On Wed, 13 Jul 1994 Jueneman(_at_)rrj0ppp(_dot_)gte(_dot_)com wrote:
I think we have a MAJOR disconnect here, for I totally disagree with your
characterization of the current users of the Internet. Maybe that is the way
people do things in Australia, but I think that it is far from the norm with
respect to the potential users of PEM. At least in the US, the major e-mail
packages tend to be commercial offerings such as cc:mail and others, primarily
operating over LANs via gateways. Using a pure SMTP package is relatively
uncommon for the vast majority of e-mail users. And almost no one uses dial up
access, except for people on the road. Even there, protocols like SLIP and PPP
are beginning to dominate.
I think you underestimate the huge number of non-business people that are
now attacking the Internet in force. These are coming in over dial-up
links, usually to a Unix prompt or some ANSI front-end. There are some
friendly graphical interfaces in this area: CIM, First Class, RIP, Oz
E-mail, and so on. But nothing approaching a standard. X doesn't count
because running it over a modem is painful in the extreme. Some companies
are now packaging SLIP software in an easy to install way, but the dial-in
account remains king, and will for some time to come. I know, because I
myself am the President of an Internet access provider here in Brisbane.
Singling out Australia as you do is a cheap shot: there are a lot of
countries, including those in Europe, that are in a worse position than Oz.
(BTW, Australia has the highest Internet usage per-capita in the world).
A business user running on a LAN is the _least_ of my worries. If PEM
requires a style of user interface for minimal support which is not
available in the currently available minimal access method to the
Internet, then it has missed the boat IMHO.
Internet standards should not be pegged to any particular user interface
or implementation style wherever possible. Some user interfaces make more
sense than others, but choosing a graphical standard and designing a system
with that, and just that, in mind is highly stupid. Some applications for
PEM (for example, automatically encrypting in the MTA) do not have a user
interface at all.
In no way would I consider myself to be an Internet guru -- I just don't have
the time or energy to deal with all of the innovations coming out. But I use a
package with a fairly decent MIME interface, I have several DUAs running using
commercial packages, I am learing to navigate using Mosaic, and I steadfastly
REFUSE to use any UNIX system that forces me to learn an arcane command
language with greps and awk and curses, etc. I can't stand to use the UNIX
mail
interface on our mail host (that's a major reason why I haven't gotten on
board
Well good for you. But for many people, a Unix account or some ANSI
front-end interface is their first introduction to the Internet. It is
possible to create a good MIME implementation for curses (I'm using one
right now called Pine). A graphical system is not required for minimal
support.
My objections were levelled at X.400 because it requires a great deal of
programming effort to get to first base. This eliminates many shareware
and freeware authors from donating their time to do something about
improving the status quo. Hence, RFC-822 is more popular than X.400
because people donated their time to add bits and pieces over time to
RFC-822 and ignored X.400.
I could understand if you said you were having problems getting X Windows to
run, or if some of the Windows or Macintosh applications didn't have quite the
finese you would like. But to be stumbling around using a 1970's command line
interface to UNIX in this day and age blows my mind.
And it blows my mind when people assume that one must have the latest and
greatest to be part of the Internet. I'll say it again: if a usuable system
can not be produced in the minimal reasonable access method currently
available, then PEM has failed. RIPEM and PGP already succeed in this area.
It may be widely used in Europe and in the academic community, but it has seen
very little use in the hundreds of thousand or even millions of commercial
e-mail users. so I wouldn't worry so much who is ahead in the very early
polling results, but rather who is leading in the major markets.
In the end, it will probably be some proprietry system invented by
Microsoft or Lotus to fit in with their proprietry mail systems rather
than PEM.
There is a time to panic, but I don't think it's here yet. What we should
panic
about, however, is the perception that is beginning to set in among the
uninvolved observers that the PEM community is badly in disarray, and the
The PEM community is in disarray because of the never ending debate over
DN's and CA's, mostly fueled by yourself. The CA structure doesn't work
because of chickens and eggs, and the whole PEM process is sitting idle
simply because people like myself cannot implement the top-down CA design
without breaking it. The CA structure is desirable certainly, but not
while it depends on the powers that be to get off their backsides before
it can be used!
This is getting way off track now. Now that the mailing list has activated
again, let's design a minimal PEM that does not rely upon huge CA structures
or particular user interface design choices, implement that widely, and then
move onto the "trendy bits". Then we might get somewhere.
Cheers,
Rhys.