It fell apart, because "each tool" has to work with every other tool in existance.
You make it sound like this is impossible to achieve. All that you need is a well defined API. What do traditional unix command line tools all have in common and use to communicate with eachother with? STDIN, STDOUT and STDERR!
Sure, for GUIs you need a bit more: copy and paste and drag and drop, and maybe something along the lines of Mac OS X's "Services" but I don't think the unix "each tool does one thing well" paradigm is doomed to failure.
how can you predict statically how often you are going to have collisions with other Wi-Fi devices?
Read over my previous posts. I never said anything about "predicting statically." It sounds like what you are saying is that the noise is non-stationary. That's fine, and I'm sure it's true. It has nothing to do with the issue of linear vs non-linear.
Your rocket ship has a pendulum on it? Cool! (-:
In all seriousness, I don't mean to give the wrong impression. I don't have a problem with modeling your system properly and accurately. But I still wonder how much performance their RNN gives over a linear filter. It must be fairly significant or they wouldn't have used it, eh?
I did read it and I'm still surprised. The quote you gave me is meaningless. What sort of random noise levels and from what distribution? I'm willing to accept that they aren't gaussian (white) but you haven't told me why.
Nature may be non-linear but more often than not linear is a damn good approximation. F=-kx anyone?
Frankly, I'm surprised they're even using a non-linear filter. I bet there would be significant performance using simply LMS. I mean, you'd think that WiFi noise would be pretty white, no? Ah, according to the article they are using a Recursive Neural Network which are related to Echo State Machines, so perhaps they have a high dimensional imput space which would lead a linear filter to overfitting problems.
As for not using SVMs, I seem to recall that they are better suited for classification than for parameter estimation.
And this is exactly why I won't use Debian. It's hard to know just what your codebase is. I once had a program that wouldn't work because it wanted a version number of perl beyond a certain patch release, due to a known security bug in the "version" that debian shipped.
I think you are looking at it the wrong way. Debian has made a commitment to supporting their stable branch, specifically to fix known security vulnerabilities. There are two ways to do this:
Update to the next version of the software, which includes the fix and quite possibly many other features, API changes, etc.
Backport the chage to the current version.
Debian developers have decided to go with the latter plan. This is by far the safest route. If they simply upgraded to the "current" version there could be other unforseen consequences which might be worse than the original vulnerability.
When I do a 'perl -v', I expect the stated version to actually be the version listed, not the version + some unknown patches.
I don't understand this. These are not "some unknown patches." If you download the source.deb you will see that it is packaged as the original sources plus a number of patches which are applied at compile time. If you don't want these patches (or you want different ones) it is a simple matter to make your own changes and make your own custom package. Furthermore every package has a changelog.Debian.gz file in/usr/share/doc/$packagename/ which describes pretty well these mysterious patches with cross-references to the debian bug tracker no less!
This is certainly possible, but I'd consider it unlikely barring other evidence. Further, would that even be legal? In any rate, they'd both have your purchase history if they shared information whether you swaped cards or not.
This may not work everywhere, you'll need to do your own experimenting but I find that my Kroger Card works at Food Lion and vice versa (ie it makes a nice beep and I get the discount on my receipt.) You obviously have to do this at the self-checkout lines. I imagine my purchases are being recorded as some other customers records... that or the loyalty card software doesn't do any sanity checking... Hmmmm buffer overflow, anyone?
Try Debian. Just do a base install to get a very minimal but bootable system and then apt-get install foo your way to whatever configuration you need.
As for a low resource using window manager, check out XFCE4. It has the look and feel of Gnome but is far more zippy on old hardware. I run it (and occasionally fluxbox) on a P2 300 laptop with 128 MB RAM.
Right: the bread and wine change to flesh and blood in essence but not in substance.
I've heard that there are obscure rules in cannon law to prevent, say, a priest getting really drunk and wandering in front of a bakery and yelling out "THIS IS MY BODY!!!!" and that being a valid transubstantiative event. (Because then you'd have to send in an army of priests to eat all the God in the bakery, you can't just throw it away or anything...:)
Excellent point. You are right about the computational flexibility of neurons. They can represent a wide range of logical functions, although I believe that the single neuron is incapable of doing an XOR.
Actually, I think it can be done (or at least a partially working XOR.) Imagine a neuron with two inputs and an output. But these inputs are not both excitatory: one is excitatory and the other is inhibitory. So, input only from the excitatory branch produces an action potential, and input from both branches yields no output. Unfortunately, input from just the inhibitory branch would produce no output either.
You can solve this by putting labels in/etc/fstab rather than a direct reference to/dev/whatever (at least for certain fs types.
