Suggestions for a Strange Project?

Kevin Riggle vj$ri^gg#guess#l55e at mt$*guess#&cn1e#t.!
Sun Jun 17 03:35:06 CEST 2001


> Is [a time server] a Python-based solution, or did you have other
> apps in mind?

If you don't mind the hassle of changing your gateway machine over to Linux
(and it can be *quite* a chore, although the security is much better than
Windows), your distribution should come with ntpd, the network time protocol
daemon, and the corresponding server application, which you can use to
update the clocks in conjunction with the US Naval Observatory's time
servers.  This is definately the way to go for a 'large cluster of Linux
machines,' as they should have everything you need on the disk if not
preinstalled.

Basic ntpd info:  http://www.eecis.udel.edu/~ntp/ntp_spool/html/ntpd.htm
USN Naval Observatory NTP Servers:  http://tycho.usno.navy.mil/ntp.html

There are some apps for Windows that use ntp.  A How-To at GeodSoft has info
on that, plus a lot on *nix systems and NTP in general.
http://geodsoft.com/howto/timesync/wininstall.htm

Python would probably work, but it's a lot easier just to install and
configure a piece of software than to "roll your own."  :)

> I have always wanted more flexible features than the
> clipboard Windows offers.

It would be best if you could set up a central 'clipboard server' and query
the machine every time you copy or paste -- much easier on Linux, as you can
directly modify the code.  For Windows, a set polling interval might work
best; some order of precedence would have to be established (ie. if machine
A has one thing in the clipboard and machine B has another, which do I send
to the central server).  Time stamps might be needed.

> Probably the biggest challenge is going to be sharing the
> graphics pad mouse-cursor messages.

Also, the Wacom tablet transmits intensity as well -- you won't just be able
to tap the Windows API.  Speed is another consideration.  I don't know if
there are USB switches out there, but it might be something to look into.

Hope that helps.
- Kevin


"Mega Hurts" <fictitious at bogus.moc> wrote in message
news:3b2b7745_2 at news3...
>
> > I believe that a 'time server' may solve the clock problem.  Security
may
> be
> > an issue -- I've never tried to set one up myself.
>
> Is this a Python-based solution, or did you have other
> apps in mind?
>
> I'm hoping to handle this with a Python applet.  My
> four personal machines would benefit from a daily auto-
> mated clock correction, but I would ventually be applying
> the solution to a large Linux cluster.
>
> Hopefully, security would be no more an issue for this
> than for any other scripted tasks either cluster would
> handle.
>
> > The others may require a
> > bit more coding.  You might be able to run a modification of the
> > Python-based chat program at
> > http://strout.net/python/server.py
> > and have each computer post the clipboard contents periodically.  > Just
a
> thought.
> >
> > - Kevin
>
> Thanks for the link!  Looks like a good starting point
> for the communications need.
>
> I have always wanted more flexible features than the
> clipboard Windows offers.  The ability to hold multiple
> items comes to mind, especially when I'm pasting
> chunks of boilerplate in standard responses to email
> inquiries.  I believe there are clipboard apps for
> Windows, but I've not taken the time to test any.
> Making a shared-clipboard app for a cluster would seem
> a natural for multiple items.
>
> Probably the biggest challenge is going to be sharing the
> graphics pad mouse-cursor messages.  Sensing and trans-
> mission of actions may be less of a challenge than
> INTERCEPTION, so a tap or movement does not effect the
> machine to which the hardware is actually connected unless
> the software is 'switched' to that machine.
>
> At best, sounds like Windows API issues.  The pad is
> a USB unit, and I'd be surprised if there was any way
> to communicate more directly with it- but maybe there
> is a hardware abstraction layer in software for these,
> and a way to tap into it.
>
> I can dream, can't I?  !-)
>





More information about the Python-list mailing list