I voted for bash as the most useful to learn, but for a lot of tasks I often find I have built my own application specific language.

The advantage is that you can tailor it you specific requirements - but this must not be allowed to grow into a general purpose language. I have seen an attempt to create a C like interpreted language that was unbelievably slow. A few optimisations and it was 100 time faster, and using an actual compiler got another 100 times improvement !

is in the teaching of algebra, not the need for it.

The TFA starts by complaining about how so many students fail the subject, but by the end is describing all sorts of scenarios where maths is a requirements. It would appear that the problem, then, is to find a better way of teaching the subject.

The best results that I have seen come when students can work together to explain stuff to each other, and where the maths is related back to real world problems. A student that has just grasped some concept is better able to explain it than someone for whom that concept is second nature. Some practical use can make the subject seem more real and can aid in the understanding - I never really understood matrices until I started doing some 3D graphics stuff - the classes never related them to anything solid - they were just an abstraction.

I have been using Gnome 3 for a few months now, and with the extension tools available it can be made almost useable.

The two things that really need to change are the dynamic workspaces and the launcher.

The idea of dynamic workspaces is great for the casual user, but for the more experienced user we would like a set of named static workspaces. I have been able to kludge this by having a terminal session on each workspace - and making sure I never close them. This keeps the workspaces alive and fixed. GConf still has the ability to name workspaces so I use the workspace menu to switch when using the mouse (I am a developer, but I use the mouse a LOT).

The idea of always jumping to the running program when you hit a launcher is really annoying. I can see that it might be useful for some large programs, like Eclipse, but for terminal sessions or a calculator I want to stay on the current workspace and open a new instance. This needs to be made into a configuable option since different users will have different requirements for this.

Those two changes and it would be a viable DE.

Well politics is the big one, but there is also the finance/economic problem. It would be uneconomic to build such a plant unless we had a lot of reactors, and an export market. Our good friends, the USA, would not like their markets being taken by us - so its more politics.

You are quite correct. I based my original comment on a TV news story that I vaguely remembered: http://www.abc.net.au/news/2010-04-13/australian-laser-threatens-nuclear-security/2570568.

I do note that the process is being developed by one of the existing refiners, so it will be interesting to see if it results in significantly cheaper fuel rods.

Some estimates are that inside of 400 years, if we continue at the present rate of increase, should see us nicely toasted.
http://articles.businessinsider.com/2011-07-21/home/29979154_1_growth-trend-energy-growth-solar-panels

I don't think you understand the politics behind the Uranium cycle.

A couple of very smart Australian scientists have developed a laser based process for refining Uranium that is far more effective and cheaper than the existing processes. It has been buried. The reason given was that it is too dangerous because it would allow Iran, N.Korea, and others to build weapons, but my suspicion is that it would obsolete too much existing investment by the current refiners.

One of the issues that is often not mentioned by proponents of nuclear power is the need to refine the Uranium ore into fuel rods for the reactors and this can only be done in a very few places (at, I suspect, a significant cost). This is not an issue for USA or France or Russia, but for a country like Australia we would be putting our energy generation capability in the hands of overseas providers.

I agree with most of your points, and that for northern Europe, and much of USA nuclear might be a better answer than solar. However, for places like South Africa (the topic of the article) and Australia we should be able to get massive amounts of solar power up and running long before the first reactor even gets approval to look for a site.

Well parhaps we could start mass-producing these: http://en.wikipedia.org/wiki/Solar_Tres_Power_Tower

As a one off it is almost economical - if we make all that parts in China and set it up in outback Australia (where the land cost is minimal and there is a lot more sunshine than in Spain) we should be able to supply the entire worlds energy. (I know, transportation is an issue, but one problem at a time.)

Except of course that it takes about a decade* to get a nuclear power station up and running, but you could power from one of these http://en.wikipedia.org/wiki/Solar_Tres_Power_Tower inside of two years.

* Assuming you can get on the list to get a containment vessel - they can only make about 10 per year.

If I had to bet on the replacement for the keyboard it would be something like the Leap Motion paired with sign language - perhaps a one handed version.

and the mess when they finally do go down is going to be catastrophic for those industries.

Books like this are at least a warning to managers that putting all their eggs in a single basket might not be the best strategy.

I guess a cluster of RaspberryPis is going to be the desktop computer in the not too distant future.

I agree completely with your analysis.

The marketing peolple are looking at the rapid rise in tablet sales, rather than the actual numbers sold which is still tiny in comparison to sales of PCs.

However, while the first computer for most people is likely to be a laptop, (and later we may get little servers like the FreedomBox idea), I can see a use for several tablets if the price is low enough. For example, one to use in the kitchen for recipies, one for the lounge to surf the web, and one for the bedroom for reading, or watching a movie.

A more likely killer for the PC is the smartphone. If Google mandated a specific connector and protocol for docking then businesses (and even consumers) would invest in fitting out their offices with docking stations confident that they would be useful for a reasonable period of time. Similarly there would be a rush of docks hitting the market.