Unfortunately not. The flight I was on appeared not to be formally cancelled on the system, so I wasn't allowed to rebook at the time - and of course their system was jammers anyway. I tried to help out the people either side of me but was pretty unsuccessful.
2) The thing is - and believe me, I do shop around - none of the other options are very much better.:( The national flag carrier has remodelled itself to a low-cost airline and now matches ryanair feature for misfeature. BMI (Baby) are quite good where they fly, but to many other locations options seem to be very expensive, even accounting for ultimate cost including disasters like the above, or nonexistent.
I was due to fly the evening it all went wrong. Here's a lesson: if you're standing in a three-hour queue for the Ryanair desk, and they tell people to rebook on the web, and you take out a laptop and 3G modem, be prepared for a stampede.
The two biggest carriers in IE have, in recent months, both moved to discourage checked baggage on short-haul flights. One already charges per bag checked in, the other recently announced plans to do so. Both are modelled these days as low-cost airlines.
One of the first questions asked on the radio yesterday morning was how these policies will be affected by yesterday's events. Remains to be seen.
Right, I see where you're coming from. Couple of things I can point out. You're right that there's a difference in the amount of time a home user will saturate their connection compared with the traffic profile of a large content provider. That's why ISPs have contention ratios, and the contention ratio that a home user will find perfectly acceptable, even during the busy times, might be far from acceptable at peak times for a site like LJ. So the amount of bandwidth feeding into the user's home is just a single part of the equation, and sharing upstream bandwidth by means of contention ratios means you're already reducing the cost thanks to the fact that most users will spend most of their time idling.
(You can start to see here, btw, why home users who do hammer their connection with peer to peer traffic cause such problem - they're outliers on that model, and may be using much more of the available bandwidth than others while paying the same share.)
With a big content provider, the equation is different because their requirements are different, but at the end of the day they have a bandwidth bill too. Just like the user, it's mostly not really under their control, they take their pick from the available providers with the available service levels.
So whether it's the users or the content provider who pay >50pc of the cost of shipping the bytes isn't I think very important. We each have our bills to pay, and we each have to find our own ways to fund that; your ISP bill doesn't pay for LJ's, or anyone else who runs a web serer.
If a site tries ads, then, that's one way to pay the bills, but it'll only work of course if people can see them. You're perfectly entitled not to use the service if you don't consent to that method of funding, of course. In the case of LJ you can stick with the old free level of service where you'll rarely see any ads at all.
But is it kosher to take the level of service that they offer with ads, then remove the ads, consuming the service anyway? Not just for LJ - if a site uses ads to pay its bills, and you don't consent to the ads, is it okay remove them but continue to consume the service?
Ok, you're not far from the truth there (full disclosure: I work in the ISP industry, but I framed the question that way because I wanted to work out where you were coming from) but I still have a couple of questions.
You said a couple of posts ago that LJ only have to pay for their path to the next router along. Now I think you say - correct me if I'm wrong - that the cost of the whole path is passed on to both parties and is shared between LJ and the individuals. That's cool, I can agree with that. And you're spot on that cost per meg goes down when you buy in bulk, much like cost per unit goes down when you buy in bulk in any industry.
Thing is, when you're buying in bulk, you're still handing over more actual hard cash than the tiddlers. So LJ and the user both get a bill to access the internet, but LJ's bill, I think it's fair to assume, is a heck of a lot bigger than mine. That's a cost of doing business, just like the cost of maintaining their servers and paying their staff, and I suspect we can both agree on that.
So my questions are: do you reckon they have a right to recoup this cost? And if so, should they really get off the internet in preference to taking on ads?
It's a nice theory, but it hinges on this assumption:
the divisions should be geographic rather than corporate
...which is an oversimplification of the ISP market. Networks have ISPs, and this is transitive up to Tier 1s. Think carefully about the relationship here - you pay ISPs to transit your packets in both directions, and they pay bigger ISPs for onward transit.
This is not a geographical relationship. Two ISPs in a given country may (and are likely to) have independent connectivity. If someone in the US sends a packet bound for "Ireland", it gets sent to the upstream ISP of that user, not to to "Ireland".
So we come to the consequence of what you propose. Do you really want to regulate that every ISP in a geographic region must share upstream connectivity?
