I am wrestling with Excel, and it is winning. I have not yet made it to the breakthrough in R (I think I need someone to teach me R for dummies, as I am somehow shattering on the most basic things, like importing an Excel workbook with several sheets and saving that as a dataframe) but I need some quick results. Not much, mind you; not the full analysis... but at least something. Something! And I would know how to get it! But... I have made the mistake of saving the formulas in one workbook, and now I need them in another one. And due to the perverseness of my system, and that of Excel, and that of the worksheets... I get an out-of-memory error every time. No, the workarounds all do not work. Yes, I have tried them all (at least all that I could find). Arrr...


And since two people posed the same question yesterday: From what I remember of Swedish coffee (real Swedish coffee drunk in Sweden, from a cafe or a selling point for caffeinated hot drinks) they are large, they are strong (but not overly so) and they are delicious to my taste - with neither too much acidity nor too much bitterness. Not like the thing we termed "Swedish coffee" at the Forum - though of course there are different strength and quality coffee experiences to be found in any country.
German plain coffee roasts often have a tendency to be quite acidic, and the sourness is something I neither like nor can stomach well, so I am fond of the "milder" kinds of beans here, and I generally prefer the fancy coffees with lots of milk (latte macchiato, anyone?). And the coffee I tasted in the Czech Republic was not really acidic, but always quite, quite strong, plus the beans and/or roast and extraction must have been different, since it was very bitter (at least to my taste).
Your coffee-when-traveling observations are very welcome in the comments - I'd be interested to hear what experiences you have had!

EXARC is running a series of interviews with persons you might call "VIPs" regarding experimental archaeology, and the latest instance is an interview with Rosemarie Leineweber.

It's an interesting thing to read about how seasoned experimental archaeologists became involved with this area of research, and who they regard as influential for themselves. There is no separate tag or a list for all the interviews, but you can find them plus a few other articles by searching for "interview with" on the site (or clicking this link).

At the Uni of Exeter, a project to build a Bronze Age boat is under way - and they now have a time-lapse video of the first four weeks:



They're planning to publish another video in a few weeks' time - so everybody can sort of keep up with the process.

Since I had some more little troubles with xmgrace yesterday, I used the Phone-a-Friend Lifeline and did just that. And thus got myself a nice little one-on-one, flesh-and-blood Grace tutorial.

Where I learned the following Two Rules for Getting Grace:
Rule one: If plotting or importing data does not work, you probably have a dumb user problem and screwed up your dataset. Which is easy to do - just add a line break at the wrong place, or leave a string in.
Rule two: When working with grace, pretend it's a jump-and-run game or an egoshooter - save early, save often, save lots of different versions. Grace has no undo function, and hitting the wrong button (or even the right button) at the wrong time can permanently screw up things.

I'm not sure whether there's a third rule - insist on grace being totally cool and a very good tool no matter how much trouble it can make - but I will find that out (probably).

For now, though, I know how to get stuff like this:


so I'm perfectly content.

It really, really is amazing what one can learn (and has to learn) just because of some spinning. I have learned on Thursday last week that Excel will find the median of a list of numbers by... choosing the value in the middle of the list. Which means that instead of finding the statistical median (half of the values are larger than this number and half of the values are smaller than it), it finds the number in the middle of a list. Thank you, Excel, I could have done that myself. By placing a simple link to the cell in the middle of the list. Because yes, I can count! So I worked on a little more, using the average instead, and grinding my teeth (figuratively speaking only, though).


In case you do not have your own version of a spinning experiment that you need to evaluate and analyse, and have no clue what I'm talking about: the average is made by summing up all numbers in a dataset and then dividing it by the amount of numbers. The median is the value in the middle of the dataset. Why do I want the median instead of the average? Median is way less vulnerable to outliers in the data. Case in point - let's assume we have a list of values that goes 1, 2, 3, 4, 100. Average is (1+2+3+4+100) divided by 5, which is 22. In contrast to this, the median of the same dataset is 3.

And Friday morning, I had a little private time with Mr. Google again, and I found out that yes, you can force Excel to get the statistical median. You can tell it to find the x-largest number in a bunch of cells. Which means that if you know how big your list of data is (I do) and see if it's an odd or even number of data (I do) and, if it's even, can divide said number giving the length by 2 (I do) and can make Excel calculate an average (I do) you can tell it to calculate the average between the two numbers in the middle of the list. Gotcha.

Oh, and I need the median... because after some fiddling and trying out and looking at data and getting brain-dead and trying something else, the most patient husband of them all and me came up with a possibility to give relative variations of thread thickness. Which uses a simple formula built around either the median (good) or the average (not quite as good). And now please excuse me while I try to get my now nicely-sized histogram bins into a good graph form.

Which will include giving my computer a split personality - because the programme I plot to do my plotting with is only running on Linux machines.

It really is amazing how much time a single small macro can gobble up.

Especially if, when you are almost through, you decide that the process might be even better suited to the aims if done a little differently. Because it does make a difference whether you run a median filter first, or whether you filter out outliers first, or make a transformation to binary first and run filters later on.

Anyways, I now have a method to turn a scan of thread samples like this:

into this:


and then read out the thread thickness of every single one of these threads.

Now I only need to wrangle the gazillions of datapoints into something resembling histograms or some other form of legible visualisation. And be amazed again at how much time a single little spinning experiment can eat.

Experimental Archaeology is cool. About as cool as the Internet, I would say - there are so many questions about our history still unanswered, and quite a few of them are suitable for designing an archaeological experiment around them.

Plus it's absolutely exhilarating to try out things and find out things at the same time - even if the trying out and finding out involves lengthy mind-numbing and tedious work. (It's a scientific method, after all - you don't get only the fun of doing, you get the work of analysing too.)

And then, impressive things can happen. Plus a successful archaeological experiment is almost always of interest for the public - so it also helps getting people understand how varied and interesting and yet unknown the past is. Case in point? This article here, exploring old alcoholic beverages reconstructed from organic residue in vessels.

Here's to archaeology - cheers!