I especially agree with the third point - guessable URLs. Unlike some sites such as Microsoft where their 'easy' urls are along the lines of http://go.microsoft.com/j483dk4, and even worse sites like Nokia where the URLs take the guise of http://www.nokia.co.uk/nokia/0,8764,46609,00.html.
Totally agree. copy/pastable urls (ones 75 chars) are nice, so are readable ones (reflect site structure), so are 'editable' ones (i can go one level up by removing the last entry on the url instead of the application pissing itself because the url isn't supposed to exist or something). It never seems to happen.
(The most fascinating way to explore the Mandelbrot set is to take a single point in the set and plot how the values change. Some escape to infinity, others seem to be "pulled" towards one or more regions which they then orbit.
I was a bit relieved to find out that, according to a book (Ross Anderson's Security Engineering I think) I've read, the implied method is slapping a rubber hose against the subject's feet.. I was worried about more awful forms of rubber hose application to the human body.
I suppose this is plenty awful soon enough though.
I couldn't agree more. dd has plenty more features cp doesn't have. I was just saying it's silly to use dd as cp, just because it involves a device node (as seems to be the habit).
Actually, I use zsh, and it doesn't tabcomplete for me. Doubtlessly that can be enabled in my case as well, but that wasn't my main point. My main point was that I think it's silly to use dd as cp just because it's about a device.
Ok, you're right about being able to set block sizes, as cp might be more conservative with that than you'd like. However, I did qualify my 'cp=dd' statement by saying that they're only the same if you don't specify skips and truncations and whatnot, as cp can't do that. block is are part of that, even though they won't affect the result (it's just that the copying may be more efficient with a larger blocksize).
I wonder why people like to use 'dd' when it comes to devices. Without skipping, conversion, truncating, etc., (i.e., just naming an if= and an of=) dd is the same as cp, but trickier. (e.g., you can't use tab completion in the shell easily due to the if=.. syntax).
In short,
cp/dev/fb0 kde.fbdump
is exactly the same as
dd if=/dev/fb0 of=./kde.fbdump
but simpler and easier to use. No good reason to use dd..
No, you're wrong. All those processes you see running on other systems won't be doing a single thing unless something is asked of them. They'll even largely be swapped out so the memory can be used for cache if they don't run. (This is presuming Linuxes paging algorithm is worth its salt.)
Same story for things in the kernel you don't need, it won't be cached so will take up memory at the most (as kernel memory isn't paged). It makes zero difference for performance what is a module and what isn't.
I suppose recompiling your kernel with your CPU specified and higher optimization than a stock kernel will show performance improvement.. a few percent i would guess. But 50%? You are pulling that out of your ass.
"If you copy (or just write) really big files, then there should be no reason for the raid-5 driver to read all the other sectors in the parity chunk (don't know official terminology for it offhand) since, if done right, you will be writing all of the sectors with brand new information anyway."
Usually true, but I think the GP is talking about updating the parity block when just one block is updated. Still, all other reads aren't necessary as the parity block is updated (XORred) with the XOR of the old and new data block.
Your case is even better of course, if the block device logic is smart enough to group transactions like these together (safely), which involves (safely) waiting to see if any 'close' transactions will be happening..
" Um, no. Since we are writing, we already know what the data is. Just write. No reads."
Presumably, we are talking about the parity disk here, otherwise there aren't any reads involved at all. In that case, you need to read the old parity in order to compute the new one.
My xbox recently stopped wanting to play DVD's; instead of starting the DVD, it would give me the 'service required' screen with an error code indicating an 'application error' (that's a crash right?). Anyone know how I can fix this? "It may be modded", not sure if that's relevant.
I think you and grandparent are right on the money. Not sure I agree with the middle-class-disease explanation, even though what the explanation is doesn't matter right now.
I think it's because of the large amount of exposure professionals get to laymen in that profession; be it programming, lawyering, managaring, or what-have-you. They've all spent many years becoming what they are (university etc) and get much exposure to (a) laypeople in that profession (their customers), and (b) the clique of other professionals, with which they can chitchat about the clueless ones out there.
Technical people thinking they're so much smarter than the rest, e.g. commercial people (managers, marketing, sales) who are needed every bit as much as the technical people, just because they understand a technical thing others don't, really annoy me. It's your job to understand these things, and it's others' job to understand their things!
I know of at least 2 exceptional professional poker players with extensive computer background: Chris "Jesus" Ferguson has a PhD in computer science, and you'll often hear him talking about how his studies in game theory have helped him at the poker table (and I'm thinking he's right, since he won the WSOP main event in 2000).
Hm.. A quote in ony of my game theory books says, loosely, "Watching a bunch of poker players play poker optimally is about as interesting as watching paint dry", and I wholeheartedly agree.
Then again, I didn't win the WSOP main event in 2000.
I agree strongly with 3770 saying we aren't smarter. I think this story reeks of self-stroking. Analytical thinking, indeed..
Not to run RC5, but to crack RC5. From the website I made "RC5-72" a link to in #10549995:
Project: RC5-72
The project to crack a message encrypted with the 72-bit RC5 cipher.
I hope this clarifies my point.
Not very much. Because
In this (unsophisticated) case, cracking == running (as they're brute forcing the key)
Linked from the very page you linked to, the FAQ entry:
Why doesn't an FPU make my computer crack RC5-72 faster?
RC5 involves a large number of integer additions, rotates and XORs. It doesn't require floating point calculations and won't, in general, benefit from them. There has been quite a lot of recent discussion on whether or not it might be possible to boost keyrates (on x86 architectures at least) by taking advantage of the fact that there are separate pipelines for integer and floating point instructions. (We leave it to the reader to figure out how to do floating-point XORs and rotates!)
Which was my original remark to back up someone you followed up to with this:
Amazing indeed... Since when do you need floating point operations for text matching?!
Not for text matching, but for numbercrunching.
Numbercrunching as in RC5-72 [distributed.net]. I think you get the idea...
People who watch The Daily Show did better on a quiz about their political knowledge than people who watch any of the cable news shows - FOX, CNN, MSNBC, etc.
Doesn't that mean he's doing his duty to inform people?
Slashdot Mirror
Totally agree. copy/pastable urls (ones 75 chars) are nice, so are readable ones (reflect site structure), so are 'editable' ones (i can go one level up by removing the last entry on the url instead of the application pissing itself because the url isn't supposed to exist or something). It never seems to happen.
They're julias aren't they?
Nitpick: Factoring primes eh?
The technical term for one such method is rubber-hose cryptanalysis..
I was a bit relieved to find out that, according to a book (Ross Anderson's Security Engineering I think) I've read, the implied method is slapping a rubber hose against the subject's feet.. I was worried about more awful forms of rubber hose application to the human body.
I suppose this is plenty awful soon enough though.
sounds redundant to me.
Troll or uninformed bullshit.
I couldn't agree more. dd has plenty more features cp doesn't have. I was just saying it's silly to use dd as cp, just because it involves a device node (as seems to be the habit).
Actually, I use zsh, and it doesn't tabcomplete for me. Doubtlessly that can be enabled in my case as well, but that wasn't my main point. My main point was that I think it's silly to use dd as cp just because it's about a device.
Ok, you're right about being able to set block sizes, as cp might be more conservative with that than you'd like. However, I did qualify my 'cp=dd' statement by saying that they're only the same if you don't specify skips and truncations and whatnot, as cp can't do that. block is are part of that, even though they won't affect the result (it's just that the copying may be more efficient with a larger blocksize).
In short,
is exactly the same asbut simpler and easier to use. No good reason to use dd..I've been told I am
Doesn't mean he was right.
No, you're wrong. All those processes you see running on other systems won't be doing a single thing unless something is asked of them. They'll even largely be swapped out so the memory can be used for cache if they don't run. (This is presuming Linuxes paging algorithm is worth its salt.)
Same story for things in the kernel you don't need, it won't be cached so will take up memory at the most (as kernel memory isn't paged). It makes zero difference for performance what is a module and what isn't.
I suppose recompiling your kernel with your CPU specified and higher optimization than a stock kernel will show performance improvement.. a few percent i would guess. But 50%? You are pulling that out of your ass.
This post is full of nonsense (see replies below, some of mine)
Usually true, but I think the GP is talking about updating the parity block when just one block is updated. Still, all other reads aren't necessary as the parity block is updated (XORred) with the XOR of the old and new data block.
Your case is even better of course, if the block device logic is smart enough to group transactions like these together (safely), which involves (safely) waiting to see if any 'close' transactions will be happening..
Presumably, we are talking about the parity disk here, otherwise there aren't any reads involved at all. In that case, you need to read the old parity in order to compute the new one.
(Otherwise I'm really happy with it btw.)
You may want to read about this: small world phenomenon. It's not necessarily the right explanation of course.
So when it suits us, bill gates is a real geek.. right.
I think it's because of the large amount of exposure professionals get to laymen in that profession; be it programming, lawyering, managaring, or what-have-you. They've all spent many years becoming what they are (university etc) and get much exposure to (a) laypeople in that profession (their customers), and (b) the clique of other professionals, with which they can chitchat about the clueless ones out there.
Technical people thinking they're so much smarter than the rest, e.g. commercial people (managers, marketing, sales) who are needed every bit as much as the technical people, just because they understand a technical thing others don't, really annoy me. It's your job to understand these things, and it's others' job to understand their things!
Thank you for listening. :)
I agree strongly with 3770 saying we aren't smarter. I think this story reeks of self-stroking. Analytical thinking, indeed..
hhgttg of course! zz 9 pluarl z alpha, does that mean anything to you zaphod, what does the z stand for, which z, any z, etc.
Project: RC5-72 The project to crack a message encrypted with the 72-bit RC5 cipher.
I hope this clarifies my point.
Not very much. Because
- In this (unsophisticated) case, cracking == running (as they're brute forcing the key)
- Linked from the very page you linked to, the FAQ entry:
Which was my original remark to back up someone you followed up to with this:
I hope it's clear now..Not for text matching, but for numbercrunching. Numbercrunching as in RC5-72. I think you get the idea...
Excuse me? floating point operations to run RC5? I don't think you get the idea..
forgetting strcmp() returns 0 when strings match are we?
Doesn't that mean he's doing his duty to inform people?
Course it doesn't necessarily mean that.