too much time to sort it out (few hours), and ended up with some missing functionality (the system is so bloated that with one Gb, you still cannot run the graphical network config tool:(
Gentoo: network (PCMCIA ethernet card) did not come up. On the mailing list, I found a workaround for the problem. Installation went on fine, except
that I needed about half a Gb for compiling X and
that installed daemons by default are not set up to start on boot (somewhat irritating). Could
not compile a working kernel, until I figured out that I had to disable USB, but that is probably not Gentoos fault.
FreeBSD: (I) install kernel did not boot, needed to
disable drivers manually
(Ia) keymap was not configured when manual
intervention was needed first (see above). I have
a german keyboard, i.e. 'y' and 'z' are exchanged.
Result: the system appeared to hang when I typed 'y' - no error message. Very irritating, took some time to figure out.
(Ib) When 'sysinstall' finally came up, I could
choose a keymap, but it did not work. On my laptop, some keys double as numerical keypad, and for all these keys, only the numerical function worked. Very annoying (it worked, however, after the first reboot).
(II) I run out of free inodes when compiling the
kernel. The installation instructions did not
explain that FreeBSD has a problem with inodes..
A few days later, I had to reconfigure my network address, and by accident introduced an error in the routing (wrong subnet for gateway). Results:
Debian: fixed the routing automagically, network worked
Redhat: detected the error, did not add any routes, no network
Gentoo, FreeBSD: did not detect the error, added the route as specified, no network
The kernel can easily know whether you are the apache program or not, however, because it knows the inode backing your executable -- there is no way to forge that.
like that (overwriting will not change the inode) ?:
cp apache apache.tmp
cat rogue_executable > apache
I've always thought it was obvious that super-massive blackholes lie at the center of galaxies. The intense gravity at the center should create one,...
Not necessarily. The point is that massive black holes, wherever in a galaxy they have formed, will
'sink' to the center of the galaxy very fast (at
least compared to the total age of the universe).
The main reason is dynamical friction - as the
BH moves through the stars (and molecular clouds),
it will alter the orbit of these other objects, and create a small overdensity behind itself. This will slow down the black hole, and cause it
to sink towards the center.
(1) The Chandrasekhar limit gives the size limit for a star to collapse and produce a white dwarf.
The Chandrasekhar limit (about 1.4 times the solar mass) gives the upper mass limit
for a white dwarf (i.e. a star where gravity is balanced by electron degeneracy pressure). If a
white dwarf crosses the Chandrasekhar limit (e.g. by accretion from a companion in a binary star system), it will become a neutron star (an object about as dense as an atomic nucleus). For neutron stars, there is also an upper mass limit, which is about 2 solar masses, but not very well known (because the equation of state at such high
densities is not well known). A neutron star will collapse to a black hole if it crosses that upper mass limit.
(2) Finally, recent research shows that the universe is inflating, due to Einstein's cosmological constant (which, he ironically labelled as his "worst mistake"). That is, Hubble's constant is increasing. There will be no Big Crunch. The universe will expand at a faster and faster rate into nothingness.
That is a result that is currently still under
debate. It is not clear whether systematic errors are fully understood, and the majority of cosmologists probably regard this result as
interesting, but unproven (read: it needs
verification by independent groups of researchers,
and possibly independent methods, before it will be widely accepted).
Separate filesystems are bad if you have only little
diskspace (this one out of five OSes on my laptop).
Putting everything in / at least guarantees that no space is wasted. And the installation manual did not point out the inodes problem, neither how to solve it...
Have installed FreeBSD. Downloaded the ports tree.
Tried to install the first port. Error message:
out of free inodes. (Fixed it by deleting major
parts of the ports tree).
FreeBSDs ports system takes a huge amount of disk
space, and takes away countless inodes which apparently
are a scarce resource in FreeBSDs filesystem.
On the other hand, my Debian installation is rock-solid, was the only Linux distro that figured out the X server configuration properly, and software upgrades are as simple as 'apt-get install xyz'...