The problem is that if IP addresses are well aggregated, all a BGP-speaking router (that's the big ones in the core) needs to know is "for this very large block of addresses, use interface A; for that very large block of addresses, use interface B; for this other very large block of addresses, use interface A again." That is your routing table, it takes processor time to traverse for every packet, and it's growing; and if sizeof(routing table)*sizeof(traffic throughput) grows faster than Moore's law, it gets rather troublesome for the internet.
If you route geographically or per end-user or (shudder) per person, the number of entries that your core router has to potentially traverse explodes. This is the essence of CIDR, and we have separate naming (i.e. DNS) and routing (i.e. IP addresses) specifically so that end users may have a portable name irrespective of the routing infrastructure.
In the phone system, where naming and addressing are both conflated into your phone number, it's a lot more painful. (All of a sudden there isn't a simple programmatic way of mapping a three-digit prefix to to the operator that will handle the call.)
The problem of routing table size remains regardless of the size of the IP space - IPv6 will solve a lot of problems, but this isn't one of them.
Forwarding a packet out an interface is a fairly straightforward task. Things get interesting, though, when you start scaling up - either speed or size. Putting together hardware and software that can handle 120,000-odd routes (and growing) in the IPv4 routing table, all of which may go up and down independently and quickly, is a challenge. Handling multiple different protocol types is a challenge. Implementing shit-fast access lists, custom queues, QoS features, multicast, packet counters, sensible reactions to line down - these all add complexity. These aren't unnecessary features; we use every single one of these where I work.
Meanwhile, the amount of traffic one expects to be able to pass is increasing at a faster rate than Moore's Law. (In 1996, I had a 33Mhz PC at my desk and a 2Mbps uplink to the internet. In 2003, I have a 1500Mhz machine on my desk and a bunch of 2.5Gbps uplinks).
Thanks for being honest.:D I gotta admit, I was feeling a bit ratty last night when I posted too.
Thing is, in theory the upgrades would go fine; in practice they won't - you'll hit SOME funny that you couldn't have predicted, and the consequences are just too serious to let it go. The longer you have to plan and enact the upgrade, the smoother it'll go, and the less hurt you'll cause your customers.
I've got an 8-hour day planned tomorrow (Sunday) do upgrade our network at work, and we're looking at that much time just to do the critical boxes (20-odd) with leeway for funnies, and in an order that lets us recover if it goes to shit somewhere. We'll be doing a little parallel stuff, but not much. In theory I could upload the IOSes to the flash cards tonight, log in tomorrow from home and run./reload-all -- but I'd be a bit screwed if any one of them had a funny I didn't know about.
You're right though. Cisco CCIEs are one step from godhood. I fear them.:)
You serious? Sure, go nuts. I look forward to seeing what happens when the build you pick for a router three hops down doesn't support the STM-4 card you had in there, and stops you reaching the 20 networks behind it. Oh, and there's one over here that's running a new build with a BGP bug, so these 100 have fallen off the network. And three of these six in New York just plain didn't come back, we're not sure why yet. You've got out of band access to them all, right? right?
Upgrade with care. Even the most reliable kit develops problems a small percentage of the time; a small percentage of a lot of kit is a lot of kit.
I'm really, truly trying not to troll here, but this attitude pisses me off.
I work for an ISP. We have about 40-odd routers of various sizes. Six months ago we began upgrading their IOSes to handle IPv6. Last Wednesday we finished. We weren't pissing about; we were picking builds, checking to make sure they supported the features we needed, checking for critical known bugs, deploying them, finding bugs, sometimes scaling back. Some of these problems didn't reveal themselves for a week or two after deployment.
Pretty much none of them were due to IPv6, they're just changes in behaviour that you get when jumping from one release to another. It happens. You upgrade with care.
And because I haven't pressed the button to start and finish this process inside of two days (and instead spent the two days planning the job and trying to divine safe ACLs to apply to tide us over until we push that button) you're telling me I'm not doing my job.
A colleague keeps wondering why we use these expensive Ciscos and Junipers when Linux would technically fulfil a lot of what we want to do. He's right, but for one thing - typical server uptimes and reliability aren't good enough for the stuff that routers do. Even Linux, which is pretty damn good for uptime. A 5 minute reboot of a web server is annoying, but a 5 minute reboot of a router will get customers on the phone. An hour's outage of a web server is trouble; an hour's outage of a router is broken SLAs.
Please, don't assume that a large network is a small one scaled up. There are a million reasons why that's not the case.
I'm not sure that having party leaders nosh with arbitrary members of the public is really the best way to run a representative democracy. These guys can't be experts on every issue that passes their way; that's why they have advisers. Make contact, get yourself talking to these advisers, and make a strong case why your position is good for everyone and the alternative is bad.
Accept that there are other points of view here, and accept that anything which harms industry, or appears on the surface to harm industry, is going to need a hell of a good reason to explain to journalists when they're asking why $megacorp is pulling out of $country with the loss of $number jobs.
These guys don't read slashdot; stuff that is obviously FUD to you or I isn't necessarily going to be self-evident to them. And other people will have alternate issues which are in conflict with yours. That's life. Laurence Lessig keeps ranting about this; we make wonderful cases why Thing Is Bad on slashdot, but damned if we follow through and start explaining to our representatives. Go, gain their ear.
I saw this yesterday; I emailed two of the MEPs in my constituency whose parties I was aware were in support of software patents (I would have liked to write, but with a 30 June deadline the letter wouldn't have time to get there). I also emailed Pat Cox, President of the European Parliament and another Irish MEP. Within six minutes, I had a response from his assistant.
The gist of it is this: the European Parliament report will be taken during the session 1-4 September. The European Liberal Group (european meta-party of which Mr. Cox is a member) hasn't taken a definitive position on the report yet, but as it has been going through the committee system they have taken a very restrictive position regarding what can be patented.
I won't copy-paste the euro-speak here.:) I don't fully understand it, but have replied asking for clarification. So far, I'm pleased that he has (a) taken a position and (b) taken one that regards the issue with some care. I also thanked his assistant for correcting me about the vote on 30 June.
This means there's plenty of time. Write to your MEP explaining, politely, how you think software patents would harm our industry in your country and in Europe as a whole, and perhaps explaining the problems inappropriate use software patents have caused elsewhere. The idea of patents is to encourage innovation - explain why software patents don't do this.
No response from the other two yet; I'll be writing to them and following up by phone.
What is it with this absurd need to see every continent as a monolith? Yes, I read the RTFA. Yes, I knew long before reading the FA that US institutions were among the first adopters of IP, and so were first in the queue for addresses with the old Class A, B, C policies that were abandoned in favour of CIDR. These assignments are made, to those particular networks, and that's it. Anyone else who wants addresses has to go through the new policies for their region, and there isn't a whole pile of difference between the policies in North America, South America, Europe and Asia Pacific.
You seem to believe that "America" is holding on to this pile of addresses and can share it out amongst "itself" when the pool dries up for the rest of us. That's not how allocation works. Lots of those networks have more addresses than they need, sure; but either they hang on to them (in which case no one else gets to use them) or they return them to the pool (in which case they're fair game for the rest of us).
IP addresses aren't property. They're only useful as long as other people recognise the allocation and are willing to route it, so they're pretty useless when split up and taken out of that context. Sure, some US early adopters have a head start here - as do early adopters all over the world - but as addresses get more scarce, any redistibution will happen according to a global policy, not some sort of deal confined to the US.
I dont like IPv6 either though, too many numbers to make it managable. The new network admins are going to have to carry around a phone book just to know where all the ip addresses are in their network.
Um. Um. DNS?:)
I'm a network admin. I'm damned if I try to remember the name and location, and IP addresses of several hundred devices in my head.:)
Speaking of phones, why can't we simply augment the current IP system with an Area code feature? Seems like it'd be a lot easier than adding a billion bits to the IP address and it'd be a whole lot more managable.
Phone numbers are variable length (where I come from anyway) and changing them involves the telco upgrading their equipment, and me typing some extra digits on my phone. This isn't trivial, but can be handled by the telco itself.
IP addresses are a fixed 32 bit field, and changing this in any way involves upgrading all IP-capable equipment out there. Yours, mine, Timmy the dog's. Every last slashdot reader and embedded device. If we're going to do this, let's do it properly and make sure there are so many bits in the address field that we'll never have to do it again.
We're talking about the policies of the Regional Internet Registry in the Asia Pacific region, which are decided by the consensus of the industry and implemented by the body (based in Australia) and the National Internet Registries (based on their respective countries), using exactly the same IP space which the rest of us use.
That's an important one, people - we're all going to run out at the same time. I never got this "IP Shortage in Asia" stuff because their shortage is our shortage, whoever "we" are.
And this guy posts a one liner comparing it to the reaction of the Chinese government to SARS, hits "submit" and gets modded +5, Insightful.
