opencog-dev team mailing list archive

["Joel Pitt" <joel.pitt@xxxxxxxxx>]

Doh, yeah, I realised that it wasn't OS after I'd posted it.

It's definitely close to what I was envisioning for an AtomSpace
navigator, and it mentions having a very configurable interface.
Unfortunately it can only display a few hundred nodes currently, so
it's not currently quick enough for the expected number of atoms we'll
be dealing with regularly.

If nothing else it's good benchmark/comparison to what can be created
for OpenCog.

J

On Sat, Sep 20, 2008 at 6:00 PM, David Hart <hart@xxxxxxxxxxxx> wrote:
> Really nice, but bummer that it's not opensource! Australian universities
> breed some weird FUD about FOSS, which is odd since so many great FOSS
> projects have come from them (well, back in the days when the 'business
> (lawyers) office' didn't have the fear of god struck into everyone!)
>
> If you're really keen on using the software, I can provide Yose with some
> guidance about dealing with the university, i.e. 1. choose a license and
> make a plan for copyrights, 2. get support from your professor, 3. approach
> the biz office, 4. hope & pray...  :-)
>
> -dave
>
> On Sun, Sep 21, 2008 at 3:21 AM, Joel Pitt <joel.pitt@xxxxxxxxx> wrote:
>>
>> Hi all,
>>
>> I've found something very cool for AtomSpace visualisation
>>
>> http://cgi.cse.unsw.edu.au/~wyos/skyrails/
>>
>> Youtube example here:
>> http://www.youtube.com/watch?v=I2d312_dXEs
>>
>> Some plusses:
>> * It can run as a server, accepting live graph updates pushed to it.
>> * It has an in built console for modifying visualisation and for
>> finding nodes/links.
>>
>> Might be impressive for demos - even if it's got nothing to do with
>> intelligence ;)
>>
>> Cheers,
>> J
>>
>>
>> On Fri, Sep 19, 2008 at 4:18 PM, Linas Vepstas <linasvepstas@xxxxxxxxx>
>> wrote:
>> > 2008/9/18 Linas Vepstas <linasvepstas@xxxxxxxxx>:
>> >
>> >>>> here is Tulip's format
>> >>>>
>> >>>> http://tulip.labri.fr/tlpformat.php
>> >
>> > Two more files, these are what I should have sent. This
>> > time, there are word pairs that are closely linked *and*
>> > frequently occuring (the previous datasets had word
>> > pairs that were closely linked, but almost all were very rare.
>> >
>> > Start with the file "1e-4.tlp.gz", which has about 1K nodes,
>> > and try
>> >
>> >> Algorithm->Layout->Force Directed -> GEM(Fricke)
>> >
>> > and you'll see the relatedness working well.
>> >
>> > The 1e-5.tlp.gz file has 10K nodes. It does load. However,
>> > after picking the GEM visualization technique, its taken
>> > my computer severl hours, and its only 7% finished processing
>> > the graph.
>> >
>> > --linas
>> >
>>
>> _______________________________________________
>> Mailing list: https://launchpad.net/~opencog-dev
>> Post to     : opencog-dev@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~opencog-dev
>> More help   : https://help.launchpad.net/ListHelp
>
>