On 02/08/12 10:40, Rob Arthan wrote:

On 2 Aug 2012, at 06:52, Roger Bishop Jones wrote:

On Thursday 02 Aug 2012 00:33, you wrote:
On 01/08/12 14:35, Roger Bishop Jones wrote:
The real challenge (for me at least) is to get xpp
and/or emacs to run in the cloud with a display here
on earth, I don't have much clue how to do that.

I've been thinking about this.  To me, it seems
conceptually wrong to be running Xpp remotely.
Well actually, it is exactly what was expected to happen in the original
X Windows concept - run applications (called X clients, because they are
clients of the local GUI services) remotely and have the GUI appear on
local machine (running an X server, because it supplies GUI services to
the remote applications). In that model, the client connects to the
server using the standard network protocols for client-server
connection. But this model is fine on a well-managed intranet but raises
a lot of security issues in the wild world of the Internet.

We often made use of X Windows' network transparency to run Xpp remotely over a LAN, without issue. My views were less security-related and more concerned with usability given the likely longer latencies to a remote host in the cloud.

I don't know about "conceptually", but there is a question
here about what problem I am trying to solve.

Exactly.  This is the old debate about thick clients v. fat client. Both
have there merits, but it depends what your are trying to achieve.

Right, I now understand Roger's motivation is to avoid installing ProofPower on various devices, rather than taking advantage of the resources offered by the cloud.

Would it
not make more sense to run a local (earth-based) Xpp
whose journal window contains a remote ProofPower shell
(up in the cloud) via e.g. ssh?

If I were just interested in running ProofPower in the cloud
then that would be correct. And this something that one
might like to investigate.

Unfortunately with things stand, Phil's fat clients would hold the
user's source files locally while the ProofPower theory database is held
remotely. That makes sense if you just want the server to offer a
stateless service whereby you uploaded your specs and proofs each time,
It would also make sense for collaborative work, but to make that really
work, you would want to adapt ProofPower to have a model supporting
multiple concurrent users accessing a shared database.
I don't see how that would make sense.

Such a fat client can easily work with remote source files, if a user so wishes - just use sshfs to make the remote account's filesystem appear local via ssh with e.g.

  sshfs user@host:dir localdir

It will be interesting to see how running Xpp remotely works out. If I get a chance, I'll try it over a long distance SSH connection (with X11 forwarding).


Proofpower mailing list

Reply via email to