idea topology and distancing

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

idea topology and distancing

Paul Sheldon-2
I poked around . In wikipedia article, Aristotle and categories got me
off to find wikipedia category theory was newer. Googling, I found
Poincare called father of algebraic topology for which category theory
gave a more vital meaning. I like Poincare from his "Science and
Hypothesis". I wish I could fathom the depths of my childhood
inspiration into science with mere echos of this master .

I have been trying to find nerve to check out what category theory is
but faltered in my nerve in my pokings .

I fell back :

I wanted the mutual information algorithm by Fraser I studied that had
relation to kd tree metrics to surface . I knew they were related to
what google does. Googling, I fell on something in genetics :

This abstract speaks of an article about comparing mutual information in
a pair of genes to find them expressing mostly the same thing . They
"hypothesize" the two genes to be "related biologically". I felt that a
very covered statement, almost tautologous, merely apologizing for a
vast amount of systematic work to get a pattern in a so called relevance
network. The periodic exposition of the article ended, however, with a
"zowie" : "The biological significance of each Relevance Network is
explained."

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=10902190

The use of the non-cynacure term "zowie" may inspire me to cover myself .

;-)

The early days of phylum class order family genus species weren't a
science, merely a systematic organization of living things. Somewhere in
high school someone said that, nonetheless, this was an initial first
step to a science (perhaps the introduction to chemistry with a periodic
table). This systematic form of abstraction may be my weakness, this
first step . My mother helped me long ago with systematics which
inspired an tumult of experiment/theory (symbiotic relationships
sometimes work) .

On national public radio in the states, the human genome project was
more recently stated, as a dictionary looking for an encyclopedia. The
systematic presentation of definitions in graduate level mathematics
scared me, probably because of my weakness . This weakness is Piaget
stuff; is there, instead, interfering brain patches as in the google
video "Doing with Images makes Symbols"?

Systematic definitions...

But, what is a dictionary looking for an encyclopedia?

What if genetic algorithms did the looking? Its a topic to focus and
maybe, then, a hope .

I think the science of the genome project may be mirroring what it is
studying in some way .

But, from "Dragons of Eden" by Paul Davies, I thought the mind was a
step beyond genetic encoding of information . Could it not be that it
also recapitulates some sort of pattern, this biological significance of
a relevance network ? What do this folks mean by relevance network .
GA's are proposed, at times, as competitive, I suppose with other sorts
of algorithms or old ways of thinking about thinking .

There is no flesh, perhaps, on my composition here because I fear to put
flesh on it .

I googled google and genetic algorithm and came up with flesh (it's
alive, it's alive) :
http://www.perlmonks.org/?node_id=477072

I had no idea google had an API that could be mixed up with genetics .

It is good you all are mumbling about these things too so the air won't
be thin on top of this mountain .

I try to make an inspirational summary of my mumblings :
How do you subtly connect things for meaning and scientific mindfulness
. That's, according to Godel Esher Bach, one of the metaphors fighting
for dominance at Stanford in meaningful understanding of "vision".
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Florent THIERY-2
You might also be interested in :

Using Wikipedia and the Yahoo API to give structure to flat lists
http://www.hackdiary.com/archives/000070.html

There is the dbpedia database too, which may be used to insert
keywords in a semantics context.
http://dbpedia.org/docs/
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Paul Sheldon-2
In reply to this post by Paul Sheldon-2
Thank you, Florent THIERY, for the connections which I will speed read
looking for
wisdom.

I noted, yesterday, that Mac browser, safari, surfing at nabble :
http://www.nabble.com/Croquet---User-f14183.html

exhibited thread structure but not an aspired
to holy grail genetic semantic network.

That is, if I clicked on a post with a number of replies,
the browser window would split into frames.
The top frame shows a so called thread structure of post subjects
I don't fully fathom except hierarchy has to do with date and subject.

The bottom frame, at first, is latest post in this thread.

Later, clicking on subject lines of posts in top frame
changes the bottom frame.

So, this thread structure makes croquet forum
somewhat knowledge navigable.

I put  the forum in a folder in netscape mail and can search it from another
windows for words, but don't have any so called
computer natural language understanding
afforded me in these searches.

Thanks again.
 

Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Les Howell
On Fri, 2007-05-11 at 08:21 -0800, [hidden email] wrote:

> Thank you, Florent THIERY, for the connections which I will speed read
> looking for
> wisdom.
>
> I noted, yesterday, that Mac browser, safari, surfing at nabble :
> http://www.nabble.com/Croquet---User-f14183.html
>
> exhibited thread structure but not an aspired
> to holy grail genetic semantic network.
>
> That is, if I clicked on a post with a number of replies,
> the browser window would split into frames.
> The top frame shows a so called thread structure of post subjects
> I don't fully fathom except hierarchy has to do with date and subject.
>
> The bottom frame, at first, is latest post in this thread.
>
> Later, clicking on subject lines of posts in top frame
> changes the bottom frame.
>
> So, this thread structure makes croquet forum
> somewhat knowledge navigable.
>
> I put  the forum in a folder in netscape mail and can search it from another
> windows for words, but don't have any so called
> computer natural language understanding
> afforded me in these searches.
>
> Thanks again.
>  
>
I am slowly reading my way through the internet, awaiting that message:
"You have reached the end of the internet!".  Not there yet!
    But Paul and Florent have indeed found the general references to
relevance networks.  And this information is part of the idea I have in
mind.  However, presentation is the key.  2D maps, Ishakawa diagrams,
relevance circles, and other "flat" displays simply do not convey either
the content, the linkages, nor the visual appeal to draw one from one
topic to another.  Even the process we are implementing via email, that
of threaded discourse is both enabling yet limiting.  Unfortunately I
cannot place into visual context quite the vision of mental gymnastics
that permits me to associate tools from plumbing, or woodworking, or
electronics to apply to some particular question in say mechanical
engineering.  Or for others such as Paul to allow him to form and
understand the relationship between the genetic relevance mapping to the
concept at hand, that of linking data and locations in a
multidimensional world space to form a map that one can use for
navigational purposes.  Moreover to be able to dynamically modify the
structure to permit new relevant information to enter the matrix (for
lack of a better word) and establish new links and new linkages to prior
unaccessed information.  Moreover, how to sort the paths to both
visualize the "mass path" and the "path less taken" to see where it
might lead.  

    Seldom is a new process, means or method simply lying about the
common accepted knowledge, just waiting to be picked up and examined.
Rather accessing the radical involves risk taking, exposing one to the
paths others could not, or would not travel or could simply not see the
need or desire of examining the information in that particular path.  If
no one goes there can't be useful is the traditional thought.  I liken
it to those who discard data from a histogram set.  What they get is not
a histogram, but a designed plot, and vital, pertinent information is
missing.  Without the minor linkages, one can miss some vital point that
leads to new and vital discoveries.  

    The questions before us are:
    1.  How should relevance be derived?
    2.  How should the map appear visually?
    3.  How can the navigation be executed, and the path retained?
    4.  How can new information be added to an existing map?
    5.  How can this be resolved so that others may experience the
discovery process inherent in such browsing?

    At least that is how I see it.  Relevance is sometimes calculable
via the existing knowledge on the web, similar to the map generated by
the BBC article, and iin our case, we could add the paths of folks
visiting the world spaces.  This would be useful for a relevance network
type solution.  Other means of evaluation could be time domain based.
Some knowledge gets obliterated when new information comes to light, and
falls into obscurity.  It is no less useful (think of the dead sea
scrolls for one example), other knowledge is proven false, but that
proof and the knowledge point are still relevant to prevent future
browsers from falling into the same trap, or perhaps permitting them to
re-examine the knowledge to add new and pertinent information that will
re-open the investigation, and perhaps open a new and more pertinent
investigation.

    Moreover there could be an additional possibility, that of
establishing the processes of creative thought, perhaps yielding new
insight to education processes to permit more people to discover their
own creativity.  

    I am sorry that I am unable to add more structure to the debate, and
feel that my thoughts here, while mine and mine alone are sometimes
hollow echoes of other work or even possibly faded memories of prior art
penetrating my consciousness.