Some of you may be thinking there's no need to recompile the kernel if you can just use insmod. Have you heard of the module-based rootkits? My hardened system has loadable modules disabled. If I need to compile something, I do it on another system. A little paranoia pays off in this world.
This will not help you at all. One can modify
the kernel at runtime using/dev/kmem, and
you cannot protect against that (for a detailed
discussion, see this article
from Phrack 58). There are rootkits out there that
use this technique.
I have some 50+ peer reviewed papers, and I also have reviewed several papers by others (you are expected to act as referee now and then, if you publish in peer reviewed publications). I can tell you that:
(a) repeating experiments is not the duty of the referee (yes, you get a list of things you should do when reviewing). Actually, repeating an
experiment takes much more time, effort, and money than a referee could ever afford. Repeating
an experiment of this size requires several man-years (and the money to pay for the job).
(b) as a referee you can, and should, watch out for bad style,
bad grammar/spelling, and obscure or needlessly
complex wording. you should also watch out for inconsistencies, and/or conclusions that are not supported by the facts given.
(c) repeating the experiment, and judging on the validity of the results, is the responsibility of the scientific community as a whole. If it is important, someone will step up and repeat it.
(d) the difference between submitted date and published date is due to a combination of lazy
referees, lazy authors (resubmitting the paper months after receiving the referees comments),
and publication backlog.
Talking about accuracy:
his program estimates 11.71 person-years to build
one of the applications I have developed. Actually, I am working three years in my spare time on it... maybe I have unknowingly figured out how to warp time ?
I am signing all my sources with GnuPG. However,
the problem is that it is not enought to verify the signature - if you want to trust signed source, you would also need to verify the key (fingerprint), and according to my experience, almost nobody does that, presumably out of lazyness.
anyone who complains about a lack of configurability apparently never had to deal with:
people who managed to tear off a taskbar by accident, and could not figure out how to put it back in place,
people who managed to switch off a taskbar by accident (this evil M$ Word...), and could not
figure out how to switch it on again,
countless other examples...
Many, perhaps most, users use their PC only occasionally, are not familiar with configuration options, cannot 'fix' even the most trivial issues, and would rather need a well thought out configuration that cannot be modified by any
means.
I code in C. I only use safe functions (like snprintf, and OpenBSD's strlcpy/strlcat functions).
I admit that still some bugs show up now and then, but these are
almost invariably caused by flawed logic in the
code (e.g. the control flow), and I don't see
how changing the language would be of any help with that.
Is there any serious study that proves the superiority of a particular language with respect to (say) bugs per function point ? It seems that
lots of people claim that Java or C++ are better
than C, but nobody ever quotes a reference for this statement...
Popular science is often dumbed down to the max, and it is really difficult to say whether Gonzales' conclusions are really as unsubstantiated as they appear from the article.
Just to point out some of the problems:
bombardement by comets/meteorites: He argues that in dense environments, this would be much more serious. However, it has been more serious in the past (known from crater counts on moon). This has depopulated the inner solar system from comets/meteorites long ago. One could argue that dense environments would favour very fast depopulation out to large radii, followed by a much calmer environment than our own solar system. You really can't say anything without solid numerical simulations, which do not exist so far.
spiral arms: according to current wisdom, the spiral arms represent a wave pattern moving with a different speed than the rotation of stars around the galactic centre. This implies that all stars in the galactik disk (including the Sun) cross the arms at regular intervals.
star clusters most probably the Sun has formed in some kind of cluster as well. And while planet frequency in clusters is unknown, the frequency of binary stars is known to be high,
although it was thought previously that the environment of massive stars could be hostile for binary star formation...
What's wrong with that expectation ? Would you
like to have a car without an engine ?
The point is that users buy computers in order to
perform certain tasks, just like they buy cars
in order to perform certain tasks, and therefore
they expect that the thing is ready to be used.
You may find this amazing, but from a consumers
perspective, it is logical to expect that what
you buy can actually be used for the intended
purpose.
