tonja
Hi,

Why it isn't possible with PB 5.5 to design multiple links and with individual weights per linktype?

In the attacement you find the following model:

Thoughts: I, You
Relationships: I_hate, I_love

Remember Shakira's song: I hate You - I love You

If I want to model this thinking, I must dublicate You, so I can hate you and I can love you. This looks in the datamodelling view quite stupid but it works.
Now I build up the relationships I hate You and I love You, even I can determine the strength of the relationships as link type strength, which is something like a weightening.

Ok, but now I look how I can proceed with my relationship to Icecream. If I want to reuse the old reationships {I_hate, I_love} I run into the disadvantage that I inherit the strength vom I hate You  and I love You. So I hate und love Icecream the same like You, this is nonsense.

An alternative would be I define the realationships after an cross-table {I_hate, I_love} {You, Icecream}. This would be a totaly ugliy datamodell, because I need to dublicate all Thoughts so define at least 2 relationships {I_hate, I_love} and on the top I multiply the  type of the relationship {I_hate, I_love} with all Thoughts {You, Icecream, etc.}.

Why the PB-datamodell cannot be enhanced in this way?

- To define N relationships between Thoughts (knots) will be possible.
- Each relationship can have her own weigthening which will shown in the dashing strengs too. If there is more than one relationship I see an double line and if I hover ofer this doble-line an popup shows me the information in full kardinality: N relationships with N weightenings and orientations.
- With the reporting it is possible to filter after the multiple relationships and/or weights. It is possible to save this filtered views (What I hate most, What I love most...).

This would be very helpful to run the full kardinality of an flexible datamodell!

Greetings,




Song: I hate You - I love You




Quote
Adathome
tonja wrote:
Hi,

Why it isn't possible with PB 5.5 to design multiple links and with individual weights per linktype?



See my request "13-01-2008" in post: http://forums.thebrain.com/post?id=2418770

please put your vote here:
http://thebrain.uservoice.com/forums/4597-thebrain/suggestions/74550-more-advanced-link-types?ref=title

and your vote here:
http://thebrain.uservoice.com/forums/4597-thebrain/suggestions/83489-typed-links-have-a-direction?ref=title
Regards, Ad Divide knowledge = multiply knowledge (Windows 10 -  TB8 / TB9)
Quote
tonja
Hi Ad,

thanks for the hint and concept to enhance the link types, where I like:

- the switch from simple 2 complex
- weightening of link types
- multiple types in parallel and the visibility flags
- differentiating between link types (categories) and relationship types (verbs)

To complete this modell there is only missing:

- "Link Thoughts" a connection from the middle of an Link to a new Thought
   (Why I like icecream?.... Because...in my tiny I hate I love you model)
- Timedependencies: "I like xy ... from 2010 until 2012"
- Geo-Tag: "I meet him ... in Zürich in the trainstation"

Greetings,




Quote
rhodes
Quote:
"Link Thoughts" a connection from the middle of an Link to a new Thought


I think this could be overcome by having independent, permanently visible link labels, a feature which has long been requested.
Alan Rhodes
Quote
tonja

Hi,

the last which is missing is an selfdepended like-type.

which is also the message of this song...

"I hate to love You",

this is an self-relationship.

P.S. It would be quite funny to use PB also for Freud's psychology-modells. 

Quote
rhodes

Hi tonja

In your first post in this thread you asked:

Quote:
Why the PB-datamodell cannot be enhanced in this way?

You call is for for multiple, weighted links per Thought. This topic periodically comes up, raised by like-minded PB users such as Spacenexus (Jim), who would like to see PB act not just as a passive database of loose associations, but as an active participant in our mental logic. I love your phrase "topology of thinking" - that sums it up so well.

I'm sure many people try out PB hoping it will assist their reasoning in the same way as mind maps, concept maps, etc. They are looking to build an AI, expert system shell out of PB. This is a fascinating holy grail. There's no reason it can't be done in the distant future. Even though PB's marketing department and User Guide use terms such as "Brain", "Thought", "forgetting", "remembering", etc., rather loosely IMO, this does not diminish PB's unique abilities as a data storehouse. If you Google for "knowledge visualization" as you have probably already done, you will find numerous attempts to create topologies of thinking.

PB's structure is based on a few very simple ideas - multiparent hierarchy, jump-linked remote thoughts, etc. Keeping it simple also means keeping it fast. Human reasoning is far more complex, and the effort needed to translate this into a individualized visualization database would probably consume far more time than the average human can spare. I would like to see the developers offer an SDK interface for AI programmers to create something truly unique and wonderful, but doubt it will be allowed to happen.

Organizing and displaying data is what PB is best at. Your thread has made us more aware of the steepness of the climb to the next level to which PB can aspire.

Alan Rhodes
Quote
tonja
Hi Alan Rhodes,

and PB-API for developing expertsytsems would be realy cool¨!

A superfunction would be:
An API which is fits to Mathematica, so this "Lisp++" will be a strong logical layer. The prob is only to transport the clowds of Information their class, superclasses and metaconnections. But their should be some technical solution.

I'm very keen to see the result.

Greetings,



Quote

Newsletter Signup  Newsletter        Visit TheBrain Blog   Blog       Follow us on Twitter   Twitter       Like Us on Facebook   Facebook         Watch Us on Youtube  YouTube       

TheBrain Mind Map & Mindmapping Software     Download TheBrain Mind Mapping Software