Aaron McDaid

The use of modularity for community finding is based on what I would call a minimum likelihood approach. The idea behind modularity involves creating a model that is as wrong as possible, called a null model, then find the parameters for which the fit between the data and the model is a bad as possible.

Community finding on a network is a scientific problem. We have data in the form of nodes and links, and we want to find out where the data came from. It is not really a mathemical or computer science question, any more than physics research is. Community finding on observed networks, and similar questions, is a scientific question. And there is no good science without good statistics.

Ever wish you could do something like this in C++? Create a function that takes one parameter and returns two values?

        const char * fileBegin;
        const char * fileEnd;
        (fileBegin, fileEnd) = AFunctionThatReturnsAPairOfStrings(fileName);

It is possible, with a slight modification. Returning two items in a struct or class is, of course, not a new idea. But I want the maximum ease for extracting the results and assigning them to variables.

Update: This is exactly what Boost tie does.

        const char * fileBegin;
        const char * fileEnd;
        make_refpair(fileBegin, fileEnd) = mmapFile(fileName);

I've defined a simple class and an accompanying function make_refpair to allow this, which I describe below. It seems like an obvious solution, but I can't quickly find anybody who's done this before. E.g. they miss this on stackoverflow. Of course, there is nothing novel in returning a pair<>, but I want the convenience of assigning both variables to the contents of the returned pair. Anybody seen something similar/better?

Aaron

Thinking recently about the relative importance of various skills required by scientific researchers, I wondered if expert programming skills would become more important in future, rather than less so. The same question goes for expertise with statistics.

Within weeks, I'll be starting a PhD in UCD in Dublin. I'm delighted to be returning to university again. I'll be studying and researching for about 4 years as part of the Clique research project at UCD. I intend to blog everything about my research.

I've just quit the teacher training course I did for the last 6 months in Bristol. Being a secondary school maths teachers just isn't for me. I enjoyed it for months, but once the novelty wore off I realised that it required was too little maths and too much parenting.

I'd be happiest doing a PhD. I'm got enough qualifications and experience in maths and software development. I could return to full time employment also, but I want to investigate the PhD first.

Some of the complaints in the US and elsewhere against stimulus plans for the economy are that the expenditure today, and the debt repayments in the future, will somehow 'crowd out' wonderful efficient private activity and investment.

I mentioned previously that I am doing a one-year teaching training course. I'm loving it now. I've enjoyed it all so far, but I'm particularly happy just in the last couple of weeks as I'm settling into my second teaching practice quite well. My mentor is really fantastic, a really nice person who knows how to train a teacher. As a result of a few of the lessons this week and a couple of interesting discussions with Denise, my mentor, I think I've learnt more about teaching than I learnt in the entire course since September!

It's easy to recommend straightforward Keynesian responses for the Republic of Ireland today, such as David McWilliams. But this neglects one crucial factor, Ireland is in the euro now. I hope David will go into more detail sometime on what we should do given that we can't print punts any more.

When Ireland had its own currency, the punt, creating full employment and fixing the economy was relatively easy. The government could borrow as much money as it liked, by printing punts if necessary. Keynesianism tells us that full employment will cause the recovery, not the other way around; so any move to increase employment is a good thing. Hence deficit-funded infrastructure programs are a no-brainer. And any mild inflation that results is probably a good thing, because the alternative is much worse; deflation can be the final straw that turns a recession into a depression.

'Capital growth prospects' are very often misunderstood. It is often interpreted as 'house prices rocket forever' and drivel such as 'Microsoft will return to its traditional, very large, price earnings ratio'.

The price of an asset cannot be any more or any less that the sum total of all future income. And future incomes aren't as valuable as immediate consumption; you're not going to pay a full year's rent in 2009 to be able to use a house during the year 2020. So a house is only ever going to be worth approximately 20 years' rent.