The job of the creators of the Sept. 11 memorial at the former site of the World Trade Center in New York was enormous: to arrange 2,983 names not alphabetically or by company, nation, or hometown, but by social and personal connections. It took a huge amount of sometimes-changing input, and unlike New York City's memorial ceremony, with which it seems just about nobody is pleased, the designers were able to manage that data with algorithms that produced an elegantly organized list. A story in Scientific American on Wednesday explains how those algorithms worked, and how they tackled such a large amount of hugely significant data:

The names algorithm works in two stages. The first stage, really an algorithm unto itself, builds clusters of names from the adjacency requests. If person A needs to be near person B, and person B near person C, those three names will form a cluster of names. "That kind of results in a pile of really irregularly shaped puzzle pieces," Thorp says. Among the various indivisible bunches formed by the clustering algorithm were blocks with as many as 70-odd names.

A second, space-filling algorithm takes those puzzle pieces and fits them into place within the confines of the 76 bronze panels enclosing each memorial pool. [See below for a NOVA video about the fabrication and etching of the panels.] "In general it takes the big pieces and finds a place for them, and then fills in the gaps with the small pieces," Thorp explains. He says it took about a month to get the algorithms working, followed by months of tweaks as design requirements shifted.

When the memorial opens after the dedication on Sept. 11, it will include an online guide that will help visitors find specific names and understand the logic behind the groupings.

Want to add to this story? Let us know in comments or send an email to the author at amartin at theatlantic dot com. You can share ideas for stories on the Open Wire.
Adam Martin