Path: utzoo!attcan!uunet!husc6!rutgers!mailrus!caen.engin.umich.edu!falken
From: falken@caen.engin.umich.edu (David R  Falkenburg)
Newsgroups: comp.protocols.appletalk
Subject: Re: Macintosh TCP/IP, what is there?
Summary: Who programs TCP/IP w/o sockets?
Message-ID: <3fcd4c3b.1285f@maize.engin.umich.edu>
Date: 21 Nov 88 17:33:00 GMT
References: <3fc47e41.129dc@blue.engin.umich.edu> <8811201926.AA27073@cayman.Cayman.COM>
Organization: U of M Engineering, Ann Arbor, Mich.
Lines: 41

In article <8811201926.AA27073@cayman.Cayman.COM>, brad%cayman@HARVARD.HARVARD.EDU (Brad Parker) writes:
> 
> Dave Falkenburg @ Univ of Michigan Computer Aided Engineering Network says:
> 
>    What I know about MacTCP:
>       Apple currently has no plans for berkley sockets
>       ( ^ THIS IS STUPID and we should all let Apple know)
> 
> Well, that is certainly one point of view. Another is that the berkeley socket
> model is not correct for the macintosh environment.
> 
> I have heard people express that a socket library is needed for the mac; it
> seems they want to port code which uses sockets. I submit that porting
> this code using sockets is not a good idea. The mac is not unix and blocking
> is not a good idea on the mac...
> 
> The asynchronous nature of networking would suggest that using the mac's
> async driver calls ("iopb's" as we call them) would work better. You end up
> with a model which looks much more like so called "upcalls". It would be
> interesting to hear what John Romkey's (and others) experience with upcalls
> and blocking were/are on the IBM PC (Romkey authored PC/IP which ported IP/TCP
> functionality to the IBM PC and led later to MacIP, the first IP on the Mac).
> 
> But, I am often wrong about these things ;-)
> 
> -brad

Ever wonder why not everyone understands/programs things which rely on
big parameter blocks, etc. on the Mac?   Berkely Sockets are the ACCEPTED
way of programming for networking applications with as little pain as
possible.  Ever HAVE to write a usefule ioCompletion routine or DDP
SOcket Listener?

Apple's alternative would be to provide "high-level" calls which are similar
in nature to BSD sockets, but aren't...  People want a high level interface
for networking, otherwise they won't program for it.

-dave falkenburg
-- 
Dave Falkenburg @ University of Michigan Computer Aided Engineering Network
ARPA: falken@caen.engin.umich.edu    UUCP: umix!caen.engin.umich.edu!falken