I'm not in the habit of criticising moderation - it's supposed to knock out trolls, not decide whose IQ is biggest - but jeez. Am I the only one who doesn't get this?
Devices communicate peer to peer, or routed via other people's idle devices.
Nice assumption. Send code.:-)
The job of an ISP is to take your traffic and deliver it to [another network that can carry it to]* its destination. It's called routing, and it's a reasonably intensive task. We have large, expensive machines dedicated to the task of mapping IP addresses to the correct next hop, and we have highly paid individuals (cough) whose job is to lay out the network that carries all this traffic. That's what you're paying for. Everyone wants to optimise this, but it's turning out to be somewhat intractable.
It's funny, people seem to make a lot of weird assumptions about what happens to a packet as it transits the network. There's quite a lot going on there, and there are a fair number of tradeoffs at work that aren't visible to the casual observer. I suppose this is good, it means we're doing our job properly;-) and I think that if you're familiar with LAN networking, it's easy to assume that it scales up with the rest of the internet. Nothing could be further from the truth.
If you can come up with an algorithm that obsoletes routing as we know it, you've got it made. Best of luck.:)
I'm curious to know how you'd rewrite IP in order to subsume the routing problem and still have it be IP. Got any proposals we could read? The only ways I can imagine to do that are less like IP and more like ATM or stuff like that, which is interesting and useful, but probably not a viable replacement for IPv4.
Usually (in this part of the world at least), a public service means one which is necessary or useful for the community as a whole, but not popular enough to fund directly. Services like flat rate postage to rural areas, and public transport on loss-making routes, are good examples. They're not popular, but it's good for everyone if they exist, so you fund them centrally.
My thinking with a license-funded TV station (and I'm not going to defend the BBC's implementation of this) is that it's important not for the service to be popular, but for the service to exist so that when we really, really need it, it's there.
I think it's also important that there is some media stream out there which can make programming decisions based on criteria other than "how many advertisers will this get us", because what's important isn't necessarily the same as what's popular.
Ok, I hear you, but consider for a second - the alternative to a TV license is commercial funding. We have an abundance of commercial channels on digital transmission now. Is it really desirable that every media stream should be commercially funded? Is there not still a place channels whose content does not depend on attracting mainstream advertisers?
Slashdot Mirror
Actually, that's not true - it did do me good. I booked an Easyjet flight from Belfast for the next day. :-)
Unfortunately not. The flight I was on appeared not to be formally cancelled on the system, so I wasn't allowed to rebook at the time - and of course their system was jammers anyway. I tried to help out the people either side of me but was pretty unsuccessful.
1) I wasn't leaving that queue. :-)
2) The thing is - and believe me, I do shop around - none of the other options are very much better. :( The national flag carrier has remodelled itself to a low-cost airline and now matches ryanair feature for misfeature. BMI (Baby) are quite good where they fly, but to many other locations options seem to be very expensive, even accounting for ultimate cost including disasters like the above, or nonexistent.
I was due to fly the evening it all went wrong. Here's a lesson: if you're standing in a three-hour queue for the Ryanair desk, and they tell people to rebook on the web, and you take out a laptop and 3G modem, be prepared for a stampede.
The two biggest carriers in IE have, in recent months, both moved to discourage checked baggage on short-haul flights. One already charges per bag checked in, the other recently announced plans to do so. Both are modelled these days as low-cost airlines.
One of the first questions asked on the radio yesterday morning was how these policies will be affected by yesterday's events. Remains to be seen.
He means the two ends. The two starts don't count. ;-)
Right, I see where you're coming from. Couple of things I can point out. You're right that there's a difference in the amount of time a home user will saturate their connection compared with the traffic profile of a large content provider. That's why ISPs have contention ratios, and the contention ratio that a home user will find perfectly acceptable, even during the busy times, might be far from acceptable at peak times for a site like LJ. So the amount of bandwidth feeding into the user's home is just a single part of the equation, and sharing upstream bandwidth by means of contention ratios means you're already reducing the cost thanks to the fact that most users will spend most of their time idling.
(You can start to see here, btw, why home users who do hammer their connection with peer to peer traffic cause such problem - they're outliers on that model, and may be using much more of the available bandwidth than others while paying the same share.)
With a big content provider, the equation is different because their requirements are different, but at the end of the day they have a bandwidth bill too. Just like the user, it's mostly not really under their control, they take their pick from the available providers with the available service levels.
