GSoC: Global File Transfer Queue

[Tad Fisher <tmf pdx edu>]

Cross-posted from gnome-soc-list:

Hello, I'm a potential applicant looking for feedback on a project idea.

Basically, the plan is to develop a desktop-agnostic specification for
the queuing of file transfers, much as Nautilus does now on its own. The
spec will include details on specifying file sources and destinations,
both of which may be located on the local machine or a network resource;
adding/removing transfers from the queue; and running multiple transfers
concurrently.

After the spec is completed, a server will be implemented, using D-BUS
as the communication mechanism. The server should be designed around a
pluggable architecture, so that it could potentially handle the
following:

Definitely doable:
- Simple file transfers (move/copy)
- HTTP / FTP downloads (supplanting browser-specific download managers)
- rsync
- Empathy file transfers

Far-out ideas:
- Bittorrent downloads
  It would be really cool to have Bittorrent downloads behave like   
  regular file operations, e.g. drag-and-drop of a .torrent link starts 
  a BT transfer to that destination
- VCS
  Some basic operations could be supported, like pushing/pulling.

In addition, the server could benefit from heuristics to determine the
most efficient transfer behavior; for example, network transfers could
be scheduled based on the amount of bandwidth available.

Implementing a simple command-line client would be a cinch using D-BUS,
and GUI clients could be implemented using various toolkits (GTK, Qt,
Fox, straight Xlib, etc.). The idea is that all user-initiated file
transfers should be accessible and configurable through a single
interface. This would be basically a separation+extension of Nautilus'
File Operations window. Additional options, such as transfer
prioritization and scheduling, would then apply for all supported
transfer types.

So the question is, is this a good fit for Gnome/FD.o, and would this be
an ideal project for a Gnome beginner to propose? In addition, does this
project mesh well with the overall goals of the Gnome project?

Thanks for any feedback,
-------------------------
Tad Fisher
Portland State University