As an aside, in case you didn’t realize it, Stanford University’s Visualization of the Republic of Letters project was based on the letters from Electronic Enlightenment (EE).
The letters in EE are extracted from modern critical editions (published by Oxford University Press and other scholarly / academic presses) and manually keyed into a database. Since EE is not based on original manuscripts, there are no digital images of the letters; only text.
Contrast the EE project to something like Eighteenth Century Collections Online (ECCO) which offers only the digital image of the original manuscript, or a very early edition if available. There are no text files for readers to rely on; no help with the awkward fonts or various typographical errors.
Which type of database is more useful in the classroom? Does it depend on the classroom? Does it matter whether it’s literature being studied or history? Should we librarians and scholars press companies to offer both images and text, and what would be the cost? If I’m correct, the projects that are based on keyed-in text (such as the Women’s Writers Project based at Brown University) generally start out as scholarly projects funded by granting agencies. Eventually if the money dries up, they usually go subscription-based, assuming there is enough demand. EE is a little bit like that, although I’m not sure whether there was ever an initial grant. It was intended as a scholarly project. ECCO, on the other hand, I think was intended to fill in the collection gaps at universities with thin historical collections. Both have their uses, but is it enough to offer only images, or only text?