Regards,
Les H

Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Hans N Beck-2
Hi,


>
>     Moreover there could be an additional possibility, that of
> establishing the processes of creative thought, perhaps yielding new
> insight to education processes to permit more people to discover their
> own creativity.
>
>     I am sorry that I am unable to add more structure to the  
> debate, and
> feel that my thoughts here, while mine and mine alone are sometimes
> hollow echoes of other work or even possibly faded memories of  
> prior art
> penetrating my consciousness.
>


I hope to find time to make experiments like taking the 3 axis of  
euclidean space and give them some meaning and just try what happens.  
Croquet should be the ideal tool for this. I always must think of  
Disneys move "Tron" which its interpretation of space in the universe  
of programs :-)

Regards

Hans

Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Les Howell
On Fri, 2007-05-11 at 21:52 +0200, Hans N Beck wrote:

> Hi,
>
>
> >
> >     Moreover there could be an additional possibility, that of
> > establishing the processes of creative thought, perhaps yielding new
> > insight to education processes to permit more people to discover their
> > own creativity.
> >
> >     I am sorry that I am unable to add more structure to the  
> > debate, and
> > feel that my thoughts here, while mine and mine alone are sometimes
> > hollow echoes of other work or even possibly faded memories of  
> > prior art
> > penetrating my consciousness.
> >
>
>
> I hope to find time to make experiments like taking the 3 axis of  
> euclidean space and give them some meaning and just try what happens.  
> Croquet should be the ideal tool for this. I always must think of  
> Disneys move "Tron" which its interpretation of space in the universe  
> of programs :-)
>
> Regards
>
> Hans
>
I think that is a pretty good idea.  I was thinking of some kind of
"hall of doors" paradigm, but that would be very bulky and not provide
much improvement over what is available currently.

Perhaps TRON had a good visualization because it was done by people who
specialize in visualization.  I keep thinking that maybe some commercial
artist(s) might be able to add dimension to this debate.  It would be
good to get a perspective from someone used to designing interfaces as
well.  I think of the book "The Art of Human Computer Interface Design",
especially the one about the gorilla who got upset and threw the works
against the wall.  Maybe something should stick from that.  I seem
continually constricted by the interface, while my mental images evolve
in some form of n-dimensional state, I am unable to place them into the
system without considerable effort.  A continual struggle of expression.
I cannot even begin to imagine the effort required by non-english
speakers in this computing environment.  Worse we don't even have tools
to help translate.  

    One profound experience I had was designing a database query system
for non-english speakers.  My parsing algorithms would fail miserably
when faced with non-standard context or structure.  Thus an Asian who
had the least ability in English got the most amount of material
returned from the searches, which made the search engine more of a
burden than an assistance.

    The search for a means of browsing rather than searching (is that as
bad as I think?) continues.  I did read a bit on Neural net
relationships and "cluster" theory, and while cluster theory seems
promising, the literature seems to indicate it is a bit beyond mere
mortals to figure out.

    Are any of the really gifted interface types following this
discourse?  Do you have any suggestions?

Regards,
Les H

Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

tmeller
Hi:

I'm doing my masters thesis on a topic that touches on classification and
clustering. Neural networks are to abstract for the user to understand. Could
you create a 3-D dimensional space for Ontologies?

When I'm done this December, I'd like to work on that. I'll do the business
intelligence stuff. I don't know much about 3D stuff. Mere mortals can't
figure that out.

For clustering technology check out Affinity Propogation from the University
of Toronto.

Cheers,

Thomas

Quoting Les <[hidden email]>:

