Re: [NTLK] Zero Configuration Networking (Zeroconf)/Rendezvous for the Newton?

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Fri Jul 02 2004 - 02:20:11 PDT


Aux environs du 2/07/04 à 8:34 +0200, sous le titre "[NTLK] Zero
Configuration Networking (Zeroconf)/Rendezv", DJ Vollkasko prit sa
plus belle plume pour écrire les mots suivants:
>On above website Apple publishes the source code (also executable for
>PocketPCs); here http://developer.apple.com/macosx/rendezvous Apple has
>code and executables for Windows 2000, Windows XP, Java, Linux, Solaris,
>FreeBSD,

multicast dns was available on the Newton before PocketPC.
In fact, AFAIK, there are three implementations of mdns:
- Apple's mDNSResponder
- JmDNS (previously known as Jrendezvous), the first Java implementation
- DCL's Dock TCP/IP extension (now exports the mDNS APIs to any third
party application, and with latest NIE Patch, you can resolve .local
names transparently).

There is some special version of Courier that does some _http browsing.

>I believe this Zeroconf technology will spread thanks to this relabeling by
>Apple. Think of Quicktime - it's by Apple, and it's everywhere, and it's
>just a *player*. This is going to be much more massive and useful, esp.
>with wireless networking via Wifi and Bluetooth and all that good stuff.
>Tivo, HP, Lexmark, Epson and Brother are already announcing
>Rendezvous-compatible products, e.g. zero-configuration shared printers.

I've been working with Rendezvous for months now, for both Newton and
non-Newton projects (mainly my PhD research). Rendezvous has a lot of
advantages, but also a lot of limits.

Among these are:
- MacOS X networking stack isn't very friendly (the fact that the
interface is shut down immediatly when nothing is connected makes
Rendezvous WiFi exchanges between two macs difficult)
- it makes some bits of networking easy but too many bits are
difficult (and sometimes aggravated by Apple's own hardware/choices).
Maybe IPv6 will change this.
- the most widespread implementation is Apple's which means that
Apple can make implementation choices that become de facto standard
without being discussed or whatever. My opinion is that very often
these choices are poor because they focus on a very simple use of the
technology.
- Apple's mDNSResponder daemon/client architecture suffers from a lot
of design problems.

>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.

Additionally, Zeroconf is three-layered with at the bottom LocalLink
which isn't available on the Newton (another source of problems for
us).

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