It seems evident that with the move towards a linked data internet that CMap should follow in the same spirit, as it is well placed to take advantage of its 'linked data' visual format. For those who don't know what linked data is - see Tim Berners-Lee (father of the internet) work and that of his partners dpmedia...
In the context of CMap, when a person types a search in the search engine, the results are given back as a CMap of linked data, with the search word in the middle, and the most appropriate near the centre and then links less appropriate moving out from that. This means it is possible that CMap builds the CMaps and the links (based on tags, and or links made by users); therefore making it possible to have an auto-generated CMap.
Then, maybe it would be possible to go even further - once the user has selected the 'auto CMap Search map', it can be manipulated with other links and user made links, and then saved as a new cmap.
Any thoughts?
Linked Data
-
acanas
- Posts: 753
- Joined: Tue Mar 17, 2009 5:52 pm
Re: Linked Data
How do you suggest that the linking phrases be generated? CmapTools was designed to construct concept maps, and concept maps have linking phrase. You suggest tags, or made by the user, would this mean that the linking phrases be added to the result of the search by the user manually?
Have you tried the "Concept Suggester" feature of CmapTools? Given a (preferrably small) Cmap it suggests related concepts --- it you drag these concepts to the Cmap, are you getting what you have suggested? (of course, the algorithm for obtaining the concepts would have to be improved. The work you are suggesting is probably working out of some ontology).
(If you haven't tried the concept suggester, on an open and edited Cmap -- that is, once you have done some editing --, click on the icon on the right of the Cmap window that looks like a chat blurb with a Cmap inside. Click on the 'green light' of the semaphore. Suggested concept should show after a while. Or click on the circle-arrows on the top right besides the semaphore and you will get suggestions. Our experiments show that just a couple of suggested concepts that trigger new ideas can help a lot if making the Cmap. As this shows, the tough part is getting enough relevant concepts.)
Regards,
Alberto
Have you tried the "Concept Suggester" feature of CmapTools? Given a (preferrably small) Cmap it suggests related concepts --- it you drag these concepts to the Cmap, are you getting what you have suggested? (of course, the algorithm for obtaining the concepts would have to be improved. The work you are suggesting is probably working out of some ontology).
(If you haven't tried the concept suggester, on an open and edited Cmap -- that is, once you have done some editing --, click on the icon on the right of the Cmap window that looks like a chat blurb with a Cmap inside. Click on the 'green light' of the semaphore. Suggested concept should show after a while. Or click on the circle-arrows on the top right besides the semaphore and you will get suggestions. Our experiments show that just a couple of suggested concepts that trigger new ideas can help a lot if making the Cmap. As this shows, the tough part is getting enough relevant concepts.)
Regards,
Alberto
-
fridemar
- Posts: 56
- Joined: Wed Apr 29, 2009 12:11 pm
- Location: Germany
- Contact:
(Partial) Semantic Triples for a CmapSearcher as Queries
A CmapSearcher in its most elementary would accept the following queries:
1. single concepts C ( for which there may (or may not) exist completing linking labels or objects D with (C,l,D) or (D,l,C) )
2. pairs of concepts (C,D) ( for which there may (or may not) exist a linking label l with (C,l,D) )
3. pairs of concepts and linking labels (C,l) ( for which there may (or may not) exist a concept D with (C,l,D) )
4. pairs of linking labels and concepts (l,D) ( for which there may (or may not) exist a concept C with (C,l,D) )
5. linking phrases l ( for which there may (or may not) exist concepts C,D with (C,l,D) )
6. propositions (C,l,D) as complete Semantic Triples
As result should come a (possibly empty) list of concept maps of all public available cmaps (of all public servers), that contain a search-hit.
If you click on an entry you get as result a concept map with the cursor positioned to 1.- .. 4. accordingly ... ready for input (depending on access-rights).
A preview of the search hit context, like Google does it with text only or imbedded thumbnails, might be indispensable.
I could even imagine tiny annotatable, clickable and rated HTML or "real" Cmap "thumbnails" in the preview.
The network effects of synergies ( http://en.wikipedia.org/wiki/Metcalfe%27s_law ) coming from that, would be a real blessing for the growth of the Cmappers Community (and beyond).
This would (imho) substantially contribute to the forming of the GlobalBrain towards our way to http://google.com/search?=TheSingularity
Best
Fridemar
PS.: Suggested linking labels would probably help to optimize the indexing of the CmapSearcher.
1. single concepts C ( for which there may (or may not) exist completing linking labels or objects D with (C,l,D) or (D,l,C) )
2. pairs of concepts (C,D) ( for which there may (or may not) exist a linking label l with (C,l,D) )
3. pairs of concepts and linking labels (C,l) ( for which there may (or may not) exist a concept D with (C,l,D) )
4. pairs of linking labels and concepts (l,D) ( for which there may (or may not) exist a concept C with (C,l,D) )
5. linking phrases l ( for which there may (or may not) exist concepts C,D with (C,l,D) )
6. propositions (C,l,D) as complete Semantic Triples
As result should come a (possibly empty) list of concept maps of all public available cmaps (of all public servers), that contain a search-hit.
If you click on an entry you get as result a concept map with the cursor positioned to 1.- .. 4. accordingly ... ready for input (depending on access-rights).
A preview of the search hit context, like Google does it with text only or imbedded thumbnails, might be indispensable.
I could even imagine tiny annotatable, clickable and rated HTML or "real" Cmap "thumbnails" in the preview.
The network effects of synergies ( http://en.wikipedia.org/wiki/Metcalfe%27s_law ) coming from that, would be a real blessing for the growth of the Cmappers Community (and beyond).
This would (imho) substantially contribute to the forming of the GlobalBrain towards our way to http://google.com/search?=TheSingularity
Best
Fridemar
PS.: Suggested linking labels would probably help to optimize the indexing of the CmapSearcher.
Let's FosterEachOther ( <- Google )
-- Fridemar
-- Fridemar
-
fridemar
- Posts: 56
- Joined: Wed Apr 29, 2009 12:11 pm
- Location: Germany
- Contact:
Link correction: http://google.com/query?q=TheSingularity. E
Link correction: http://google.com/search?q=TheSingularity.
EOM=EndOfMessage.
Sorry for the inconvenience.
By the way: A wiki or a CMap would be more convenient than a forum to make such corrections on the fly
EOM=EndOfMessage.
Sorry for the inconvenience.
By the way: A wiki or a CMap would be more convenient than a forum to make such corrections on the fly
Let's FosterEachOther ( <- Google )
-- Fridemar
-- Fridemar