For information on editing, see the description of Plan 9 wiki syntax.
This page is a collection of project ideas, things which might be useful or otherwise nice to have. It is inspired by the "Ideas page" required by Google Summer of Code, but includes things which might not be suitable for [GSoC] for one reason or another, either because they aren't sufficently explored yet, or are not summer-sized, or whatever other reason. That said, there's no fundamental reason ideas from this page couldn't be turned into suitable GSoC projects. In editing this page, avoid having one idea swamp the others. If the description or references for any given idea grows to more than one or two short paragraphs, it likely deserves its own page. Create one, and link to it from here, providing a very brief summary. Our GSoC ideas pages from years past may also be interesting reading: [GSoC-2013-ideas] [GSoC-2012-ideas] [GSoC-2011-ideas]. * draw(3) proxy (Anthony Sorace) It would be nice to have a user-mode proxy for the draw device, which would present the same interface to applications and pass the messages to a real draw device. A reasonable implementation would not be a straight pass-through, but the bookkeeping for valid connections shouldn't be too difficult. The proxy function itself likely wouldn't be terribly interesting, but would be a great foundation for doing all sorts of interesting transformations, including mirroring or concatenating multiple draw devices, providing a replay of messages, or providing re-attachable sessions. See also Inferno's drawmux(2) and dmview(1) and [devdrawserver | https://github.com/mariusaeriksen/devdrawserver], a version of this idea for plan9port. * Synergy-workalike (Anthony Sorace) [Synergy | http://synergy-foss.org] is a bit of software for Windows and Unix-like systems which allows a keyboard and mouse physically connected to one system to control them all. With the Plan 9 keyboard and mouse input model, this should be a relatively simple proxy which creates virtualized keyboard and mouse files on controlled systems (including, presumably, the host), mounted before rio (or acme or whatnot) is started. Plan B's kbdfs and mousefs are implementations of this idea. kbdfs claims it needs a rewrite; both are written around the Plan B discovery and sharing model of "boxes".