I don't fully agree with your conclusions. In particular, while make is bloated with features,
they don't get in your way - if you want simplicity,
it is straightforward to create a Makefile as simple
as you want.
Just downloaded the patch. After download, a security info gets displayed, and it says that the patch was signed 24.04.02 21:04... not really sure what to think about that, but there is nothing really important on the box anyway.
Don't know PM Mail 2000, but I have tried the
other three and found them rather poor
(Eudora lite: nasty ads, could not find out whether/how to configure for multiple accounts,
Netscape et al: bloated, slow, limited functionality, Pegasus: horrible user interface).
By far
the best one I have found so far is
Sylpheed (yes, it is based on GTK+, but there is a Windows binary available, and it even supports GnuPG, as well as SSL connections for POP3/IMAP).
Still, that's only about 4/day which seems very conservative to me.
True - it's about what I get daily. The problem is: I can pretty well get rid of spam in my private mail using one-time mail aliases for most purposes, but I can't do that at work.
The fatal flaw is that you can't stick all your heat into something small without using a lot of extra energy.
How is that a fatal flaw ? It is (well, at least should be) pretty obvious that you need some extra
energy to power an active system that gets you rid of the IR emission. But I agree that BB emission is suboptimal - radiating away the energy in a narrow wavelength range would be more appropriate.
Your 'high flow rate' for convection would also
require substantial energy. Probably it could work, but only if the throughput is so high that rise in air temperature would stay within the normal range of fluctuations. Just spreading out the heat will
not help much - as you point out yourself, brightness goes as T^4, so even a small temperature difference creates an easily visible
'cloud' around a person.
This is certainly one of the more interesting
contributions to this topic.
One note however: there is an ongoing debate in
the scientific community about the quality of
the peer-review process of journals, and not all
people are happy with it. Peer reviewers are not
alway as careful as they should be, authors have
a tendency to simply submit an article for another
journal, if the first journal they tried has rejected it, and especially medical publications are often sponsored by interested parties that have a lot of money to win or loose. Note that the qouted abstracts represent four papers on apparently only two different studies, and one of these studies is sponsored by the military.
Which means: reading an abstract often doesn't tell you the real story. You need to read the full article to understand what hidden assumptions have been made, how the selection of participants in the study helped to influence the outcome, and how significant the results really are. Unfortunately,
access to the original publication may be rather expensive.
The peak wavelenght of thermal radiation depends on the
temperature of the radiating object (Planck's law
on blackbody radiation). So, while it
is clearly necessary to get rid of the excess heat
generated by the human body, it should be possible
to do this in a different wavelenght range (e.g. build a thermos outfit, collect the body heat with a cooling system, and radiate away in the ultraviolet rather than in the IR).
It should also be possible to radiate away the
heat in a suitably directed narrow beam, rather than uniformly.
On the other hand, using convection
would not help at all (the heated air would radiate in the IR just like the human body does).
Re:Not as bad as all that
on
Google Juice
·
· Score: 1
At least some religious communities clearly do
not use "Google bombing"... e.g., if you search for samhain, you get about 91,600 results, but the
first (neo-)pagan site is only the fourth in the list.
Slashdot Mirror
First, because many/most users do not know any other editor than Word (in fact, for many/most Word is the only piece of software they know - you would be surprised how many users never have used the file manager, or even know it exists).
Second, because most Word users don't know that Word can export into other formats than .doc
Debian: installed like a charm, network ok, X configuration ok, no problems at all.
Redhat: (a) installer died when probing the X configuration (on two years old hardware). Any newbie would have thrown it out of the window at that point ... :(
(b) I did not have too much diskspace for each system (about one Gb), and it was very difficult to trim down the bloat. I wasn't really successful (everything depends on python and gnome and hundreds of obscure libraries), did not have too much time to sort it out (few hours), and ended up with some missing functionality (the system is so bloated that with one Gb, you still cannot run the graphical network config tool
Gentoo: network (PCMCIA ethernet card) did not come up. On the mailing list, I found a workaround for the problem. Installation went on fine, except that I needed about half a Gb for compiling X and that installed daemons by default are not set up to start on boot (somewhat irritating). Could not compile a working kernel, until I figured out that I had to disable USB, but that is probably not Gentoos fault.