From the fstab manpage:
Instead of giving the device explicitly, one may indicate the (ext2 or
xfs) filesystem that is to be mounted by its UUID or volume label (cf.
e2label(8) or xfs_admin(8)), writing LABEL= or UUID=,
e.g., `LABEL=Boot' or `UUID=3e6be9de-8139-11d1-9106-a43f08d823a6'.
This will make the system more robust: adding or removing a SCSI disk
changes the disk device name but not the filesystem volume label.
Slashdot Mirror
Sure, for GUIs you need a bit more: copy and paste and drag and drop, and maybe something along the lines of Mac OS X's "Services" but I don't think the unix "each tool does one thing well" paradigm is doomed to failure.
Your rocket ship has a pendulum on it? Cool! (-: In all seriousness, I don't mean to give the wrong impression. I don't have a problem with modeling your system properly and accurately. But I still wonder how much performance their RNN gives over a linear filter. It must be fairly significant or they wouldn't have used it, eh?
which part of approximation don't you understand?
Nature may be non-linear but more often than not linear is a damn good approximation. F=-kx anyone?
As for not using SVMs, I seem to recall that they are better suited for classification than for parameter estimation.
I think you are looking at it the wrong way. Debian has made a commitment to supporting their stable branch, specifically to fix known security vulnerabilities. There are two ways to do this:
- Update to the next version of the software, which includes the fix and quite possibly many other features, API changes, etc.
- Backport the chage to the current version.
Debian developers have decided to go with the latter plan. This is by far the safest route. If they simply upgraded to the "current" version there could be other unforseen consequences which might be worse than the original vulnerability.When I do a 'perl -v', I expect the stated version to actually be the version listed, not the version + some unknown patches.
I don't understand this. These are not "some unknown patches." If you download the source .deb you will see that it is packaged as the original sources plus a number of patches which are applied at compile time. If you don't want these patches (or you want different ones) it is a simple matter to make your own changes and make your own custom package. Furthermore every package has a changelog.Debian.gz file in /usr/share/doc/$packagename/ which describes pretty well these mysterious patches with cross-references to the debian bug tracker no less!
Hot Grits? Good lord man! 2000 called, they want you back!
I believe "Speaker to Animals" was his original name. His noble name was something like "Chimee."
Actually the Soviet Union did use nukes for mining and creating canals... Check out this link (scroll down to the bottom.)
This is certainly possible, but I'd consider it unlikely barring other evidence. Further, would that even be legal? In any rate, they'd both have your purchase history if they shared information whether you swaped cards or not.
This may not work everywhere, you'll need to do your own experimenting but I find that my Kroger Card works at Food Lion and vice versa (ie it makes a nice beep and I get the discount on my receipt.) You obviously have to do this at the self-checkout lines. I imagine my purchases are being recorded as some other customers records... that or the loyalty card software doesn't do any sanity checking... Hmmmm buffer overflow, anyone?
As for a low resource using window manager, check out XFCE4. It has the look and feel of Gnome but is far more zippy on old hardware. I run it (and occasionally fluxbox) on a P2 300 laptop with 128 MB RAM.
As if theft of projectors isn't bad enough already!
<sarcasm> Sure, cause you want to make sure you vote for the winning team!!!!!!!!</sarcasm>
Good point. However, do you remember how long it took you to learn how to walk? ;)
/joeyo
I've heard that there are obscure rules in cannon law to prevent, say, a priest getting really drunk and wandering in front of a bakery and yelling out "THIS IS MY BODY!!!!" and that being a valid transubstantiative event. (Because then you'd have to send in an army of priests to eat all the God in the bakery, you can't just throw it away or anything... :)
Actually, I think it can be done (or at least a partially working XOR.) Imagine a neuron with two inputs and an output. But these inputs are not both excitatory: one is excitatory and the other is inhibitory. So, input only from the excitatory branch produces an action potential, and input from both branches yields no output. Unfortunately, input from just the inhibitory branch would produce no output either.
"Ooooooh, If you want possessive, type Y-O-U-R, but if it's supposed to be a contraction type Y-O-U-APOSTROPHE-R-E.
scalawag!"
:)
From the fstab manpage:
Instead of giving the device explicitly, one may indicate the (ext2 or xfs) filesystem that is to be mounted by its UUID or volume label (cf. e2label(8) or xfs_admin(8)), writing LABEL= or UUID=, e.g., `LABEL=Boot' or `UUID=3e6be9de-8139-11d1-9106-a43f08d823a6'. This will make the system more robust: adding or removing a SCSI disk changes the disk device name but not the filesystem volume label.
24
The only problem I have is that file permissions are not preserved. My solution is to run:
ls -Rl / | /usr/bin/bzip2 > /root/perms.txt.bz2
prior to each backup so that there is at least a record of the permissions. Does anyone know a better way?