Not withstanding a need for performance.
If this is mearly a list of titles there is less value in visualising it beyond a title list, if there is a complex set of relationships between a thousand items a thousand nodes can easily become 10,000 relationships and whilst you may be able to represent this in a graph, the value for communicating with the human visual system and brain is questionable. It seem better to build a solution where we focus on a subset of the total set and navigate from there.
- With large numbers there are other representations such as column graphs, expandable lists, statistical summaries and click to expand that would serve better.
See my last reply that references here which includes this idea and more, and @Flibbles is reviewing.
- Somtimes a simple list of titles is best for this kind of relationship.
There is an argument that could be put forward that this is under using tiddlywikis capabilities, the tiddler can be like a record in a database and to suggest there is no record suggests you are bypassing a key structure. Sure there will be exceptions and these can be handled like we often do using virtual tiddlers, or titles drawn from a list of some kind that are not actual titles. Including;
- Using any lis as input titles need not have underlying tiddlers
- There are powerful features and tricks available for handling missing tiddlers
Mapping nodes to titles, if they exist or not, perhaps arguably also fields (or their contents) map to relationships, helps keep a tiddlywiki flavor and ease of comprehension.