FreeBSD: (I) install kernel did not boot, needed to disable drivers manually ..
(Ia) keymap was not configured when manual intervention was needed first (see above). I have a german keyboard, i.e. 'y' and 'z' are exchanged. Result: the system appeared to hang when I typed 'y' - no error message. Very irritating, took some time to figure out.
(Ib) When 'sysinstall' finally came up, I could choose a keymap, but it did not work. On my laptop, some keys double as numerical keypad, and for all these keys, only the numerical function worked. Very annoying (it worked, however, after the first reboot).
(II) I run out of free inodes when compiling the kernel. The installation instructions did not explain that FreeBSD has a problem with inodes
A few days later, I had to reconfigure my network address, and by accident introduced an error in the routing (wrong subnet for gateway). Results:
Debian: fixed the routing automagically, network worked
Redhat: detected the error, did not add any routes, no network
Gentoo, FreeBSD: did not detect the error, added the route as specified, no network
like that (overwriting will not change the inode) ?:
cp apache apache.tmp
cat rogue_executable > apache
Not necessarily. The point is that massive black holes, wherever in a galaxy they have formed, will 'sink' to the center of the galaxy very fast (at least compared to the total age of the universe). The main reason is dynamical friction - as the BH moves through the stars (and molecular clouds), it will alter the orbit of these other objects, and create a small overdensity behind itself. This will slow down the black hole, and cause it to sink towards the center.
(1) The Chandrasekhar limit gives the size limit for a star to collapse and produce a white dwarf.
The Chandrasekhar limit (about 1.4 times the solar mass) gives the upper mass limit for a white dwarf (i.e. a star where gravity is balanced by electron degeneracy pressure). If a white dwarf crosses the Chandrasekhar limit (e.g. by accretion from a companion in a binary star system), it will become a neutron star (an object about as dense as an atomic nucleus). For neutron stars, there is also an upper mass limit, which is about 2 solar masses, but not very well known (because the equation of state at such high densities is not well known). A neutron star will collapse to a black hole if it crosses that upper mass limit.
(2) Finally, recent research shows that the universe is inflating, due to Einstein's cosmological constant (which, he ironically labelled as his "worst mistake"). That is, Hubble's constant is increasing. There will be no Big Crunch. The universe will expand at a faster and faster rate into nothingness.
That is a result that is currently still under debate. It is not clear whether systematic errors are fully understood, and the majority of cosmologists probably regard this result as interesting, but unproven (read: it needs verification by independent groups of researchers, and possibly independent methods, before it will be widely accepted).
Separate filesystems are bad if you have only little diskspace (this one out of five OSes on my laptop). Putting everything in / at least guarantees that no space is wasted. And the installation manual did not point out the inodes problem, neither how to solve it ...
FreeBSDs ports system takes a huge amount of disk space, and takes away countless inodes which apparently are a scarce resource in FreeBSDs filesystem.
On the other hand, my Debian installation is rock-solid, was the only Linux distro that figured out the X server configuration properly, and software upgrades are as simple as 'apt-get install xyz' ...
This will not help you at all. One can modify the kernel at runtime using /dev/kmem, and
you cannot protect against that (for a detailed
discussion, see this article
from Phrack 58). There are rootkits out there that
use this technique.
What particular pharmacutical company is this ? My wife is a doctor, and she would certainly be interested :-)
(a) repeating experiments is not the duty of the referee (yes, you get a list of things you should do when reviewing). Actually, repeating an experiment takes much more time, effort, and money than a referee could ever afford. Repeating an experiment of this size requires several man-years (and the money to pay for the job).
(b) as a referee you can, and should, watch out for bad style, bad grammar/spelling, and obscure or needlessly complex wording. you should also watch out for inconsistencies, and/or conclusions that are not supported by the facts given.
