It is scalable as far as CPU power goes, as long as you can parallelise your application - many can't easily be made to run in a parallel environment. Additionally, it is not scalable in terms of addressable memory (at least if you want huge bandwidth to that memory), and in many HPC environments you really do need a big shared memory model with massive bandwidth - something that a cluster of boxes just isn't going to give you without some serious customisation (ala Cray).
Mainframes (ok, I'm abusing the term here to include the like of HP Superdomes) are just as popular as ever, possibly more so. It is just that small scale commodity computing is now massive whereas before it was a small player.
When I visit the USA I just flip the switch. Probably because here in the UK I have two way circuits, except in the bedrooms, so I'm used to the switch sometimes working the wrong way.
If this guy can do it, then any other user or, more likely, kids can do it which means it is more likely to do it again when in flight. If it's likely to overheat (after that Swiss disaster I doubt it would be designed in a way that it could) then this guy finding it and bringing it to the publics attention has to be a good thing.
As an aside, I've had the whole plane entertainment system on Emirates crash on me twice (two different planes on the way to Australia) - and I didn't do anything other than try and watch a film.
With respect to Verified By Visa - most people haven't got a clue what it is. It isn't well marketed by Visa and banks using Visa cards, and I suspect most users suspect it to be some kind of scam.
But regardless, as you say there is little need for the consumer to bother with it. I've seen things about it from my bank, but I've always ignored it as it doesn't give me anything I don't have already.
The UK has 60+ million people (it is far more than that, but 60million is close to the official figure of people registered here). Most of the unbundled lines are in the centre of London, Birmingham, Manchester and Cardiff, and unfortunately the majority of the population do not live in the centre of those cities.
3.5mbit is a pretty decent speed at the moment. If we want to start streaming HD quality video then 3.5mbit certainly will not be enough.
The population isn't *that* bad no, but it is a lot worse than it could have been if the Thatcher government had allowed BT to install fibre in the 80's. Compared to countries like Hong Kong and South Korea the UK is pretty poor - those countries still believe that investment and working hard is the way to better oneself. The UK is all about doing as little as possible for as much as possible, IMO:)
I disagree; I think you are lucky to live that close to an exchange. The majority of the population struggle to get much above 2mbit ADSL on conventional ADSL and are not expected to get anything close to 16 or 22Mbit from the new offerings. There are stats on the BT site and on sites such as SamKnows / ADSLGuide that help back this up, but I'm too tired to find the links at the moment:) I agree with you on cable, but only a fraction of the population can get that again.
WRT sticking with BT, most people have no choice. Few exchanges have been local loop unbundled outside of the major cities - it just isn't cost effective for companies to go and install their kit - and this forces people onto IPStream, and at that point you are stuck with BT. As I said, I live in Cheltenham which is an affluent area with a highly technical workforce (a number of large technical organisation have offices here) and "broadband" take up is very high. The services are still pretty poor unless you live in the town center. I'm on a new development 3km from the exchange, and I'm lucky to have a brand new copper line to get 3.5mbit. A friend slightly closer to the exchange (and with a shorter cable run) can only get 2mbit when it's wet. when it is dry his connection drops to well under 1mbit. Aluminium cables from the 60's rock.
The free broadband on subscription thing is great while it works. Loads of my colleagues have been stung by that, and have had to resort to for weeks at a time until engineers get around to fixing problems. I value my internet connection too much so at the moment I'd rather pay a premium, even if it is a rip off:(
Here in the UK we have 8mbit max typically on both cable and dsl, with most people on dsl getting between 1 and 4mbit due to the poor quality cables and distances from exchanges.
We are just getting 22mbit in some areas, but even then most people get far less because of the cabling issues. Fibre is essentially non-existant.
In order to get good technical support I'm paying £25 for 20GB download cap on a "maxadsl" 8mbit package, and get a 3500kbit sync rate. Typically one could get a 50GB cap product for the money I'm paying, or an unlimited package which would seems to drop to 200kbit throughput at peak times (evenings).
I live in a very affluent area of the UK (Cheltenham), with a population of around 100,000. I envy the connections my friends get in the US.
x86_64 is big on the desktop. AMD64 is very popular, as is Core2 duo.
