There are UDP based congestion control protocols out there, so UDP doesn't necessarily have to be "send as fast as you like"; it's just up to the application developer to think a bit more.
That it is another method of distribution, a much more advanced one.
We knew that long ago. Look at the Usenet for example, running around 25 years ago. Peer-to-peer systems have been used for distribution of information for quite some time, it's just that it's been thrown back into the limelight with the illegal sharing of copyrighted content.
Peer-to-peer technology has been around and has been in common usage for many, many years. I can't name any well known peer-to-peer systems before 1979, but I'm quite sure there were some.
P2P had legitimacy long before illegal file sharing came along.
... no computers are allowed to connect to the network unless authorised by the support guys. They're very strict about letting student's personal computers or laptops connect to the network, so much so I can only remember once where they allowed the student to use their own laptop with the department's network (and they're postgrad, so they don't exactly blend into the sea of undergrads).
Presumably they're only allowing certain MAC addresses to connect, so I don't know if they can detect somebody who's reprogrammed their NIC's MAC address. That said, if they were noticed (and the labs we use are all monitored with CCTV), their place on the course would be seriously questioned, as it would probably be a substantial breach of the conditions of use all students in the department sign at the start of the academic year.
The labs are fully featured enough for most tasks anyway. There are a couple of hundred computers for undergrads, and loads of other computers littered throughout the department, and a few specialist labs. The network connects all manner of systems, from all many of Windows installs, to Mac OS X, to loads of Linux systems, and a few BSD machines running in the background, and probably others I've never encountered (it is a CS department, after all!). The department sees no need to put the network in jeopardy by letting any computer connect, rather than just the department's own:)
A lot of "bloat" comes from increasing levels of abstraction away from the hardware, and numerous levels of interfaces between different system layers to allow for more modular code, more reusable code, and generally, safer code. Coding an application for a modern machine with all its fancy hardware would be pretty tricky without these interfaces already written for people. Sure, somebody has got to write the low-level interaction gubbins in the first place, but once it's done, it's done. If software houses wrote for the hardware to reduce bloat, we'd never see any games or software. Praise the bloat. The bloat be praised.
[That's not to say that an understanding of what the machine is actually doing with the code wouldn't help people when designing software systems, I agree that it would certainly be beneficial. It's just not necessarily where all the bloat comes from. The lack of understanding is probably more and more relevant as more and more layers of abstraction are added. And there's always room for one more layer of abstraction!]
...you can see the same patterns over and over again, that were in TNG, DS9 and the original.
It's true. But of all the series, I've grown to like DS9 more than the others. For one, at the start of the series they were in this clapped out station, and certainly weren't in the same position of power the previous series had seen the central characters in. The later series had story arcs - imagine that! A story that was really carried from episode to episode. No other series has had that (well, Voyager had a goal that encapsulated the entire series, and TNG toyed with the Maquis (sp?) at one point and finished off where they started with Q, but none have really done what DS9 did).
That said, I pretty much stopped watching toward the end of DS9/mid Voyager. Not really sure why, it was probably that I was busy and just didn't want to sit in and watch TV in my free moments. I wouldn't mind going back and watching the DS9 episodes at all; I watched TNG when it was repeated ad finitum on BBC2, Voyager and Enterprise I'd only watch if I happened to catch it and I didn't have anything else to do. Same could probably be said about some ToS episodes.
/me wonders just how a DS9 film could come about. A deep, dark, star trek film is what they need. They need them vulnerable, stuck on that damned station with no weapons, food, power or toilets, and with the Cardassians attacking them. Screw the watered down philosophical crap they've been trying to throw at us for quite some time now, if they want to save Star Trek, they need guns and more women. Find some way to get Seven of Nine into a DS9 film, wearing a low-cut top and hot pants rather than getting William Shatner into Enterprise. Then they'd be onto a winner.
I've seen images that have been encoded using fractal compression; the compression ratios achievable are pretty damn good:) I seem to recall the issue being that the encode was difficult and *very* processor intensive.