So whether it's the users or the content provider who pay >50pc of the cost of shipping the bytes isn't I think very important. We each have our bills to pay, and we each have to find our own ways to fund that; your ISP bill doesn't pay for LJ's, or anyone else who runs a web serer.
If a site tries ads, then, that's one way to pay the bills, but it'll only work of course if people can see them. You're perfectly entitled not to use the service if you don't consent to that method of funding, of course. In the case of LJ you can stick with the old free level of service where you'll rarely see any ads at all.
But is it kosher to take the level of service that they offer with ads, then remove the ads, consuming the service anyway? Not just for LJ - if a site uses ads to pay its bills, and you don't consent to the ads, is it okay remove them but continue to consume the service?
Ok, you're not far from the truth there (full disclosure: I work in the ISP industry, but I framed the question that way because I wanted to work out where you were coming from) but I still have a couple of questions.
You said a couple of posts ago that LJ only have to pay for their path to the next router along. Now I think you say - correct me if I'm wrong - that the cost of the whole path is passed on to both parties and is shared between LJ and the individuals. That's cool, I can agree with that. And you're spot on that cost per meg goes down when you buy in bulk, much like cost per unit goes down when you buy in bulk in any industry.
Thing is, when you're buying in bulk, you're still handing over more actual hard cash than the tiddlers. So LJ and the user both get a bill to access the internet, but LJ's bill, I think it's fair to assume, is a heck of a lot bigger than mine. That's a cost of doing business, just like the cost of maintaining their servers and paying their staff, and I suspect we can both agree on that.
So my questions are: do you reckon they have a right to recoup this cost? And if so, should they really get off the internet in preference to taking on ads?
It's a nice theory, but it hinges on this assumption:
This is not a geographical relationship. Two ISPs in a given country may (and are likely to) have independent connectivity. If someone in the US sends a packet bound for "Ireland", it gets sent to the upstream ISP of that user, not to to "Ireland".
So we come to the consequence of what you propose. Do you really want to regulate that every ISP in a geographic region must share upstream connectivity?
Dave
The problem is that if IP addresses are well aggregated, all a BGP-speaking router (that's the big ones in the core) needs to know is "for this very large block of addresses, use interface A; for that very large block of addresses, use interface B; for this other very large block of addresses, use interface A again." That is your routing table, it takes processor time to traverse for every packet, and it's growing; and if sizeof(routing table)*sizeof(traffic throughput) grows faster than Moore's law, it gets rather troublesome for the internet.
If you route geographically or per end-user or (shudder) per person, the number of entries that your core router has to potentially traverse explodes. This is the essence of CIDR, and we have separate naming (i.e. DNS) and routing (i.e. IP addresses) specifically so that end users may have a portable name irrespective of the routing infrastructure.
In the phone system, where naming and addressing are both conflated into your phone number, it's a lot more painful. (All of a sudden there isn't a simple programmatic way of mapping a three-digit prefix to to the operator that will handle the call.)
The problem of routing table size remains regardless of the size of the IP space - IPv6 will solve a lot of problems, but this isn't one of them.
Forwarding a packet out an interface is a fairly straightforward task. Things get interesting, though, when you start scaling up - either speed or size. Putting together hardware and software that can handle 120,000-odd routes (and growing) in the IPv4 routing table, all of which may go up and down independently and quickly, is a challenge. Handling multiple different protocol types is a challenge. Implementing shit-fast access lists, custom queues, QoS features, multicast, packet counters, sensible reactions to line down - these all add complexity. These aren't unnecessary features; we use every single one of these where I work.
:)
Meanwhile, the amount of traffic one expects to be able to pass is increasing at a faster rate than Moore's Law. (In 1996, I had a 33Mhz PC at my desk and a 2Mbps uplink to the internet. In 2003, I have a 1500Mhz machine on my desk and a bunch of 2.5Gbps uplinks).
Are you scared yet?
Dave
Thanks for being honest. :D I gotta admit, I was feeling a bit ratty last night when I posted too.
./reload-all -- but I'd be a bit screwed if any one of them had a funny I didn't know about.
:)
Thing is, in theory the upgrades would go fine; in practice they won't - you'll hit SOME funny that you couldn't have predicted, and the consequences are just too serious to let it go. The longer you have to plan and enact the upgrade, the smoother it'll go, and the less hurt you'll cause your customers.