Sparc is very popular on desktops in some government arenas, and still has some strangleholds in academia. Non-x86_32 platforms are a minority, but a rapidly growing one thanks to em64t etc.
I really don't see what all the fuss is about. I read my news through the Google News "portal" because it pushes things into one place. I usually end up visiting the sites of the papers in question to view the articles directly, or to see what else was said by them in the past. I wouldn't have bothered to visit all those hundreds of news sites individually so the sites would have lost out on my hits.
The same applies with Slashdot, Digg etc. - although that's worse, because I normally just skim the summary and top comments unless I think the summary might have missed something. When is Slashdot going to get dragged through the courts?
Yes, the money situation is bad, but I don't think it will be a long term issue.
So many organisations are making use of Wikipedia data and the Mediawiki software for alsorts of things. The amount of research going on using the wikipedia corpus is staggering - check out the major journals such as the ACM (or look on Citeseer).
If wikipedia was really in trouble I think that most organisations using the data would donate - both commercial and academic. The data in Wikipedia is just too valuable for so many things - and as time goes on the project could expand into a massively near-parallel corpus of text in non-trivial subjects (it already is in some subjects). That, coupled with the semantic detail contained within, makes it a hugely valuable resource that I don't think will be allowed to die.
People are already working on this - primarily as a means of serving up video material though. No reason it couldn't be used for the articles too. I guess with P2P you have the problem of knowing whether an article is the genuine thing or not - someone could malicously remove all the previous history and say their article was the definitive etc.
There is a website about it somewhere (it has quite a novel way of navigating wikipedia content) but I cant' remember the url - it is french though, and might be in that paper
No - you running any download binaries on my computer without my permission would seriously wind me up. Install whatever crap you like on your own machines, but don't do it on mine.
To some platforms. Flash9 support only just came out for x86 linux, there is no support x86_64, PPC etc, and no support for Solaris - let alone anything more unusual.
Java applets can look exactly how you want them. If you want them to look as disfunctional as most deployed Flash and design your own widgets then you can.
Unless this was exactly what they were trying to model. Seeing how a popular service grows, and the type of person involved (geeks, researchers, "real users", etc) could be a very useful thing to do if you plan on giving away services in future and want to see the kind of infrastructure involved (think Google productivity suite or YouTube and modelling what kind of hardware would you need to support those as they grow).
I'd go absolutely bloody nuts if you used my computer and installed any software on it. I'm sure most people with a clue would feel the same - especially businesses.
hey there, sorry to bug ya, i'll be quick
mhm
bring two packs of diapers, two baguettes, ham, cheese, lettuce, half a kilo of grinded meat.
ok, i'll do that
there is no need to snap
i'm not snapping, i'm just a bit busy trying to get this algorithm to scale
oh, so now you don't even have time to talk to me?
I can talk to you when I get home, I need to work right now
oh I see, well good luck!
thanks
cya
wait - what was that list ag....
I think we just said the same thing. Money is needed to prove effectiveness (and safety, feasability, etc). The GP said that the foundation would provide money, IF it turns out to be effective. No money = no proof = no drugs.
So you signed up for the wrong course. Computer Science, or Computing Science, is really Computing / Computational Mathematics. If you didn't research what was involved in computer science then you can't blame the university, only yourself.
Computer science (and possibly some related disciplines) are dying because there are more and more dumbed down courses which people apply for instead because they are at universities that are more accesisble. People are generally lazy - the explosion in people asking homework questions on programming forums and IRC is testament to that - and if they think they can get away without doing something, generally they will.
Hard computer science problems require lots of math, you can't escape that, whether it be graphics, artificial intelligence, classification, whatever. The more I work, the more I realise just how much math I really should know, and how little that I do. There are areas of using and designing for computers that don't require math - but these are mainly psychology related. Once you get down to what is going on under the hood, it is math.
There is a reason that major employers are picky about the grades, courses and universities that they recruit from.
Slashdot Mirror
It is scalable as far as CPU power goes, as long as you can parallelise your application - many can't easily be made to run in a parallel environment. Additionally, it is not scalable in terms of addressable memory (at least if you want huge bandwidth to that memory), and in many HPC environments you really do need a big shared memory model with massive bandwidth - something that a cluster of boxes just isn't going to give you without some serious customisation (ala Cray).
