Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!samsung!think!mintaka!bloom-beacon!EXPO.LCS.MIT.EDU!rws
From: rws@EXPO.LCS.MIT.EDU (Bob Scheifler)
Newsgroups: comp.windows.x
Subject: Re: internet address as DISPLAY host field on local machine
Message-ID: <9003061412.AA03737@expire.lcs.mit.edu>
Date: 6 Mar 90 14:12:39 GMT
References: <ROSEN.90Mar5220252@cochlea.bu.edu>
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 26


    If the display is local, so no
    host field need be specified, is it less efficient if the host name or
    internet address *is* specified?

This is likely to be implementatation/release dependent.  In R4, ":0" is
supposed to mean the "best" local communication strategy.  For most
implementations right now that's chosen to be Unix Domain sockets (UDS); in
some product implementats a shared-memory connection gets chosen.  Whether
UDS is really better than "loopback TCP" depends on the OS.  It's possible
that we've made the wrong choice on some systems in R4, but not on purpose.
In R4, we shipped the Apollo config file to use UDS; since then we've released
a patch that de-configures UDS because the TCP path is faster.

    For example, is it *better* to use `:0.0' than `myws:0.0' as
    the DISPLAY varible value for the local display on myws?

In general, I would say yes.

    Another question: what is the difference between using `unix:0.0' and
    simply `:0.0'. Is the latter faster?

unix:0 has been removed from the documentation in R4, to avoid trademark
problems.  It is still supported in the R4 implementation, but should not
be depended on.  You should think of "unix:0" as meaning UDS, whereas ":0"
means best local.