This conjecture is known as the Church-Turing thesis. Hofstadter's GEB has some discussion of this.

The caps in Australia aren't that onerous anymore. Add to that that ISPs have mirrors that don't count towards the quota (eg. most open-source stuff, game updates, video, Steam content servers, etc.), and it really isn't that big a deal.

The alternative would be to be limited by congestion, which at least in my experience is practically nonexistent. Similarly, there is no incentive for providers to try to reduce usage of streaming video and Bittorrent, since the heavy users are paying a premium and there is no desire for them to switch to cheaper plans.

While I've not actually implemented this, it should be fairly easy to do by building two delay-line memories, and transmitting by copying from one to the other (i.e. replacing the feedback loop in the destination memory with the output of the source register). Sounds like time for me to sit down and build some of this stuff.

ISPs normally provide a /48 or /60 or such, so you get all of your addresses with it. You can still have a firewall at the network exit; the only difference is that the destination addresses and ports don't get rewritten by NAT, but passed on verbatim.

Every host having a public address *is* a feature, since it removes the need for port forwarding.

This was because of other considerations.

The American government had asked them to start charging so that the rest of the soldiers wouldn't feel hard done by; to provide such things free to the (relatively well-paid) Americans wasn't thought to be fair on the British/Australian/etc. soldiers of lesser salary.

If this uni is anything like Adelaide, then everything necessary to follow the course will be provided. The lecturers here usually provide their own texts that cover all of the material in the course, and failing that there is usually at least a set of slides with all of the material covered. The electrical students' society sells printed copies of everything for $5--10 if you don't want to print them yourself.

I have not had a single mandatory textbook since I started, and I doubt that Melbourne Uni will be any different. The only textbook that I can think of that most people have a copy of is Sedra and Smith's Microelectronic Circuits, the rest of the common ones being those that everyone bought in first-year not knowing better.

Beyond such exceptions, the only real reason to spend the $130 is if you need a viewpoint different to that of the lecturer to figure out what's going on, and need it often enough that it's worth spending the money not to be using a library copy constantly.

Hardly; it's practically never done, so few will think to do so in an emergency. Uncontrolled acceleration is rare enough that people don't have much in the way of instinct for it beyond the usual stopping reflex.

You're sending the same amount of data to and from each phone, and to and from each base station. Moving the packetisation from the handset to the backbone doesn't change the amount of payload.

The only way that you could reduce the bandwidth requirement would be to force people listening to the same thing to use but a single channel, and my understanding is that Pandora is an on-demand service, which rules this out.

Lustre's the standard spelling outside of the US, so it wouldn't make much difference. More likely to be just the preference of the original author.

Farnell have free next-day delivery. It's a shame that none of the local Sparkfun distributors have their full range; some of their stuff looks good, but not enough to spend $20 just on shipping.

"Drink-driving" is used in a lot of places because "drunk-driving" has the implication that one must be drunk to be dangerous, while in reality the main concern is the diminished reaction time that results from a more moderate level of consumption. Not that people aren't caught driving with 0.15.

Another way to think of this---assuming we are to have exponential population growth, then population increases exp(t), while the habitable space will only increase t^3. We would hence need to reduce growth down to t^3 at most. The effect would be that even if we are to maintain the greatest possible population growth, only an increasingly small proportion of the population would be able to have more than one child per person.

Control data goes through a different channel which is normally shared, and hence has far less capacity.

Better yet, something like this.

The details of the program at Adelaide are here. While as a EEE student I cannot speak for the sciences beyond first-year physics, most lecturers already provide sufficiently-detailed notes that textbooks are not necessary. I expect that, at least to start with, it will just be a matter of loading existing PDFs/slides/whatever that were either bought from the copy-shop or handed out during lectures.