Although I didn't see it, the lecturer talking about this at the time (he was researching in this area) said he'd seen fractal encodings of images which pull out more detail than was actually in the image that was encoded. Sounds like crazy talk to me though;)
Yes, I'm aware that there is a lot more to it than I put forward; it was a general example to the guy who thought that big O notation was a "slipshod manner" of evaluating the performance of an algorithm. As you also made clear with your last point, the line between the Big O notation and where real-life engineering blurs when you start to consider just what an algorithm is working with (your slow-medium example). That was already encapsulated within my comment about different architectures/data types/etc, although the link wasn't entirely clear, I'll admit:)
"Really? Even if you could whip up the O(n^2) one in half an hour and have a working product today rather than two weeks from now when the O(n) algorithm's complexity is finally hashed out in code?"
If it really would take you that much longer to implement the quicker algorithm, you probably shouldn't be in that engineering position you speak of;)
The discussion is not on real life engineering, however, we're talking about computing science.
Big O notation simply isn't intended to be used as an indication of just how fast the algorithm is going to run given a sample data set - there's no point, because different implementations of the same algorithm in different languages using different data structures on different architectures renders any approximation of running time redundant. Big O notation is a guideline to the performance of the algorithm given certain sizes of data set. It's a more abstract concept which goes hand in hand with the abstract concept of algorithmics, but provides some basis in reality for reasoning about the performance of an algorithm.
If you have two sort algorithms, one is O(n^2), another O(n), you'd want to implement/use the O(n) one; you can see at a glance just which algorithm is the better of the two.
It's a tool used in the field of algorithmics, a field which is definately not my speciality, and evidently not yours either;)
Slashdot Mirror
There are UDP based congestion control protocols out there, so UDP doesn't necessarily have to be "send as fast as you like"; it's just up to the application developer to think a bit more.
eg: TFRC, ftp://ftp.isi.edu/in-notes/rfc3448.txt
Like most people on slashdot?
... oh, never mind.
Thanks, I'll be
That it is another method of distribution, a much more advanced one.
We knew that long ago. Look at the Usenet for example, running around 25 years ago. Peer-to-peer systems have been used for distribution of information for quite some time, it's just that it's been thrown back into the limelight with the illegal sharing of copyrighted content.
This is the first app to give P2P legitimacy.
... Usenet, 1979, anyone?
Peer-to-peer technology has been around and has been in common usage for many, many years. I can't name any well known peer-to-peer systems before 1979, but I'm quite sure there were some.
P2P had legitimacy long before illegal file sharing came along.
In fact, scratch that.
:)
The mirror seems to have part, but not all of, the file. I get 8 and a half megs through the download each time using wget, and then it stalls.
It was going damned fast up until that point though
And, therefore:
/ news/2003_10_liamkemp/TWL360x208.mov
http://media01.cgchannel.com.nyud.net:8090/images
... Pink Floyd ...
And he's only made 91 posts in his time here :)
The article might be objective, but he who submitted the story does not seem to be quite so objective.
:)
Of course, I haven't read all of TFA, so perhaps I'm doing "Roblimo" an injustice
Having only seen Pi for the first time a few weeks ago, the parent's post was exactly my thoughts on reading the slashdot blurb :)
That struck me as being more of a teaser than a trailer.
:)
Oh well
... no computers are allowed to connect to the network unless authorised by the support guys. They're very strict about letting student's personal computers or laptops connect to the network, so much so I can only remember once where they allowed the student to use their own laptop with the department's network (and they're postgrad, so they don't exactly blend into the sea of undergrads).
:)
Presumably they're only allowing certain MAC addresses to connect, so I don't know if they can detect somebody who's reprogrammed their NIC's MAC address. That said, if they were noticed (and the labs we use are all monitored with CCTV), their place on the course would be seriously questioned, as it would probably be a substantial breach of the conditions of use all students in the department sign at the start of the academic year.
The labs are fully featured enough for most tasks anyway. There are a couple of hundred computers for undergrads, and loads of other computers littered throughout the department, and a few specialist labs. The network connects all manner of systems, from all many of Windows installs, to Mac OS X, to loads of Linux systems, and a few BSD machines running in the background, and probably others I've never encountered (it is a CS department, after all!). The department sees no need to put the network in jeopardy by letting any computer connect, rather than just the department's own
... although I seem to recall that little factoid gets overlooked because there really aren't that many homes in the UK connected via fibre.
Personally I tend to write UK as my country if asked, but Scottish as nationality is asked.
;) What really ticks us off are the folks who think England *is* the UK.
Varies from person to person
Because it makes it sound complex. :)
A lot of "bloat" comes from increasing levels of abstraction away from the hardware, and numerous levels of interfaces between different system layers to allow for more modular code, more reusable code, and generally, safer code. Coding an application for a modern machine with all its fancy hardware would be pretty tricky without these interfaces already written for people. Sure, somebody has got to write the low-level interaction gubbins in the first place, but once it's done, it's done. If software houses wrote for the hardware to reduce bloat, we'd never see any games or software. Praise the bloat. The bloat be praised.
[That's not to say that an understanding of what the machine is actually doing with the code wouldn't help people when designing software systems, I agree that it would certainly be beneficial. It's just not necessarily where all the bloat comes from. The lack of understanding is probably more and more relevant as more and more layers of abstraction are added. And there's always room for one more layer of abstraction!]
It's true. But of all the series, I've grown to like DS9 more than the others. For one, at the start of the series they were in this clapped out station, and certainly weren't in the same position of power the previous series had seen the central characters in. The later series had story arcs - imagine that! A story that was really carried from episode to episode. No other series has had that (well, Voyager had a goal that encapsulated the entire series, and TNG toyed with the Maquis (sp?) at one point and finished off where they started with Q, but none have really done what DS9 did).
That said, I pretty much stopped watching toward the end of DS9/mid Voyager. Not really sure why, it was probably that I was busy and just didn't want to sit in and watch TV in my free moments. I wouldn't mind going back and watching the DS9 episodes at all; I watched TNG when it was repeated ad finitum on BBC2, Voyager and Enterprise I'd only watch if I happened to catch it and I didn't have anything else to do. Same could probably be said about some ToS episodes.
/me wonders just how a DS9 film could come about. A deep, dark, star trek film is what they need. They need them vulnerable, stuck on that damned station with no weapons, food, power or toilets, and with the Cardassians attacking them. Screw the watered down philosophical crap they've been trying to throw at us for quite some time now, if they want to save Star Trek, they need guns and more women. Find some way to get Seven of Nine into a DS9 film, wearing a low-cut top and hot pants rather than getting William Shatner into Enterprise. Then they'd be onto a winner.
Oh God, please don't let this be another lame Slashdot joke we'll be reciting in 2 years time... ;)
I've seen images that have been encoded using fractal compression; the compression ratios achievable are pretty damn good :) I seem to recall the issue being that the encode was difficult and *very* processor intensive.
;)
Although I didn't see it, the lecturer talking about this at the time (he was researching in this area) said he'd seen fractal encodings of images which pull out more detail than was actually in the image that was encoded. Sounds like crazy talk to me though
Let's link to the article again!
http://www.phonescoop.com/news/item.php?n=853
"Actually there is more to it than that."
:)
Yes, I'm aware that there is a lot more to it than I put forward; it was a general example to the guy who thought that big O notation was a "slipshod manner" of evaluating the performance of an algorithm. As you also made clear with your last point, the line between the Big O notation and where real-life engineering blurs when you start to consider just what an algorithm is working with (your slow-medium example). That was already encapsulated within my comment about different architectures/data types/etc, although the link wasn't entirely clear, I'll admit
"Really? Even if you could whip up the O(n^2) one in half an hour and have a working product today rather than two weeks from now when the O(n) algorithm's complexity is finally hashed out in code?"
;)
If it really would take you that much longer to implement the quicker algorithm, you probably shouldn't be in that engineering position you speak of
The discussion is not on real life engineering, however, we're talking about computing science.
Big O notation simply isn't intended to be used as an indication of just how fast the algorithm is going to run given a sample data set - there's no point, because different implementations of the same algorithm in different languages using different data structures on different architectures renders any approximation of running time redundant. Big O notation is a guideline to the performance of the algorithm given certain sizes of data set. It's a more abstract concept which goes hand in hand with the abstract concept of algorithmics, but provides some basis in reality for reasoning about the performance of an algorithm.
;)
If you have two sort algorithms, one is O(n^2), another O(n), you'd want to implement/use the O(n) one; you can see at a glance just which algorithm is the better of the two.
It's a tool used in the field of algorithmics, a field which is definately not my speciality, and evidently not yours either
Who said drink excessively? Or are you saying that one sniff of alcohol, and *bang*, there goes your memory for the next 3 hours of your life? ;)
Off topic, I know, but why on Earth would anybody regularly reinstall their OS of choice once a month?