Re: Large-datagram hop-by-hop delivery mechanism

[Andrew Brown <twofsonet () graffiti com>]

> I like the idea of using email. Occasionally, though, an alternate
> transport layer will be necessary, and I'd hope that the endpoints
> wouldn't have to know too much about it.
>
> Based on the email I've gotten, it looks like lots of people have tried
> to solve this. Would anybody be interested in an informal working
> group? Let me know, if so. If nothing else, I'd appreciate some review
> of my draft specification.

i dunno about a working group, but i have done something with this
before in a production environment.  i had to set up a firm-wide data
collection mechanism (for a really large international company), stuff
the data in a database, and generate reports three times a day from it
(for tokyo, london, and new york offices).  a few people whined about
email ("it's not a production system"), but it worked just fine.  and
when my system went down one night, all the inbound data was queued
correctly "outside" and flowed in nicely when i fixed the problem.
email actually works much better than people think.

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior () daemon org             * "ah!  i see you have the internet
twofsonet () graffiti com (Andrew Brown)                that goes *ping*!"
andrew () crossbar com       * "information is power -- share the wealth."