Subj : Directly include binary data in messages
To : Tim Schattkowsky
From : Oli
Date : Sat Feb 12 2022 09:46 am
Tim wrote (2022-02-11):
TS> Hello All,
TS> w.r.t. the embedding of images I actually also consider a variant where
TS> the images are included in a more binary-style manner to conserver
TS> memory. The idea is, to introduce a Kludge for including Binary
TS> attachments directly in the mail similar to MIME but simpler to handle
TS> and more space-conserving:
TS> @BIN <filename> <CRC32 in hex> <almost binary data>
TS> Where <almost binary data> uses a simple encoding that essentially aims
TS> at avoiding $00, $0d and $0a so the resulting string still forms a valid
TS> line of 8-Bit characters. The checksum is also intended to detect any
TS> charset violence or 7-Bit fun that might have happend to the message on
TS> the way.
Hi Tim,
I was to about to reply to your first message with a very similar proposal / idea. I'm glad I didn't and you were faster. Saved me some time, redundant work and doubts to overcome ;-)
I do like the general idea of hiding binary data in kludge lines. I also would prefer ABD (Almost Binary Data) over base64 or other 7-bit encodings. (8-bit) Messages must not be changed or converted in Fidonet.
The biggest question is: would it work with current software?
Potential problems:
- kludge line too long
- tossers
- readers / editors
- message base formats (storage)
- message size too big
- maximum allowed size to small for many regular images / files
- much bigger messages
- only few programs / users can display it
- wasted bandwidth
- wasted hard disk storage
- file size limit of message bases
- gateways / NNTP server
- Internet Mail <-> FTN mail
- some software might barf on 8-bit X-FTN-BIN headers
if not properly escaped and wrapped at 998 chars
- no support for converting @BIN kludge <-> MIME attachment
- (8-bit corruption)
Then there are more general questions:
- will sysops support binaries in mail or is it annoying?
- how to display images in BBS / terminal / TUI
I haven't looked into the potential problems. Maybe we can setup an echo for testing and generate some test messages to see which software can or cannot handle it. Or create a set of VMs (Windows, Linux, DOS (?)) for a network simulation, which could also be useful for future experiments.
My gut feeling is, that it won't work satisfactory in a heterogeneous network with some old software still used on some hubs. Or there is some limitation in one of the message base formats. Or NNTP access / gateways would blow up.
Maybe we are stuck with HTTP URLs in mails and any more ambitious solution would need a new mail format.