>> um...okay.  i don't like nscd that much, but what do you have against
>> doors?  or do you simply not like things you know nothing about?
>
>I know a fair bit about Solaris "doors" -- at least as much as can be
>learned from reading the (now available) manual pages and from talking
>to friends who've done internals programming on them.

then we're about even, except that i intend to reserve my opinion
about it as an ipc mechanism until after i've tried to *use* it.

>The primary thing I don't like about them is that even their authors
>were unsure for the longest time whether they were an answer to a
>problem, or just a potential answer looking for a problem.  From what
>I've learned since I think there are much more elegant ways to do what
>they do.  Of course I really don't like Sun's RPC (in any incarnation)
>either.  Why can't people just do the easy thing and write simple IPC
>protocols (eg. using ASCII commands and data)!?!?!?

it's called innovation: trying to come up with possible solutions to
problems that you have.  doors were invented to speed up rpc
processing.  i don't like rpc much either; i tried to program using it
once and quickly became lost.  it seems like a much too heavyweight
abstraction for what it's trying to do, i'll try it again some day.

doors in and of themselves can't be bad simply because of that.  the
traditional rpc mechanism was udp, with the "redundancy" needed
supplied by that the rpc mechanism.  would you call udp bad simply
because it was used to make something you don't like?

-- 
|-----< "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."