(c) repeating the experiment, and judging on the validity of the results, is the responsibility of the scientific community as a whole. If it is important, someone will step up and repeat it.
(d) the difference between submitted date and published date is due to a combination of lazy referees, lazy authors (resubmitting the paper months after receiving the referees comments), and publication backlog.
Talking about accuracy: his program estimates 11.71 person-years to build one of the applications I have developed. Actually, I am working three years in my spare time on it ... maybe I have unknowingly figured out how to warp time ?
I am signing all my sources with GnuPG. However, the problem is that it is not enought to verify the signature - if you want to trust signed source, you would also need to verify the key (fingerprint), and according to my experience, almost nobody does that, presumably out of lazyness.
- people who managed to tear off a taskbar by accident, and could not figure out how to put it back in place,
- people who managed to switch off a taskbar by accident (this evil M$ Word
...), and could not
figure out how to switch it on again,
- countless other examples
...
Many, perhaps most, users use their PC only occasionally, are not familiar with configuration options, cannot 'fix' even the most trivial issues, and would rather need a well thought out configuration that cannot be modified by any means.I admit that still some bugs show up now and then, but these are almost invariably caused by flawed logic in the code (e.g. the control flow), and I don't see how changing the language would be of any help with that.
Is there any serious study that proves the superiority of a particular language with respect to (say) bugs per function point ? It seems that lots of people claim that Java or C++ are better than C, but nobody ever quotes a reference for this statement ...
Just to point out some of the problems:
The article also mentions that software companies need to get on some M$ program in order to get the API for the control system disclosed to them ...
The point is that users buy computers in order to perform certain tasks, just like they buy cars in order to perform certain tasks, and therefore they expect that the thing is ready to be used.
You may find this amazing, but from a consumers perspective, it is logical to expect that what you buy can actually be used for the intended purpose.
I don't fully agree with your conclusions. In particular, while make is bloated with features, they don't get in your way - if you want simplicity, it is straightforward to create a Makefile as simple as you want.
Just downloaded the patch. After download, a ... not
security info gets displayed, and it says that
the patch was signed 24.04.02 21:04
really sure what to think about that, but there
is nothing really important on the box anyway.
By far the best one I have found so far is Sylpheed (yes, it is based on GTK+, but there is a Windows binary available, and it even supports GnuPG, as well as SSL connections for POP3/IMAP).
True - it's about what I get daily. The problem is: I can pretty well get rid of spam in my private mail using one-time mail aliases for most purposes, but I can't do that at work.
How is that a fatal flaw ? It is (well, at least should be) pretty obvious that you need some extra energy to power an active system that gets you rid of the IR emission. But I agree that BB emission is suboptimal - radiating away the energy in a narrow wavelength range would be more appropriate.
Your 'high flow rate' for convection would also require substantial energy. Probably it could work, but only if the throughput is so high that rise in air temperature would stay within the normal range of fluctuations. Just spreading out the heat will not help much - as you point out yourself, brightness goes as T^4, so even a small temperature difference creates an easily visible 'cloud' around a person.
One note however: there is an ongoing debate in the scientific community about the quality of the peer-review process of journals, and not all people are happy with it. Peer reviewers are not alway as careful as they should be, authors have a tendency to simply submit an article for another journal, if the first journal they tried has rejected it, and especially medical publications are often sponsored by interested parties that have a lot of money to win or loose. Note that the qouted abstracts represent four papers on apparently only two different studies, and one of these studies is sponsored by the military.
Which means: reading an abstract often doesn't tell you the real story. You need to read the full article to understand what hidden assumptions have been made, how the selection of participants in the study helped to influence the outcome, and how significant the results really are. Unfortunately, access to the original publication may be rather expensive.
On the other hand, using convection would not help at all (the heated air would radiate in the IR just like the human body does).
At least some religious communities clearly do not use "Google bombing" ... e.g., if you search for samhain, you get about 91,600 results, but the
first (neo-)pagan site is only the fourth in the list.