Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!ncar!hsdndev!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.unix.programmer
Subject: Re: RPC vs Sockets vs TLI
Message-ID: <19930:Mar1909:03:5091@kramden.acf.nyu.edu>
Date: 19 Mar 91 09:03:50 GMT
References: <1991Mar19.004054.8840@bellcore.bellcore.com>
Organization: IR
Lines: 25

In article <1991Mar19.004054.8840@bellcore.bellcore.com> tim@grenada.bellcore.com writes:
> I have been told that sockets are falling out of favor in this area.

Sockets have been the de facto standard for years, and despite rumors
begun by certain vendors, sockets will remain the de facto standard for
the foreseeable future. Use them.

This has been thoroughly discussed before. If you want you can pick up
an archive of the latest comp.protocols.tcp-ip argument on the subject
via anonymous ftp to 128.122.128.22, pub/hier/inet/socketstream{,.MW,.Z}.

> Is
> RPC a sun-ism that will soon pass away?

Many vendors provide many different remote procedure call mechanisms. If
you applications are sufficiently constrained that they work under some
RPC then you might as well use it, but you shouldn't expect RPC to be as
portable or as widely applicable as sockets.

> Is TLI more "blessed" by
> standards organisations (ISO, X/Open, etc)?

No.

---Dan