> On Fri, 2007-05-11 at 21:52 +0200, Hans N Beck wrote:
> > Hi,
> >
> >
> > >
> > >     Moreover there could be an additional possibility, that of
> > > establishing the processes of creative thought, perhaps yielding new
> > > insight to education processes to permit more people to discover their
> > > own creativity.
> > >
> > >     I am sorry that I am unable to add more structure to the  
> > > debate, and
> > > feel that my thoughts here, while mine and mine alone are sometimes
> > > hollow echoes of other work or even possibly faded memories of  
> > > prior art
> > > penetrating my consciousness.
> > >
> >
> >
> > I hope to find time to make experiments like taking the 3 axis of  
> > euclidean space and give them some meaning and just try what happens.  
> > Croquet should be the ideal tool for this. I always must think of  
> > Disneys move "Tron" which its interpretation of space in the universe  
> > of programs :-)
> >
> > Regards
> >
> > Hans
> >
> I think that is a pretty good idea.  I was thinking of some kind of
> "hall of doors" paradigm, but that would be very bulky and not provide
> much improvement over what is available currently.
>
> Perhaps TRON had a good visualization because it was done by people who
> specialize in visualization.  I keep thinking that maybe some commercial
> artist(s) might be able to add dimension to this debate.  It would be
> good to get a perspective from someone used to designing interfaces as
> well.  I think of the book "The Art of Human Computer Interface Design",
> especially the one about the gorilla who got upset and threw the works
> against the wall.  Maybe something should stick from that.  I seem
> continually constricted by the interface, while my mental images evolve
> in some form of n-dimensional state, I am unable to place them into the
> system without considerable effort.  A continual struggle of expression.
> I cannot even begin to imagine the effort required by non-english
> speakers in this computing environment.  Worse we don't even have tools
> to help translate.  
>
>     One profound experience I had was designing a database query system
> for non-english speakers.  My parsing algorithms would fail miserably
> when faced with non-standard context or structure.  Thus an Asian who
> had the least ability in English got the most amount of material
> returned from the searches, which made the search engine more of a
> burden than an assistance.
>
>     The search for a means of browsing rather than searching (is that as
> bad as I think?) continues.  I did read a bit on Neural net
> relationships and "cluster" theory, and while cluster theory seems
> promising, the literature seems to indicate it is a bit beyond mere
> mortals to figure out.
>
>     Are any of the really gifted interface types following this
> discourse?  Do you have any suggestions?
>
> Regards,
> Les H
>
>




Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Florent THIERY-2
Did you read about Small semantic world and psearch? These are quite
up-to-date research efforts in the semantic clustering area...

"Peers in SSW are clustered according to the semantics of their local
data and self-organized as a small world overlay network."
http://www.cse.psu.edu/~wlee/Publications/wlee%20ICNP04.pdf

And pSearch
http://www.cs.rochester.edu/u/sarrmor/publications/pSearch-SIGCOMM03.pdf

---> TOREAD list :p
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Peter Quirk-2
In reply to this post by Les Howell
The discussion on this topic and the other one on semantic distance suggests
that we need an abstract space class which includes one or more methods for
measuring distance between objects, and another method to return the available
distance methods. The default space might be Euclidean, and include a default
method to measure the distance between two points. However, I could imagine one
implementing other measures of distance such as:
* How far away (in time, or distance) is that point if I walk, avoiding
obstacles? In a gridded space it might measure the Manhattan distance.
* How far away(in time, or distance) is that point if I am a winged creature
that fly over obstacles?

I want to hook Business Process Management tools into Croquet to create
simulation tools. I can imagine creating a space that represents a swim-lane
diagram. My distance functions might include:
* How long till I reach a certain step in the diagram?
* How many signatures (approvals) lie between here and the end-state?

To interact with external BPM tools, I need to communicate with them via web
services. I asked on the croquet-developer list a while back about how to
interact with web services when core distribution doesn't seem to include
support for SOAP, and the SoapClient package won't load. If anyone has an
answer or example, I would be most grateful. Should we expect standard support
for SOAP, RSS, XSD, RDF, UDDI and other Web 2.0 technologies in future releases
of Croquet?

-- Peter
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Florent THIERY-2
> To interact with external BPM tools, I need to communicate with them via web
> services. I asked on the croquet-developer list a while back about how to
> interact with web services when core distribution doesn't seem to include
> support for SOAP, and the SoapClient package won't load. If anyone has an
> answer or example, I would be most grateful. Should we expect standard support
> for SOAP, RSS, XSD, RDF, UDDI and other Web 2.0 technologies in future releases
> of Croquet?

This is a very interesting question indeed. Are there RPC types in
croquet that may allow inter-program communication, with local or web
applications? Can one render html within croquet directly?

I recently read about ogoglio, which offers this very interesting
feature already (such as displaying textures made of photos from a
flickr account); is there some inherent limitation to take opencroquet
as base for this type of "mashup", and would have led him to do his
own platform?

Thanks

Florent
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Peter Quirk-2
The Squeak community has been active in developing SOAP and RSS support, but
I have not been able to load the SOAPcore class
(http://map.squeak.org/package/dab9b621-00d2-41c3-966c-458bf62b8008) into
Croquet due to dependencies on earlier versions of Squeak.