I've got an 8-hour day planned tomorrow (Sunday) do upgrade our network at work, and we're looking at that much time just to do the critical boxes (20-odd) with leeway for funnies, and in an order that lets us recover if it goes to shit somewhere. We'll be doing a little parallel stuff, but not much. In theory I could upload the IOSes to the flash cards tonight, log in tomorrow from home and run
You're right though. Cisco CCIEs are one step from godhood. I fear them.
Dave
You serious? Sure, go nuts. I look forward to seeing what happens when the build you pick for a router three hops down doesn't support the STM-4 card you had in there, and stops you reaching the 20 networks behind it. Oh, and there's one over here that's running a new build with a BGP bug, so these 100 have fallen off the network. And three of these six in New York just plain didn't come back, we're not sure why yet. You've got out of band access to them all, right? right?
Upgrade with care. Even the most reliable kit develops problems a small percentage of the time; a small percentage of a lot of kit is a lot of kit.
Dave
I'm really, truly trying not to troll here, but this attitude pisses me off.
I work for an ISP. We have about 40-odd routers of various sizes. Six months ago we began upgrading their IOSes to handle IPv6. Last Wednesday we finished. We weren't pissing about; we were picking builds, checking to make sure they supported the features we needed, checking for critical known bugs, deploying them, finding bugs, sometimes scaling back. Some of these problems didn't reveal themselves for a week or two after deployment.
Pretty much none of them were due to IPv6, they're just changes in behaviour that you get when jumping from one release to another. It happens. You upgrade with care.
And because I haven't pressed the button to start and finish this process inside of two days (and instead spent the two days planning the job and trying to divine safe ACLs to apply to tide us over until we push that button) you're telling me I'm not doing my job.
A colleague keeps wondering why we use these expensive Ciscos and Junipers when Linux would technically fulfil a lot of what we want to do. He's right, but for one thing - typical server uptimes and reliability aren't good enough for the stuff that routers do. Even Linux, which is pretty damn good for uptime. A 5 minute reboot of a web server is annoying, but a 5 minute reboot of a router will get customers on the phone. An hour's outage of a web server is trouble; an hour's outage of a router is broken SLAs.
Please, don't assume that a large network is a small one scaled up. There are a million reasons why that's not the case.
Dave
I'm not sure that having party leaders nosh with arbitrary members of the public is really the best way to run a representative democracy. These guys can't be experts on every issue that passes their way; that's why they have advisers. Make contact, get yourself talking to these advisers, and make a strong case why your position is good for everyone and the alternative is bad.
Accept that there are other points of view here, and accept that anything which harms industry, or appears on the surface to harm industry, is going to need a hell of a good reason to explain to journalists when they're asking why $megacorp is pulling out of $country with the loss of $number jobs.
These guys don't read slashdot; stuff that is obviously FUD to you or I isn't necessarily going to be self-evident to them. And other people will have alternate issues which are in conflict with yours. That's life. Laurence Lessig keeps ranting about this; we make wonderful cases why Thing Is Bad on slashdot, but damned if we follow through and start explaining to our representatives. Go, gain their ear.
Dave
I saw this yesterday; I emailed two of the MEPs in my constituency whose parties I was aware were in support of software patents (I would have liked to write, but with a 30 June deadline the letter wouldn't have time to get there). I also emailed Pat Cox, President of the European Parliament and another Irish MEP. Within six minutes, I had a response from his assistant.
:) I don't fully understand it, but have replied asking for clarification. So far, I'm pleased that he has (a) taken a position and (b) taken one that regards the issue with some care. I also thanked his assistant for correcting me about the vote on 30 June.
The gist of it is this: the European Parliament report will be taken during the session 1-4 September. The European Liberal Group (european meta-party of which Mr. Cox is a member) hasn't taken a definitive position on the report yet, but as it has been going through the committee system they have taken a very restrictive position regarding what can be patented.
I won't copy-paste the euro-speak here.
This means there's plenty of time. Write to your MEP explaining, politely, how you think software patents would harm our industry in your country and in Europe as a whole, and perhaps explaining the problems inappropriate use software patents have caused elsewhere. The idea of patents is to encourage innovation - explain why software patents don't do this.
No response from the other two yet; I'll be writing to them and following up by phone.
