I agree with the above poster. There's two things I'd add.
Firstly, right now, the important thing is to keep a record of absolutely everything - dates/times of conversations, what was said, who you talked to etc. You'll need it.
Secondly, after you've exhausted the other options (writing to PC World, trading standards) then don't be afraid of using the small claims court. It's very easy: no lawyer is required, the fee is only £50 (which PC World will pay if you win) and the paperwork is minimal. I'd guess that in 99% of cases the company will pay up before it actually gets as far as court. In my case my old landlady returned our deposit (plus the £50) the day after I'd filed in court!
Yeah, I've got one too, it's truly dire. But I bet you've got a Vodafone one like me. They took a really old build of the M600i software, took away the good bits, and replaced them with crappy, crashing branded bits. Worse still, they've locked down the firmware so you can't reflash it with a non-crahsing version. (I have used 'plain-old' unbranded M600is for long periods too and they're a different world).
The vast majority of content is in the US, so traffic has to come across highly expensive transoceanic links.
In the UK, the cost of this is about 4p / MB. I imagine that works out at about 1/3 the cost they're charging in Australia - and they have a much bigger ocean to cross.
I ran a large ethernet in my halls of residence (dorms) for several years. We were charged 2p/MB by our upstream provider, and found that we just couldn't cover that cost within a reasonable monthly subscription.
We fought tooth and nail to pass on that cost to the consumer - that was the right option! But in the end we had to give up, and put a very harsh throttle on the whole line.
Telstra are doing the right thing here. The alternative is to limit everyone's use, even that vast majority who don't use > 3 GB. I'm sure Telstra had to work quite hard to make this solution happen, and I'm impressed they did.
Some systems exist to extract facts from language into semantic knowledge representations, and they're surprisingly good.
SNOWY is a system that "reads" the World Book Encyclopaedia and stores each fact about a concept into a hierarchic memory based on that concept. It's sufficiently sophisticated to be able to realise that "The bear digs up the nut" implies that the bear eats the nut, while "The miner digs up the coal" doesn't imply that. You can then ask it "what eats nuts" and it will reply correctly. (At least, this is my impression - I haven't used it, sadly.) As I remember it can fully understand 50-60% of the sentences in the bits of the encyclopaedia that it has been commanded to parse.
The language it works on is fairly simple, but is nevertheless text designed for humans as opposed to computers. Systems like this could be a good bridge between language and semantic based representations.
There are also, of course, dozens of systems designed to work on English text that has been specifically created to be computer-parsable, but still readable by humans.
I'm incredibly sceptical about all this sort of technology, but if the systems continue to evolve, the agents might be able to glean much of their knowledge from existing web pages.
PowerMac G4s, for example, have no analog connection coming out of the CD drive. The CD player software works by ripping the audio data across the IDE/SCSI/USB bus and then feeding it out of the sound card. That won't work with burn-proof CDs.
The problem will get bigger with, for instance, the proliferation of USB speakers, where all data has to be transferred digitally all the way.
Hopefully the population using such schemes will become large enough that the move will be politically impossible by the time the technology is there.
As I understood it the CDDB identified track names based on the overall signature of a CD, including its ID number and the lengths of all the tracks.
How can this be used to identify individual MP3 tracks?
And, if it is possible, why do we not have a tool which fills in the ID3 tags for totally unlabelled MP3 files, just from the CDDB. This would be very useful. Surely if this were possible, it would have been done?!
It uses the Linux ethertap driver to create a virtual ethernet within your computer. You can then set up a router (including, if you wish, masquerading) between this and your normal ethernet card - or PPP connection - or whatever.
If that's too complex you can instead just attach it to your regular ethernet interface, though obviously you would then need two IP addresses. I don't know whether it uses promisc mode for this.
There's been a LinuxPPC Mac virtualizer, Mac on Linux, for quite a while longer than plex86 has existed.
What have you learnt from MOL? Is it easier to virtualise an OS on PPC (if so, why?) Are the necessary kernel modifications between the two packages similar, and should/could there be some sort of generic kernel virtualiser API across all flavours of Linux? (i.e. are any bits of it not CPU dependent?)
Do you see plex86 and MOL co-operating on, for instance, the Linux side of a 3D graphics card emulator? (That would be extremely cool!) Or can you use the ethernet/ethertap code from MOL to enhance plex86's networking?
