Re: File upload in HTML forms
Wed, 12 Oct 1994 22:31:26 +0100

> In article <> you write:
> >> > * On the client end, support a FORM field which can accept input of
> >> > arbitrary MIME content (not just text/plain from the keyboard). The
> >> > most natural way would be to upgrade the behavior of TEXTAREA.
> >>
> >Paul,
> >
> >I just thought of another appearent problem with the upgraded TEXTAREA. What
if the user wants to send multiple 100MB binary image files? How would the
user put 10 or 20 of them into an upgraded TEXTAREA? Does the server have to
decide for the user exactly how many files the user can send to the server?
> >
> >Ernesto
> >
> I think you may have not understood Paul's suggestion. I believe the
> idea is that when the client has a TEXTAREA window there is an option
> for the user to type in it as is done now, or pop open a dialog box
> to enter the name of a file to send and the content-type. If you
> want to upload 10 or 20 files it is presumably up to you to get them
> in an acceptable MIME multipart format.
> This seems like something that could realistically happen. It is
> important to keep in mind that unless the changes called for are
> really minimal they just aren't going to happen. At this point it
> is not clear that all (or a majority) of browser writers can be
> persuaded to make even minimal changes.
> John Franks


Thank you for the clarification. I am assuming that, when you say: "it is
presumably up to you to get them in an acceptable MIME multipart format", with
"you" you mean the browser implementor, and not the user.

Our original thoughts on file upload were along those lines. The browser lets
the user select one or more files, and when the user presses the submit button,
the files are put into a MIME multipart format and sent to the server along
with the rest of the form data. We even have a partial implementation of this.
But we decided to modify the proposal because the new scheme: 1) allows the
server to decide whether it has the resources to accept the files before they
are actually transmitted, and 2) makes backward compatibility easier (it even
allows the file transmission piece to be implemented as an external "viewer"
application; see the proposal for details).

Since I started this thread, I have gotten two inquiries from people wanting to
implement prototypes on two different browsers. At the moment, I am somewhat
optimistic about the prospects of the proposal.

I appreciate your comments.