Mainframes (ok, I'm abusing the term here to include the like of HP Superdomes) are just as popular as ever, possibly more so. It is just that small scale commodity computing is now massive whereas before it was a small player.
When I visit the USA I just flip the switch. Probably because here in the UK I have two way circuits, except in the bedrooms, so I'm used to the switch sometimes working the wrong way.
If this guy can do it, then any other user or, more likely, kids can do it which means it is more likely to do it again when in flight. If it's likely to overheat (after that Swiss disaster I doubt it would be designed in a way that it could) then this guy finding it and bringing it to the publics attention has to be a good thing.
As an aside, I've had the whole plane entertainment system on Emirates crash on me twice (two different planes on the way to Australia) - and I didn't do anything other than try and watch a film.
With respect to Verified By Visa - most people haven't got a clue what it is. It isn't well marketed by Visa and banks using Visa cards, and I suspect most users suspect it to be some kind of scam.
But regardless, as you say there is little need for the consumer to bother with it. I've seen things about it from my bank, but I've always ignored it as it doesn't give me anything I don't have already.
The UK has 60+ million people (it is far more than that, but 60million is close to the official figure of people registered here). Most of the unbundled lines are in the centre of London, Birmingham, Manchester and Cardiff, and unfortunately the majority of the population do not live in the centre of those cities.
:)
3.5mbit is a pretty decent speed at the moment. If we want to start streaming HD quality video then 3.5mbit certainly will not be enough.
The population isn't *that* bad no, but it is a lot worse than it could have been if the Thatcher government had allowed BT to install fibre in the 80's. Compared to countries like Hong Kong and South Korea the UK is pretty poor - those countries still believe that investment and working hard is the way to better oneself. The UK is all about doing as little as possible for as much as possible, IMO
I disagree; I think you are lucky to live that close to an exchange. The majority of the population struggle to get much above 2mbit ADSL on conventional ADSL and are not expected to get anything close to 16 or 22Mbit from the new offerings. There are stats on the BT site and on sites such as SamKnows / ADSLGuide that help back this up, but I'm too tired to find the links at the moment :) I agree with you on cable, but only a fraction of the population can get that again.
:(
WRT sticking with BT, most people have no choice. Few exchanges have been local loop unbundled outside of the major cities - it just isn't cost effective for companies to go and install their kit - and this forces people onto IPStream, and at that point you are stuck with BT. As I said, I live in Cheltenham which is an affluent area with a highly technical workforce (a number of large technical organisation have offices here) and "broadband" take up is very high. The services are still pretty poor unless you live in the town center. I'm on a new development 3km from the exchange, and I'm lucky to have a brand new copper line to get 3.5mbit. A friend slightly closer to the exchange (and with a shorter cable run) can only get 2mbit when it's wet. when it is dry his connection drops to well under 1mbit. Aluminium cables from the 60's rock.
The free broadband on subscription thing is great while it works. Loads of my colleagues have been stung by that, and have had to resort to for weeks at a time until engineers get around to fixing problems. I value my internet connection too much so at the moment I'd rather pay a premium, even if it is a rip off
Here in the UK we have 8mbit max typically on both cable and dsl, with most people on dsl getting between 1 and 4mbit due to the poor quality cables and distances from exchanges.
We are just getting 22mbit in some areas, but even then most people get far less because of the cabling issues. Fibre is essentially non-existant.
In order to get good technical support I'm paying £25 for 20GB download cap on a "maxadsl" 8mbit package, and get a 3500kbit sync rate. Typically one could get a 50GB cap product for the money I'm paying, or an unlimited package which would seems to drop to 200kbit throughput at peak times (evenings).
I live in a very affluent area of the UK (Cheltenham), with a population of around 100,000. I envy the connections my friends get in the US.
x86_64 is big on the desktop. AMD64 is very popular, as is Core2 duo.
Sparc is very popular on desktops in some government arenas, and still has some strangleholds in academia. Non-x86_32 platforms are a minority, but a rapidly growing one thanks to em64t etc.
I really don't see what all the fuss is about. I read my news through the Google News "portal" because it pushes things into one place. I usually end up visiting the sites of the papers in question to view the articles directly, or to see what else was said by them in the past. I wouldn't have bothered to visit all those hundreds of news sites individually so the sites would have lost out on my hits.
