I don't understand the distinction made between "documents" and "blobs" though. The entities uploaded to Google Drive can be up to 1TB (or 1TiB?) in size, so I'd think that qualifies as large. They certainly are binaries.
Then the comments may just have been reflecting inexperience with Google Drive ;-) I think the point was that there are some things users of "textual data storage systems" expect that a typical "binary data store" would not provide: indexing, searching etc. etc.
But, in any case, that functionality is typically a *superset* of what a binary data store offers. So implementing a binary data store API on top of a "textual data store" shouldn't be a problem. And, if I'm understanding you correctly, it may be more accurate to regard Drive as a binary data store in the first place.
You may want to check on the progress of any work in this area, though. I not sure if anything concrete is happening or has even been planned.
Of course, any offers to contribute are much appreciated ;-) ap
