RE: revised fileupload proposal & SAMPLE IMPLEMENTATION

Fisher Mark (FisherM@is3.indy.tce.com)
Fri, 2 Dec 94 07:16:56 EST

(Mark puts his pedant/nitpicker's hat on ...:))

In section 3, "Suggested implementation":
File inputs may optionally identify the file name using the
"content-disposition" header. This is not required, but is as a
convenience for those cases where, for example, the uploaded files
might contain references to each other, e.g., a TeX file and its
.sty auxiliary style description.

One way to divide up file uploads is by whether they are anonymous (server
script does not know/care about filename) or named (non-anonymous) (server
script makes some use of the filename transmitted). Since browsers are
unconstrained as to the order in which they upload multiple files, there
must be a way at the server end to figure out which file is which. For
files of different types, either the browser or the server script can divine
the file type by examining some portion of the initial data of the file.
For files of the same type, there is an implicit sequence of files imposed
by the serial nature of the file upload. However, this sequence, as noted
above, is not forced into any particular order (it might be file79.txt, then
file101.txt, then file2.txt, ...). As an example, it would seem to be very
hard to upload a whole web of HTML documents simultaneously if the browser
does not use Content-Disposition and does not upload the files in a
particular order.

For the case of uploads of multiple files (and since it is cheap to provide
since the browser needs to know it anyway), I recommend that
Content-Disposition be made mandatory with wording similar to this:
File inputs identify the file name using the "Content-Disposition:"
header.
This is as a convenience for the upload of multiple files where,
for example, the uploaded files might contain references to each other
(e.g., a web of interconnected HTML documents), or there exists no other
mechanism for easily distinguishing between multiple files of the same
or similar types.
("web of interconnected" is possibly redundant.)
======================================================================
Mark Fisher Thomson Consumer Electronics
fisherm@indy.tce.com Indianapolis, IN

"Just as you should not underestimate the bandwidth of a station wagon
traveling 65 mph filled with 8mm tapes, you should not overestimate
the bandwidth of FTP by mail."