Uploading documents and browsing the ones you can reach

An ordinary member's upload page takes one or more files at once, accepts

them dropped onto the page as readily as chosen with a picker, and files

them into a compartment picked from a pulldown. The pulldown matters more

than it looks: it is built from the compartments that person may actually

write to, so the page cannot offer somewhere the upload would refuse, and

there is no free-text box left for typing an id into. Alongside it is the

list of documents that person can reach, and it says what the vault knows

about each one: its version number, the summary the vault wrote after

reading it, the metadata found inside it, when it was uploaded and how big

it is. It once said only what a document was and never what it said, and

that turned out to hide the very thing a person needs in order to trust an

upload — whether the vault actually read the document at all. What it still

never carries is a passage of the document itself. It also says whether the

file can still be taken back out exactly as it was uploaded, because a

version stored before the vault began keeping originals cannot be

downloaded and the page must say so rather than offer a button that would

fail. Anything the vault does not know is reported as unknown rather than

as a zero or a default it would be inventing.