The same applies with Slashdot, Digg etc. - although that's worse, because I normally just skim the summary and top comments unless I think the summary might have missed something. When is Slashdot going to get dragged through the courts?
Yes, the money situation is bad, but I don't think it will be a long term issue.
So many organisations are making use of Wikipedia data and the Mediawiki software for alsorts of things. The amount of research going on using the wikipedia corpus is staggering - check out the major journals such as the ACM (or look on Citeseer).
If wikipedia was really in trouble I think that most organisations using the data would donate - both commercial and academic. The data in Wikipedia is just too valuable for so many things - and as time goes on the project could expand into a massively near-parallel corpus of text in non-trivial subjects (it already is in some subjects). That, coupled with the semantic detail contained within, makes it a hugely valuable resource that I don't think will be allowed to die.
People are already working on this - primarily as a means of serving up video material though. No reason it couldn't be used for the articles too. I guess with P2P you have the problem of knowing whether an article is the genuine thing or not - someone could malicously remove all the previous history and say their article was the definitive etc.
f
Tagging in p2p wikipedia - http://www.emse.fr/OSIR06/2006-osir-p39-fokker.pd
There is a website about it somewhere (it has quite a novel way of navigating wikipedia content) but I cant' remember the url - it is french though, and might be in that paper
No - you running any download binaries on my computer without my permission would seriously wind me up. Install whatever crap you like on your own machines, but don't do it on mine.
To some platforms. Flash9 support only just came out for x86 linux, there is no support x86_64, PPC etc, and no support for Solaris - let alone anything more unusual.
Linux x86_64 (or anything else, Sparc etc)
Solaris x86, x86_64, Sparc
These are surprisingly big markets in some arenas.
Java applets can look exactly how you want them. If you want them to look as disfunctional as most deployed Flash and design your own widgets then you can.
Unless this was exactly what they were trying to model. Seeing how a popular service grows, and the type of person involved (geeks, researchers, "real users", etc) could be a very useful thing to do if you plan on giving away services in future and want to see the kind of infrastructure involved (think Google productivity suite or YouTube and modelling what kind of hardware would you need to support those as they grow).
I'd go absolutely bloody nuts if you used my computer and installed any software on it. I'm sure most people with a clue would feel the same - especially businesses.
Nah
hey there, sorry to bug ya, i'll be quick
mhm
bring two packs of diapers, two baguettes, ham, cheese, lettuce, half a kilo of grinded meat.
ok, i'll do that
there is no need to snap
i'm not snapping, i'm just a bit busy trying to get this algorithm to scale
oh, so now you don't even have time to talk to me?
I can talk to you when I get home, I need to work right now
oh I see, well good luck!
thanks
cya
wait - what was that list ag....
damn.
And unless Operating System B is even more crippled (hard to imagine, I know, but maybe it's possible).
I think we just said the same thing. Money is needed to prove effectiveness (and safety, feasability, etc). The GP said that the foundation would provide money, IF it turns out to be effective. No money = no proof = no drugs.
How is that going to help? They need the money to find out whether or not it actually is effective?
So you signed up for the wrong course. Computer Science, or Computing Science, is really Computing / Computational Mathematics. If you didn't research what was involved in computer science then you can't blame the university, only yourself.
Computer science (and possibly some related disciplines) are dying because there are more and more dumbed down courses which people apply for instead because they are at universities that are more accesisble. People are generally lazy - the explosion in people asking homework questions on programming forums and IRC is testament to that - and if they think they can get away without doing something, generally they will.
Hard computer science problems require lots of math, you can't escape that, whether it be graphics, artificial intelligence, classification, whatever. The more I work, the more I realise just how much math I really should know, and how little that I do. There are areas of using and designing for computers that don't require math - but these are mainly psychology related. Once you get down to what is going on under the hood, it is math.
There is a reason that major employers are picky about the grades, courses and universities that they recruit from.
You replied totally out of context. What on earth has your post got to do with my example of using security cameras?
Regardless, the GPL is not in the spirit you describe; BSD and MIT licences are more aligned to that. The GPL is far more restrictive.
Maybe you could help fix the bugs.
They have stolen parts of their software rampantly for years now. Can you blame those wanting to steal theirs?