[modular-debian] Re: More thoughts on Plan 9 and GUIs

  • From: Marty <martyb@xxxxxxxxxxxxx>
  • To: modular-debian@xxxxxxxxxxxxx
  • Date: Fri, 28 Nov 2014 13:09:24 -0500

On 11/27/2014 11:01 PM, Miles Fidelman wrote:
David L. Craig wrote:

With the GUI out of the way, we can focus on the user-space system.
My proposals for that foundation are Plan 9 and Go (with C
for the performance parts).

ummm folks... before starting to chant "plan 9" it would behoove us to
understand what plan 9 actually is

it's primarily an extension of core concepts, specifically:
- extending the concept of all objects as files or file systems
- incorporating network communications more thoroughly
- adding the notion of private namespaces

it's a step beyond Unix, and a good one, but probably not a starting
point for "a more modular Debian ecosystem"

I see Debian's problems on two orthogonal axes: dependencies and
compatibility. Dependencies are vertical (software stack) issues and
compatibility is horizontal (interdependence).

Plan 9 in theory addresses these in two ways

1) Dependencies (archs, FHS, POSIX, scripting, shared libs)

Plan 9's bind mounts point to the true promised land and Gobolinux
demonstrates it in practice. Library dependency issues are reduced to
version compatibility, and further alleviated by JIT compilation and IPC/RPC standardization (9p).

2) compatibility (VMs, IPC/RPC, distributed computing and filesystems)

9p is used in academic and scientific environments. Inferno was AT&T's
answer to Java:

The article misses web apps, which are the last piece of the
compatibility puzzle. MS explorer and Konqueror made tentative steps to
merge web and file browsing, with browser as app interface. Enterprise
apps are moving to the browser.

It's just hand waving at this point and I'm just looking for the big
big holes in my theory.

Other related posts: