RE: revised proposal for file upload from browser to server

Larry Masinter (masinter@parc.xerox.com)
Thu, 3 Nov 1994 07:42:31 +0100

Mark Fisher asked:

>Is it worthwhile to now consider some kind of efficiency measure for file
>upload such that uploads only occur when space exists for the uploaded
>files? If so, would this measure be best accomplished by means of a
>separate HTTP method like METRICS (<2EA55482@MSMAIL.INDY.TCE.COM>), or by
>some other means?

but he probably didn't notice a recent addition:

> Browers may optionally include a
> Content-Length header both in the reply and in individual components;
> the content-length is not intended as a replacement for the boundary
> but just as a way forwarning the server of the amount of data coming.

That is, if the client knows how big the files are, it might say. If
the server cares and wants to check, it can and (if necessary) abort
the transfer before all the data has been sent.

This seemed like a simple way to allow clients to supply information
but not to require it, and allow servers to ignore it if they didn't
want to check, all without adding any additional transactions.