The National Endowment for the Humanities has decided that digital technology is here to stay, apparently. The agency just announced that what used to be their Digital Humanities Initiative is now the Office of Digital Humanities.
“Time to get those new business cards printed, I suppose!” blogged Brett Bobley on the official NEH Web site.
In addition to the presumed new cards and new building signs (changing DHI to ODH), one real but small change will be transforming the DHI—er, ODH e-mail newsletter to a Web-based update with an RSS feed.
And it’s true that an office does have a more permanent feel than an initiative. The mission, though, remains the same: To help scholars figure out new ways to analyze, preserve and teach materials using digital formats. —Josh Fischman
3 Responses to Digital Humanities Gets Federal ‘Office’ Space
Leigh Caldwell - October 25, 2011 at 9:29 am
Fun analysis. A couple of things that may have helped the lawyers…
1. Adding up sums is a much simpler computer operation than calculating the digits of pi. 20 billion sums per hour is 300 million per minute or 5 million per second – easily within the capabilities of 15-year-old computers. With modern computers (or even using Amazon’s cloud capabilities? Not sure if you can sign up for that after business hours in NYC!) it would be easy to go well beyond 1000 times that speed. This would probably bring the problem within the realms of practicability.
2. More importantly, $150 million is a lot of money to transfer, even in seven transactions – I suspect that, if you can eliminate transfers to/from known sources such as the trading desks of European banks, there would be very few transactions left over $20 million (the average transaction must be at least this size, meaning one or probably more of them would be higher). Lichtenstein’s GDP is about $5 billion (just $20 million per working day) and bank assets, although not easy to track down, can reasonably be estimated to be somewhere between $100 and $500 billion held by 100,000 to 500,000 overseas investors (cf Switzerland which has $6 trillion of bank assets and around 100 times Liechtenstein’s population). Bank transfers and account balances are usually observed to obey a power law distribution, so if we estimate that the average bank account has $1 million and (generously) half of that balance is transferred in or out every year, the number of transactions over $20 million would probably be less than one in 2000.
Thus, in your estimated dataset of 1000 transactions in the five relevant banks, the $20 million+ transactions would stand out like a sore thumb. The only way in which the scoundrel could have hidden his tracks effectively would be to make one or two huge transfers ($50 million or $100 million+) – a giveaway in itself – and then make a few smaller transactions for the rest of the money.
Once these large transactions are identified, the problem simplifies down to a much easier combinatorical challenge, and one would imagine that an hour of computer time would be enough to finish it. Which is lucky, because they’d probably need quite a few hours to write the program, even if one of them is a decent software developer.
Edward Aboufadel - October 25, 2011 at 9:51 am
Those are great comments, Leigh. Sums would be quicker to calculate (and is, in fact, the basis of something I’ve been working the past few weeks), and the point about the power law distribution does suggest the seven deposits would stick out in the list. Seems to me the scoundrel would know better and would have broken the money up into more like seventy deposits.
johnbarnes - November 8, 2011 at 12:04 pm
I can see several ways that subtotals might be created and sorted to drastically cut the number of sums that actually needed to be computed, and once subtotals are sorted, a few iterated comparisons ought to suffice to find the right result. I’m not sure I see why anyone would actually compute all the permutations without subtotaling and sorting; seems grotesquely inefficient.