On 23-feb-04, at 16:29, Markus Stumpf wrote:
What I am really missing in this whole discussion are visions.
Don't/didn't you read/watch SciFi books/films?
Sure, but they're generally not about email. Maybe Kevin O'Donnel and
Vernor Vinge (I forget the titles, look them up if anyone's interested)
are exceptions to some degree, although it's more like Usenet there.
If you want something radical, a while ago I started thinking about a
situation where there is no general connectivity, but people roam
around with mobile clients. A way to distribute mail in such an
environment would be to exchange copies with everyone you meet and then
after some time the message should eventually be delivered.
Should a discussion not start at a more generic point? Maybe at a point
"what is email?" and "do we need something - like we now understand
email -
any more?".
The basic idea is pretty simple: transmit a message of an arbitrary
nature to zero or more people and/or automated systems. (Note the "and"
possibility.) I think the thing that makes email email is that the
sender gets to decide on when it's appropriate to send, and the
receiver gets to decide when it's appropriate to look at the message,
which can be as good as immediately or at a much later time. (Where
obviously Tsend < Trecv.)
Another interesting aspect of email is that the messages that are sent
and received are the same, so it's easy to file and subsequently
reinterpret, retransmit or retarget a message. IMNSHO this makes email
much more suitable for ecommerce and similar transactions than HTTP.
Also, the fact that everyone provides their own favorite front-end with
HTTP drastically lowers its usability in this regard. In fact, much of
what we do over the web these days could be done just as well or even
better by email or by a combination of email and web. Maybe that should
be one of our targets: create a synthesis of email and web that is
stronger than each individual protocol/concept.
Also IMHO it is very important to have an overview of how traditional
email
is used today and would be used in the future.
I'd like to disagree with Piers here, as very little of my email is
personal in nature. My job entails a lot of writing (book chapters,
magazine or website articles, drafts, router configurations,
explanations of any of these) and email is the perfect way to get this
text to the people who need to read it. Don't let the fact that this
isn't all that much in messages/day fool you: this type of email is
VERY important and would alone be reason enough to keep a connection to
the internet around (for me). Then there is the more mundane
business-related stuff such as where, what time, look at this link and
so on. And email discussions such as this very one are also an
important use of email for me.
While a few years ago text/plain was THE encoding to send emails a lot
of companies start switching to text/html as the layout provides more
possibilities to make "official email" look like a written letter, with
logos and a letter head. Maybe in three or four years they like to send
MP3s or MPEGs,
Four years?? How about TODAY. My mail program will play movies embedded
in messasges...
so IMHO discussions about ASCII/UTF-8 or whatever are pretty
useless and BINARY is the way to go ;-)
That's a transport issue. I think we all agree that base64 and similar
hacks have to go and for all fields that are meant to be read by users
and not machines Unicode is the only text encoding we want.
Email had a strong move towards a file/document transfer protocol the
last
years and the big advantage over FTP/HTTP is that it is asynchronous
(for the sender and recipient) and that it is push and not pull.
Sending a use
the information right to his box has a higher acceptance than sending
the user an URL and have him retrieve the information himself.
Actually the fact that the sender has to make this specific choice
makes no sense. The system should be smart enough to choose behavior
that is similar to one or the other automatically based on
sender/receiver capabilities and preferences.