Stored files
As a paid-for extra cost option, what about saving the full contents, original full name and path, hash, and time stamp of each filename on the server after upload? Then files already uploaded need not be uploaded again. And pre-computing a hash on a newly seen client-end file could be used to also avoid uploading duplicates. Also, the results of all comparisons could be cached too, for instant results - caching a bool equal/unequal plus the diff list. Could also perform inferences too concerning predicted result of a comparison of a new file with a cached one. I'm good at proposing work for other people when I'm not doing it myself. Would def pay extra for upload avoidance and instant comparisons tho.
Maybe some day. So far we want to stay away from storing any user content.
-
Cecil Ward commented
Understandable. Would make it lightning fast to use tho, especially good for users on expensive or slow connections. Would increase your costs and risks, so would definitely expect to pay for such an advanced service. Many thanks for the good work!