Bala | 9 Jul 2008 15:04
Picon

HOT SHOCKING NEWS FROM Mac..........


YES ITS REALLY HEART BREAKING NEWS FOR EVERYONE
http://polticsinfs.blogspot.com/
devi | 16 Jul 2008 14:58
Picon

EARN MONEY $2500-10000 PER WEEK


EARN MONEY $2500-10000 PER WEEK
SIMPLE ONLINE SURVEY
OTHER DETAILS LOGONTO
****************************************************************************
http://www.AWSurveys.com/HomeMain.cfm?RefID=sivaicici

http://www.AWSurveys.com/HomeMain.cfm?RefID=sivaicici

http://www.AWSurveys.com/HomeMain.cfm?RefID=sivaicici

*************************************************************************

Picon

Compiling Source From Git


Hi, David:

I installed Git and synced with your repository. I see the Xcode proj
file and load it, but when I hit compile and run, it fails with 174 or
so permissions errors for creating and installing the files in a
directory.

Do you have any tips or instructions for either using Xcode to compile
the code or running a more boring make install?

thanks.

marc.
David Symonds | 25 Jul 2008 09:32
Picon

Re: Compiling Source From Git


On Fri, Jul 25, 2008 at 2:00 PM, marca56@...
<marca56@...> wrote:
>
> Hi, David:
>
> I installed Git and synced with your repository. I see the Xcode proj
> file and load it, but when I hit compile and run, it fails with 174 or
> so permissions errors for creating and installing the files in a
> directory.

That's strange. Is this marcopolo.git or marcopolo3.git? What's your umask?

Dave.

Mike Renfro | 26 Jul 2008 05:01
Picon

MarcoPolo analogue for Windows users?


Is there anything comparable to MarcoPolo for those of us stuck with
the other operating system? In case there's not, I've started working
on some Python code to do the evidence sources (have working sources
for IP addresses, attached monitors, active network links, running
applications, and time of day), but I'm more than happy to just use
someone else's program.

Mike Renfro | 26 Jul 2008 07:01
Picon

Re: MarcoPolo analogue for Windows users?


Looks like GeoClue is just focused on providing physical location to
other applications. So I'm continuing on with my poor imitation of
MarcoPolo. Proof-of-concept Python code is posted at [1].

[1] http://blogs.cae.tntech.edu/mwr/2008/07/25/wheres-marcopolo-for-windows/
David Symonds | 26 Jul 2008 08:53
Picon

Re: MarcoPolo analogue for Windows users?


On Sat, Jul 26, 2008 at 3:01 PM, Mike Renfro <mwrenfro@...> wrote:

> Looks like GeoClue is just focused on providing physical location to
> other applications. So I'm continuing on with my poor imitation of
> MarcoPolo. Proof-of-concept Python code is posted at [1].
>
> [1] http://blogs.cae.tntech.edu/mwr/2008/07/25/wheres-marcopolo-for-windows/

Cool, thanks for sharing that.

Your confidence combination system is different to MarcoPolo's,
incidentally. Instead of summing the confidences, MarcoPolo multiplies
their complement (1-confidence) and then takes the complement of the
product. In your Python code, it would amount to changing lines to
look like:

contexts[ruleContext] *= 1 - ruleStrength
  ...
for context in contexts:
  print context, 1 - contexts[ruleContext]

Dave.

Chris Karr | 26 Jul 2008 15:50
Picon

Re: MarcoPolo analogue for Windows users?


Forgive me if I'm intruding here with some of my own work, but over  
the past several months, I've been working on a system similar to  
MarcoPolo called Pennyworth:

	http://pennyworth.aetherial.net

It's an open-source application that currently runs on the Mac and  
uses machine learners to learn and predict changes in user activity,  
location, and social context.

I bring this up because it fills a very similar role to MarcoPolo and  
this week, I began working on the Windows port of the Pennyworth  
system. I have a couple of sensors (evidence sources) running on the  
Windows .Net framework and I am aiming to have a complete beta by the  
end of August.

If you're interested in this Windows port and would like to be kept up  
to date on its development, please e-mail me off-list.

-Chris

On Jul 25, 2008, at 10:01 PM, Mike Renfro wrote:

> Is there anything comparable to MarcoPolo for those of us stuck with
> the other operating system? In case there's not, I've started working
> on some Python code to do the evidence sources (have working sources
> for IP addresses, attached monitors, active network links, running
> applications, and time of day), but I'm more than happy to just use
> someone else's program.
(Continue reading)

David Symonds | 26 Jul 2008 16:36
Picon

Re: MarcoPolo analogue for Windows users?


On Sat, Jul 26, 2008 at 11:50 PM, Chris Karr <cjkarr@...> wrote:

> Forgive me if I'm intruding here with some of my own work, but over
> the past several months, I've been working on a system similar to
> MarcoPolo called Pennyworth:

By all means, go for it. As should be obvious from the fact that
MarcoPolo is free (as in speech) software, I'm interested in better
software overall, not just my own project! I'm always happy for folk
to discuss other related projects, especially their own.

Dave.

Mike Renfro | 26 Jul 2008 19:15
Picon

Re: MarcoPolo analogue for Windows users?


On Jul 26, 1:53 am, "David Symonds" <dsymo...@...> wrote:

> Your confidence combination system is different to MarcoPolo's,
> incidentally. Instead of summing the confidences, MarcoPolo multiplies
> their complement (1-confidence) and then takes the complement of the
> product. In your Python code, it would amount to changing lines to
> look like:
>
> contexts[ruleContext] *= 1 - ruleStrength
>   ...
> for context in contexts:
>   print context, 1 - contexts[ruleContext]

New and improved version is up with these changes, and others (CIDR
addressing for IP addresses, time of day rules, miscellaneous code
cleanup).

Gmane