Cosmic Conservative

Rational Conservatism from the heart of America

Harvard's astronomical problem

By CosmicConservative
Astronomical in both senses of the word.
A Trip Back in Time and Space - Harvard's Cosmos - New York Times In the summer of 1889, when this was still an analog world, a young astronomer named Solon I. Bailey carefully packed two crates of glass photographic plates taken at his outpost in the Peruvian Andes for shipment to Harvard College Observatory. Carried down the mountain on muleback and across a suspension bridge to the village of Chosica, the fragile load was put on a train bound for Lima and the long voyage to Boston Harbor.
There are roughly a half-million of these glass plates taken from Harvard's observatories in the northern and southern hemispheres, covering roughly the time between 1880 and 1970, almost a full century. There is no other such collection on earth. And currently it is all glass. They are trying to raise the money to scan them and digitize them. I get concerned every time I hear about ancient libraries being digitized because digital formats come and go like the hemline of women's skirts. Paper records and glass photo plates may be fragile and may degrade over time, but if they do survive ten thousand years, all you will need to decipher them will be eyes and maybe a magnifying glass. Will we be able to read a DVD in ten thousand years? We are reaching a point in human development where we are going to have to either settle on an unchangeable data format, or else put in our plans the constant migration of our data collections as we move from format to format. The problem is that our data collections are growing exponentially. It may soon become impossible to convert data as fast as we are migrating formats. Of course all of this will be solved by Moore's Law. We hope.

Comments (2)

mtnlurker
Are you saying that each migration will lose data? Once the plates are digitized then digital conversions are much easier.
CosmicConservative
Well, it is very difficult to do any transfer without losing data quality. You can never get better quality than the original you are copying from, but it can always be worse. Data storage formats that compress data almost always lose some data, although there are some (such as TIFF) that do not. So yes, in general you will lose quality with each conversion. The problem is that it will be a massive effort to scan and store half a million photographic plates, and by the time it is done it is almost certain that a new format will exist that needs to be converted to. But you are right that the initial scanning is by far the most difficult. And that's why I mentioned Moore's Law as helping with this.