I am also interested in recording people and families in TW. I'll refrain from calling what I do genealogy - I just collect information, pictures, and stories from family members and bits I find online. I'm not too concerned about the stories being perfectly factual and write most entries (tiddlers) in a free-form format with manual links to parents and children. This does not lend itself to producing trees or graphs, or even generated lists.
I did write a script in the Ruby language to convert a gedcom to a TiddlyWiki. It works acceptably well. Only thing is the wiki is no longer human editable which is important if I want to get other people involved in the data entry and editing.
Creating diagrams or trees is extremely useful though and something I'd like to do. I also think going straight to graphs is putting the cart before the horse. There needs to be a system on place for recording basic facts about a person - relationships, relationship types, dates, locations, events, notes, sources, and so on. GEDCOM is the defacto standard for this. It's a fairly flexible format which is why I think it has survived over time, although it is not without issues. Such a system might include:
* a standard format for recording facts as fields
* standard formats for recording dates, locations, names in fields
* a set of templates for displaying and editing people and families
* a set of helper macros to "add new person", "add person as child", "add as parent", etc.
* more macros to calculate relationships between people (Person A and Person B are 2nd cousins)
* probably a lot more
It may be even possible to offer import and export of simplified gedcom files but only if said standards were rigidly adhered to.
It's a big job, and genealogy is complicated and messy. People and their relationships don't fit neatly in compartments. All of which makes me wonder if the use of TW is even a viable option.
Sorry this turned into something of a random thoughts rant. Need to run, tho.
Thoughts?