Dave
What is it with this absurd need to see every continent as a monolith? Yes, I read the RTFA. Yes, I knew long before reading the FA that US institutions were among the first adopters of IP, and so were first in the queue for addresses with the old Class A, B, C policies that were abandoned in favour of CIDR. These assignments are made, to those particular networks, and that's it. Anyone else who wants addresses has to go through the new policies for their region, and there isn't a whole pile of difference between the policies in North America, South America, Europe and Asia Pacific.
You seem to believe that "America" is holding on to this pile of addresses and can share it out amongst "itself" when the pool dries up for the rest of us. That's not how allocation works. Lots of those networks have more addresses than they need, sure; but either they hang on to them (in which case no one else gets to use them) or they return them to the pool (in which case they're fair game for the rest of us).
IP addresses aren't property. They're only useful as long as other people recognise the allocation and are willing to route it, so they're pretty useless when split up and taken out of that context. Sure, some US early adopters have a head start here - as do early adopters all over the world - but as addresses get more scarce, any redistibution will happen according to a global policy, not some sort of deal confined to the US.
Dave
This is another troll, right? Oh ok, I'll bite.
Um. Um. DNS? :)
I'm a network admin. I'm damned if I try to remember the name and location, and IP addresses of several hundred devices in my head. :)
Phone numbers are variable length (where I come from anyway) and changing them involves the telco upgrading their equipment, and me typing some extra digits on my phone. This isn't trivial, but can be handled by the telco itself.
IP addresses are a fixed 32 bit field, and changing this in any way involves upgrading all IP-capable equipment out there. Yours, mine, Timmy the dog's. Every last slashdot reader and embedded device. If we're going to do this, let's do it properly and make sure there are so many bits in the address field that we'll never have to do it again.
Dave
We're talking about the policies of the Regional Internet Registry in the Asia Pacific region, which are decided by the consensus of the industry and implemented by the body (based in Australia) and the National Internet Registries (based on their respective countries), using exactly the same IP space which the rest of us use.
That's an important one, people - we're all going to run out at the same time. I never got this "IP Shortage in Asia" stuff because their shortage is our shortage, whoever "we" are.
And this guy posts a one liner comparing it to the reaction of the Chinese government to SARS, hits "submit" and gets modded +5, Insightful.
I'm not in the habit of criticising moderation - it's supposed to knock out trolls, not decide whose IQ is biggest - but jeez. Am I the only one who doesn't get this?
Nice assumption. Send code. :-)
The job of an ISP is to take your traffic and deliver it to [another network that can carry it to]* its destination. It's called routing, and it's a reasonably intensive task. We have large, expensive machines dedicated to the task of mapping IP addresses to the correct next hop, and we have highly paid individuals (cough) whose job is to lay out the network that carries all this traffic. That's what you're paying for. Everyone wants to optimise this, but it's turning out to be somewhat intractable.
It's funny, people seem to make a lot of weird assumptions about what happens to a packet as it transits the network. There's quite a lot going on there, and there are a fair number of tradeoffs at work that aren't visible to the casual observer. I suppose this is good, it means we're doing our job properly ;-) and I think that if you're familiar with LAN networking, it's easy to assume that it scales up with the rest of the internet. Nothing could be further from the truth.
If you can come up with an algorithm that obsoletes routing as we know it, you've got it made. Best of luck. :)
Dave
I'm curious to know how you'd rewrite IP in order to subsume the routing problem and still have it be IP. Got any proposals we could read? The only ways I can imagine to do that are less like IP and more like ATM or stuff like that, which is interesting and useful, but probably not a viable replacement for IPv4.
Usually (in this part of the world at least), a public service means one which is necessary or useful for the community as a whole, but not popular enough to fund directly. Services like flat rate postage to rural areas, and public transport on loss-making routes, are good examples. They're not popular, but it's good for everyone if they exist, so you fund them centrally.
My thinking with a license-funded TV station (and I'm not going to defend the BBC's implementation of this) is that it's important not for the service to be popular, but for the service to exist so that when we really, really need it, it's there.
I think it's also important that there is some media stream out there which can make programming decisions based on criteria other than "how many advertisers will this get us", because what's important isn't necessarily the same as what's popular.
Dave
Ok, I hear you, but consider for a second - the alternative to a TV license is commercial funding. We have an abundance of commercial channels on digital transmission now. Is it really desirable that every media stream should be commercially funded? Is there not still a place channels whose content does not depend on attracting mainstream advertisers?
Dave