I'd love to, but: -- It uses shaper not IProute2 so it's obsolete -- It's so full of hacks you wouldn't believe it -- Some of the more commercially-minded professors wanted to sell our stuff to other universities(!?!) -- I don't work there any more! So that's fairly conclusive. BUT: I've gone back to studenthood for a bit, and I'm trying to persuade my tutor that something like this would be suitable for one of my projects. So it could happen yet... but it would need to be totally rewritten for IProute2. IProute2 is too flexible for its own good: because there's so many possible permutations of options, the documentation can't really help you find what you're looking for (and that's not an insult to the docs.) So a good IProute2 website would be a great start, if anyone's got loads of free time and a spare busy network to experiment with:-)
I was partly responsible for QoS at a UK university, where things are worse because we have to pay 2 pence (4 cents??) per MB for all traffic received from the US (with a 56KBps video stream that adds up!) In the first year we left it open and made a HUGE loss. (We charged 30 pounds per year per student).
It's a tough problem (obviously). The best solution we came up with was to use DummyNet under BSD or iproute2 under Linux to give each user their own allocation of bandwidth. When they run out, they can still use our proxy servers (which don't incur charges) or can buy more bandwidth for their allocation.
IProute2 is actually excellent for this. It can do just as much as your average Cisco, much more easily: source-based routing, processing of packets based on arbitrary hexadecimal strings in them, and so-on. With a powerful enough CPU and two 3Com cards, we got a decent throughput too.
We came up with a whole complex system with perl, Oracle, DBI, SNMP, shaper.o (no iproute2 in those days) and lots of other things - then ran out of time and money just as it was starting to work (though shaper.o wasn't very suited to the task). There just isn't enough money, at least in UK universities, to do this sort of thing.
Instead each ethernet segment of 100+ users squeezes through a 20Kbps throttle. This is of course totally unfair, because 2% use more than the rest put together, but on the bright side traffic through our proxies is excluded from the throttling. It's a terrible solution but there isn't money for anything better. We don't have Cisco CPU capacity for selective QoS by protocol. Any suggestions welcome:-)
Slashdot Mirror
I agree with the above poster. There's two things I'd add.
Firstly, right now, the important thing is to keep a record of absolutely everything - dates/times of conversations, what was said, who you talked to etc. You'll need it.
Secondly, after you've exhausted the other options (writing to PC World, trading standards) then don't be afraid of using the small claims court. It's very easy: no lawyer is required, the fee is only £50 (which PC World will pay if you win) and the paperwork is minimal. I'd guess that in 99% of cases the company will pay up before it actually gets as far as court. In my case my old landlady returned our deposit (plus the £50) the day after I'd filed in court!
Yeah, I've got one too, it's truly dire. But I bet you've got a Vodafone one like me. They took a really old build of the M600i software, took away the good bits, and replaced them with crappy, crashing branded bits. Worse still, they've locked down the firmware so you can't reflash it with a non-crahsing version. (I have used 'plain-old' unbranded M600is for long periods too and they're a different world).
Another Symbian OS phone with GPRS, MMS, Java MIDP, and a digital camera - but this one is tri-band so works in the US!
Page with 3D animation
News story with details and lots of pictures (on subsequent pages)
(Bias: I work for Symbian)
Apparently it was named octothorpe because it looked like the street pattern of an English village. Can't remember where I read that.
In the UK, the cost of this is about 4p / MB. I imagine that works out at about 1/3 the cost they're charging in Australia - and they have a much bigger ocean to cross.
I ran a large ethernet in my halls of residence (dorms) for several years. We were charged 2p/MB by our upstream provider, and found that we just couldn't cover that cost within a reasonable monthly subscription.
We fought tooth and nail to pass on that cost to the consumer - that was the right option! But in the end we had to give up, and put a very harsh throttle on the whole line.
Telstra are doing the right thing here. The alternative is to limit everyone's use, even that vast majority who don't use > 3 GB. I'm sure Telstra had to work quite hard to make this solution happen, and I'm impressed they did.
SNOWY is a system that "reads" the World Book Encyclopaedia and stores each fact about a concept into a hierarchic memory based on that concept. It's sufficiently sophisticated to be able to realise that "The bear digs up the nut" implies that the bear eats the nut, while "The miner digs up the coal" doesn't imply that. You can then ask it "what eats nuts" and it will reply correctly. (At least, this is my impression - I haven't used it, sadly.) As I remember it can fully understand 50-60% of the sentences in the bits of the encyclopaedia that it has been commanded to parse.
