Music, photography and publishing have all gone digital to varying extents, and the analog storage media they use are quickly losing market share or becoming obsolete altogether. The field of astronomy is no different, save perhaps for the fact that the digital revolution took hold early with astronomers. The 1969 invention of the charge-coupled device (CCD), which forms the basis of image sensors in everything from point-and-shoot digital cameras to the Hubble Space Telescope, changed the face of professional astronomy in the 1970s and 1980s.

CCDs quickly replaced their analog predecessor for astronomical observations: glass plates coated with photosensitive emulsion, which had been in use for more than a century. Like a record collector with a roomful of old vinyl LPs, many observatory archives now find their shelves filled with a storage medium that is physically cumbersome to use and far less popular than it once was. But unlike LPs, which have retained a steady niche among collectors and audiophiles, photographic plates do not look to be coming back into vogue anytime soon.

"Young scientists are just not interested in coming in and pulling plates," says Alison Doane, curator of astronomical photographs at the Harvard College Observatory. Doane presides over the largest plate archive in the world; its 525,000 photographic plates, give or take, roughly double the size of the next-largest collection.

These days a large portion of Doane's work week is given over to bringing the collection into the 21st century by scanning the plates to create a digital archive that will be accessible to researchers everywhere. The project, known as Digital Access to a Sky Century @ Harvard, or DASCH, is generating a record of changes in the sky over timescales of decades that has already produced a few scientific papers. But the project lacks funds to hire a dedicated staff, and its custom-built, high-speed plate scanner is operating at just a fraction of its design throughput.

Some smaller plate archives have already gone digital or are in the process of doing so, using commercial scanners or custom-made equipment. But the scale of the DASCH project required a faster plate scanner than was available. "It's a question of volume," Doane says. "If we were going to scan the whole collection on a commercial scanner, it would take decades."

The Harvard scanner, built with the help of a 2004 National Science Foundation (NSF) grant, can digitize two eight- by 10-inch plates in about 90 seconds. But just preparing the plates for scanning is a Herculean task on its own, notes Harvard University astrophysicist Jonathan Grindlay, the principal investigator for DASCH. "We're now developing a cleaning machine, because you can't clean a half a million plates by hand," he says.

Four years after scanning began, the DASCH task is only 2 percent complete. Productivity has improved somewhat over time, and in July the DASCH team digitized its 10,000th plate. But at the current rate it would take at least 20 years to digitize the remainder of the collection. "Even with this amazing machine and this fancy software, it still takes people," Grindlay says. With a staff of six to eight—mostly technicians, plus a systems programmer or two—the DASCH pipeline could handle 400 plates per day, he says, potentially scanning the entire archive in a matter of years rather than decades.

That is not to say that DASCH has gone without support. In addition to the $544,794 grant from the NSF in 2004, DASCH received a another NSF grant last year, this one for $722,275, that runs through 2013. But that is not enough to hire a full-time staff, and even at Harvard, which boasts the largest university endowment in the U.S., large chunks of institutional funding are not easy to come by. "There are lots of other needs here, even at this relatively wealthy university," Grindlay says.

So why bother with such a costly and manpower-intensive project? Aside from its unparalleled volume, the Harvard archive is unique for its complete sky coverage—the observations came from telescopes in both hemispheres—and its exceedingly long timescale. The plates record objects in the night sky from the 1880s to the 1980s, with a roughly 13-year gap in the middle, known as the Menzel gap after the observatory's then director, Donald Menzel, who suspended the observation campaign in 1953.

Massive survey projects such as the Sloan Digital Sky Survey, which is based on CCDs, harness more powerful technology and take more consistent observations than those the Harvard archive holds. After all, the photographic plates come from myriad different telescopes, many of them quite small, operating under different observing conditions and with varying qualities of photosensitive emulsion. But modern surveys cover much shorter time spans than DASCH's century. "It's a time domain about which we know relatively little, and nothing systematic over large areas," Grindlay says.

Even with the meager sample of plates already scanned, the archive is generating some science. A study in the October 2010 issue of The Astronomical Journal used a sample of 500 digitized plates, each containing the star cluster M44, to show that most plates in the archive can be calibrated to within 0.1 magnitude. (Magnitude is a measure of a celestial object's apparent brightness.) And a paper that appeared in The Astrophysical Journal Letters in February used the DASCH database to identify three stars with significant and unexplained variability in their luminosity, on the order of one full magnitude, over timescales of 10 to 100 years.

The group is now working on characterizing the corner of sky that NASA's Kepler spacecraft has been surveying since 2009 for the presence of extrasolar planets. "We'll know the time history of this piece of sky for the last 100 years, compared to the very short-time variability measured by this spacecraft," Grindlay says.

Once the scanning process is finished, he notes, the physical archive could be stored elsewhere. "We'll preserve them—they are the ultimate record," he says. "But we will almost certainly move them off-site." The glass plates together weigh 170 tons, Doane estimates, and occupy 465 square meters of Cambridge real estate.

For now, the huge footprint of the plate collection may represent its best hope for an accelerated digitization process, updating a historic library to modern relevance. "I keep trying to convince my Harvard fundraisers that there's half a building that we could liberate if we could get these things scanned," Grindlay says.