Re: [NTLK] Zero Configuration Networking (Zeroconf)/Rendezvous

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Wed Jul 07 2004 - 15:23:31 PDT


Aux environs du 3/07/04 à 14:43 +0200, sous le titre "Re: [NTLK] Zero
Configuration Networking (Zeroconf)/Ren", DJ Vollkasko prit sa plus
belle plume pour écrire les mots suivants:
>Does it also hold promises for us? I think a cross-platform standard in the
>connectivity area would be very sweet... Running on all kinds of desktops
>and our green buddies would rock, wouldn't it?

Apparently, you are a little bit confused by Apple Marketese.

Zeroconf is three-layered
1) LocalLink
2) Multicast DNS
3) DNS Service discovery.

The Newton already does #2 and #3, with Dock TCP/IP from the DCL
project, that exports these features. Multicast DNS is made available
in any Newton application by latest NIE Patch (and Dock TCP/IP).

There is no promise. It's just here. Now. Actually, it has been there
for several weeks.

>Admitted, but the news

There is no news. This is pure Apple Marketese.
They just wrote a Java wrapper around their code, but this wrapper
doesn't bring *anything* at all compared to what already existed
(jRendezvous & Java mDNS).

Actually, I may be a little bit negative here. There are some
advantage of their code (which is yet to be really released and which
will only be available in Tiger in *9* months), but it's complex and
explaining this would be too long here. You can consider for now that
this Java wrapper doesn't bring anything at all.

>Well, making it public might amend some things... Hoepfully.

Apple's mDNSResponder has been open source for a long time.

Being open source doesn't mean that things improve. The DCL also has
been open source for a long time. Still, many users are writing us
asking for this or that feature, which we cannot implement because we
are lacking time.

> > >It would seem to me as a good idea to get this technology onto the Newton,
> > >too, if only via a Waba adaptation of the Java Code. What do you think?
> >
> >Apple's Java interface is just a wrapper around their native code.
>
>Hmh, wouldn't this wrappe be implementable in Waba?

No.
Apparently, you lack understanding of what a wrapper is. Please don't
think that wherever you read Java the thing can be imported on the
Newton via Waba.

mDNSResponder is Apple's C code to provide multicast dns and dns
service discovery. This code doesn't run on NewtonOS.
Their Java wrapper is a bit of Java code to talk to a process (a
daemon) based on their native C code that doesn't run on NewtonOS.

Porting mDNSResponder to NewtonOS is difficult and would be
completely stupid since we already have a Multicast DNS and a DNS
Service Discovery implementation working on our platform.

Implementing a Waba wrapper around Apple's mDNSResponder is
_completely_ useless if mDNSResponder doesn't run on the Newton. Plus
I highly doubt that the connection between their Java wrapper and
mDNSResponder daemon is possible on the Newton's architecture, it is
way too much Unix-like.

Implementing a Waba wrapper for the implementation we have on the
Newton is doable but would be striclty useless because nobody wrote a
program on the Newton to take advantage of the Zeroconf APIs exported
by Dock TCP/IP, I doubt that anyone will with Waba APIs. I also doubt
that our wrapper could be made compatible with Apple's. Waba is very
limited in respect of communications and threads (compared to Java).

>Okay, you kung fu is better than my kung fu as concerns these matters, so I
>take it this whole thing is of no value for us?

Apple's Marketese is of no value for us indeed.

Paul

-- 
Philosophie de baignoire - consultations sur rendez-vous.
NPDS/NewtonOS: http://newton.kallisys.net:8080/
Apache/FreeBSD: http://www.kallisys.com/
-- 
This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries
Official Newton FAQ: http://www.chuma.org/newton/faq/
WikiWikiNewt for all kinds of articles: http://tools.unna.org/wikiwikinewt/


This archive was generated by hypermail 2.1.5 : Fri Jul 16 2004 - 08:03:20 PDT