From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Wed, 27 Aug 1997 08:15:37 +0100
From: Steve_Kilbane@cegelecproj.co.uk Steve_Kilbane@cegelecproj.co.uk
Subject: [9fans] Re: Anyone still running plan9?
Topicbox-Message-UUID: 5fc519aa-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <19970827071537.A0DfrSMBiLidc3nV4ozvmdtEz7ySn1yP35RkZFDQ3cM@z>

> WHy won't somebody take things like the namespacing stuff,
> the 9P and IL protocols, the fileserver/cpuserver/terminal concepts... all
> the good "internals" of Plan 9 and make an effort to bring those ideas
> into modern freeware *nix clones.

Lots of reasons come to mind...

It's harder. Writing system-level code is almost always harder than
writing an application in the warm, fuzzy environment that the OS
provides. Especially if you've got to make a radical change to the
nature of a core system service, such as the namespace.

It's not just a question of adding a "thing" in isolation. How useful
would per-process namespaces be in Plan 9, if every single application
assumed the namespace was global? Some use, sure, but not nearly as
much.

9P and IL are ok, providing there're other machines to talk them
to, and IL in itself doesn't add any value at all apart from performance.
TCP *could* have been used, with modifications to the handlers, but IL
does the job better. Few sites would be willing to put their main filestores
onto machines with a hacked-up kernel, so that they can talk 9P when
NFS is doing the job for them.

Most compelling: someone's already done the job. There's a lot of effort
involved in shoe-horning this lot into *nix. Given the cost of the real
thing, it makes far more sense to just go out and buy it.