The language it works on is fairly simple, but is nevertheless text designed for humans as opposed to computers. Systems like this could be a good bridge between language and semantic based representations.
This is the best link I can find, unfortunately.
There are also, of course, dozens of systems designed to work on English text that has been specifically created to be computer-parsable, but still readable by humans.
I'm incredibly sceptical about all this sort of technology, but if the systems continue to evolve, the agents might be able to glean much of their knowledge from existing web pages.
PowerMac G4s, for example, have no analog connection coming out of the CD drive. The CD player software works by ripping the audio data across the IDE/SCSI/USB bus and then feeding it out of the sound card. That won't work with burn-proof CDs.
The problem will get bigger with, for instance, the proliferation of USB speakers, where all data has to be transferred digitally all the way.
Hopefully the population using such schemes will become large enough that the move will be politically impossible by the time the technology is there.
Aha! I was thinking more along the lines that somebody would write an open-source tool to do this, using FreeDB if not CDDB.
As I understood it the CDDB identified track names based on the overall signature of a CD, including its ID number and the lengths of all the tracks. How can this be used to identify individual MP3 tracks? And, if it is possible, why do we not have a tool which fills in the ID3 tags for totally unlabelled MP3 files, just from the CDDB. This would be very useful. Surely if this were possible, it would have been done?!
Have a look at Mac-on-Linux, the LinuxPPC virtualizer equivalent:
Mac-on-Linux
It uses the Linux ethertap driver to create a virtual ethernet within your computer. You can then set up a router (including, if you wish, masquerading) between this and your normal ethernet card - or PPP connection - or whatever.
If that's too complex you can instead just attach it to your regular ethernet interface, though obviously you would then need two IP addresses. I don't know whether it uses promisc mode for this.
There's been a LinuxPPC Mac virtualizer, Mac on Linux, for quite a while longer than plex86 has existed.
What have you learnt from MOL? Is it easier to virtualise an OS on PPC (if so, why?) Are the necessary kernel modifications between the two packages similar, and should/could there be some sort of generic kernel virtualiser API across all flavours of Linux? (i.e. are any bits of it not CPU dependent?)
Do you see plex86 and MOL co-operating on, for instance, the Linux side of a 3D graphics card emulator? (That would be extremely cool!) Or can you use the ethernet/ethertap code from MOL to enhance plex86's networking?
What, in general, do you think of MOL?
I'd love to, but: :-)
-- It uses shaper not IProute2 so it's obsolete
-- It's so full of hacks you wouldn't believe it
-- Some of the more commercially-minded professors wanted to sell our stuff to other universities(!?!)
-- I don't work there any more!
So that's fairly conclusive. BUT: I've gone back to studenthood for a bit, and I'm trying to persuade my tutor that something like this would be suitable for one of my projects. So it could happen yet... but it would need to be totally rewritten for IProute2. IProute2 is too flexible for its own good: because there's so many possible permutations of options, the documentation can't really help you find what you're looking for (and that's not an insult to the docs.) So a good IProute2 website would be a great start, if anyone's got loads of free time and a spare busy network to experiment with
It's a tough problem (obviously). The best solution we came up with was to use DummyNet under BSD or iproute2 under Linux to give each user their own allocation of bandwidth. When they run out, they can still use our proxy servers (which don't incur charges) or can buy more bandwidth for their allocation.
IProute2 is actually excellent for this. It can do just as much as your average Cisco, much more easily: source-based routing, processing of packets based on arbitrary hexadecimal strings in them, and so-on. With a powerful enough CPU and two 3Com cards, we got a decent throughput too.
We came up with a whole complex system with perl, Oracle, DBI, SNMP, shaper.o (no iproute2 in those days) and lots of other things - then ran out of time and money just as it was starting to work (though shaper.o wasn't very suited to the task). There just isn't enough money, at least in UK universities, to do this sort of thing.
Instead each ethernet segment of 100+ users squeezes through a 20Kbps throttle. This is of course totally unfair, because 2% use more than the rest put together, but on the bright side traffic through our proxies is excluded from the throttling. It's a terrible solution but there isn't money for anything better. We don't have Cisco CPU capacity for selective QoS by protocol. Any suggestions welcome :-)