Soren,
I have read much of this and will continue to do so (and reread), I think
it is important for tiddlywiki itself that we explore knowledge and
information modeling/management. This is a sizable contribution to the
discussion. As I read I have seen a few alternate paths both converging and
diverging from your ideas, perhaps I will discuss these in details later
but here are a few high level of concepts;
- Indexes can be built automatically from non-trivial words and capture
frequency information
- Perhaps this is how one generates potential keywords
- I imagine when saving a tiddler a keyword field opens with
suggestions you can accept/reject/add to.
- Comparison between keywords the author selected and those found in
the text could discover new relationships
- There are algorithms such as google language translations that
determine how closely related one word is to another statistically to
learn
about relationships.
- Is a "related" word found in the same sentence, paragraph,
tiddler etc... semantic structure
- In fact one way to store text is to store every word in an index
and store the link to the word not the word itself when we think like this
we can see new possibilities.
- Access to simple synonyms and antonyms in the language used and
extending searches to also search for the use of the synonym's in addition
to the named word.
- It is possibly to expose content to the google search engine and use
google to search your own content.
- TOCs and Outlines, Hierarchies in effect, I am a fan but also aware of
the apparent limitations, however if you recognise the same data can exist
in multiple hierarchies, or not in a hierarchies.
- Learning how to flag "missing data" or tentative knowledge/links
etc... helps exploration and consistency.
- We should find ways to ensure for all content we know the semantic
structure from tiddler title to tag hierarchies and html headings,
paragraphs sections...
- It is easy to extend Tiddlywiki's search into other fields buy
changing the scope of a search, even comparing the results of different
searches can produce useful insights
- ...
Very interesting
Tones
On Tuesday, 10 August 2021 at 11:17:00 UTC+10 Soren Bjornstad wrote:
> Some of you all might be interested in this new post on my blog:
>
>
> https://controlaltbackspace.org/notes/better-indexes-through-semantic-modeling/
>
> It's a proposal for a system for indexing large documents based on a
> hypertext graph, including a discussion of a possible TiddlyWiki prototype.
> Warning: 6,000+ words.
>
--
You received this message because you are subscribed to the Google Groups
"TiddlyWiki" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/tiddlywiki/8c9b6b74-595a-4a3b-86c4-3fc6056c2e82n%40googlegroups.com.