Rendering of HTML (as found in the wild) is weak. The scamper browser has
not kept up with the evolving web and cannot handle JavaScript and
AJAX-enabled sites. Most users on this list will suggest using a VNC
connection to a real browser. VNC is banned in my organization, particularly
across the firewall. If you use VNC, then you are clearly interacting with
the presentation of the data, not the data itself.

-- Peter

----- Original Message -----
From: "Florent THIERY" <[hidden email]>
To: <[hidden email]>; "Peter Quirk" <[hidden email]>
Sent: Sunday, May 13, 2007 6:36 PM
Subject: Re: [croquet-user] idea topology and distancing

>> To interact with external BPM tools, I need to communicate with them via
>> web
>> services. I asked on the croquet-developer list a while back about how to
>> interact with web services when core distribution doesn't seem to include
>> support for SOAP, and the SoapClient package won't load. If anyone has an
>> answer or example, I would be most grateful. Should we expect standard
>> support
>> for SOAP, RSS, XSD, RDF, UDDI and other Web 2.0 technologies in future
>> releases
>> of Croquet?
>
> This is a very interesting question indeed. Are there RPC types in
> croquet that may allow inter-program communication, with local or web
> applications? Can one render html within croquet directly?
>
> I recently read about ogoglio, which offers this very interesting
> feature already (such as displaying textures made of photos from a
> flickr account); is there some inherent limitation to take opencroquet
> as base for this type of "mashup", and would have led him to do his
> own platform?
>
> Thanks
>
> Florent
>
Reply | Threaded
Open this post in threaded view
|

Re: idea topology and distancing

Hans N Beck-2
Hi,

now, there were many interesting thoughts in this thread and related,  
but what does this all mean for Croquet ? Have we a consensus on what  
Croquet should provide or support ? Can someone do a summary :-) ?

  I have the feeling that until today  Croquet misses a founded  
concept on

- searching/browsing
- navigation
- distance

this all related to information. Maybe I'm wrong, but as far it  
belongs to me, Croquet is worth to be more than a replication of our  
physical environment.  There is potential for more :-)

Regards

Hans


Am 14.05.2007 um 19:10 schrieb <[hidden email]>:

> The Squeak community has been active in developing SOAP and RSS  
> support, but I have not been able to load the SOAPcore class  
> (http://map.squeak.org/package/ 
> dab9b621-00d2-41c3-966c-458bf62b8008) into Croquet due to  
> dependencies on earlier versions of Squeak.
>
> Rendering of HTML (as found in the wild) is weak. The scamper  
> browser has not kept up with the evolving web and cannot handle  
> JavaScript and AJAX-enabled sites. Most users on this list will  
> suggest using a VNC connection to a real browser. VNC is banned in  
> my organization, particularly across the firewall. If you use VNC,  
> then you are clearly interacting with the presentation of the data,  
> not the data itself.
>
> -- Peter
>
> ----- Original Message -----
> From: "Florent THIERY" <[hidden email]>
> To: <[hidden email]>; "Peter Quirk" <[hidden email]>
> Sent: Sunday, May 13, 2007 6:36 PM
> Subject: Re: [croquet-user] idea topology and distancing
>
>>> To interact with external BPM tools, I need to communicate with  
>>> them via web
>>> services. I asked on the croquet-developer list a while back  
>>> about how to
>>> interact with web services when core distribution doesn't seem to  
>>> include
>>> support for SOAP, and the SoapClient package won't load. If  
>>> anyone has an
>>> answer or example, I would be most grateful. Should we expect  
>>> standard support
>>> for SOAP, RSS, XSD, RDF, UDDI and other Web 2.0 technologies in  
>>> future releases
>>> of Croquet?
>>
>> This is a very interesting question indeed. Are there RPC types in
>> croquet that may allow inter-program communication, with local or web
>> applications? Can one render html within croquet directly?
>>
>> I recently read about ogoglio, which offers this very interesting
>> feature already (such as displaying textures made of photos from a
>> flickr account); is there some inherent limitation to take  
>> opencroquet
>> as base for this type of "mashup", and would have led him to do his
>> own platform?
>>
>> Thanks
>>
>> Florent