Baby, it's cold outside

Just after having decided that I would not try to adjust further to Canadian time, this morning for the first time I woke up after 6:00; in fact I was awakened by my neighbour using our shared bathroom, at 8:00. (A first door leads from the corridor to a front porch with further doors for one bathroom and two bedrooms.) Of course that meant I couldn't use it, which was annoying. I have no complaints about our accommodation otherwise, but this particular aspect is a minus.

I thought we had a cold day yesterday, but according to the official sources, when I made my way to the breaksfast hall this morning it was an astounding -36 degrees Celcius. That is by a large margin the coldest I have ever experienced. I can't even say I really experienced it, because I do not think our skin is equipped with the right receptors to distinguish between, say, -20 and -30 or -36; between +15 and +20 we can feel very fine differences, but below freezing point it is just cold, or very cold, or maybe even extremely cold, none of which are very accurate terms. I do remember a Lithuanian postdoc telling me that at -40 it starts to be dangerous exposing bare skin to the air for more than a minute or so; it will blister very quickly.  I never thought to be in a position to verify that statement, and in fact I didn't try to verify it now. All I can say is that my lips do not like the combined exposure of the last few days.

We had an interesting day today. In the morning I gave my position statement, which was well received and generated some interesting discussion. Essentially the message was that in transforming between models you should not only get your results right on the syntactical level, but also on the level of semantics, or meaning. The database example I used in my yesterday post to illustrate the principle of model transformation is not really appropriate to explain what I mean by meaning: for that you have to imagine different types of models that do not only describe data but also actions. Think for instance of the old tax forms, before it started to be routine to submit them electronically: they typically had flow charts showing under what circumstances you had to fill in which questions. Such a flow chart is also a model, and you might want to transform it, for instance to a program that tests those conditions automatically and selects the appropriate questions for you. But then you want to make sure that the meaning of the original flow chart is preserved in the program: you do not want to miss tax deduction opportunities! Thus, semantic-preserving model transformations; an interesting complication to a problem that is already far from trivial.

Remains of the day

In the afternoon we continued, in the same smaller group, the discussion and taxonomy of benchmarks started on Tuesday, this time with the aim of at least filling the space of BX benchmarks with at least one concrete instance. Hammering this out revealed further misunderstandings, but resolving those is also a good way to learn other people's perspectives. We agreed to continue this evening, in fact a few minutes from now.

Later that same evening

A fruitful continuation of our discussion later, we even have a concrete plan to submit a paper to the BX workshop, for which the deadline is next week. That's very tight, but we might just be able to work it. Almost all of my colleagues are very deadline-driven, so aiming to write a paper is about the only way to make progress anyway; if in the end we don't make it, still we may have gotten far enough to continue and try to get it published elsewhere. A tangible outcome of this workshop!

Tomorrow I will be travelling back; do not expect any post. In fact, I will have to wait and see how fit I feel on Saturday. It's another Beyond Monopoly! day, if at all possible I plan to go. This blog comes second to that.