Composing stuff from smaller pieces definitely lurks in the back of my mind when I think about TiddlyWeb. It's not in the forefront at the moment because I'm trying to get TiddlyWeb to hit its original use case (composing TiddlyWikis). Once that's nailed it should be possible to expand to more stuff.
I think Ian's idea is a good one, especially if you are working in a WSGI stack it is dead easy.
Where I find the challenge with these sorts of things is how to communicate (between client and server and server and storage) the notion of the graph that represents the document or partial document involved. I've tried all kinds of things and it has all been rather complex. Which seems a bad smell.
I guess you have to make a decision one way or another whether your document is a tree, in which case a path to the sub-tree or filter criteria is sufficient, or whether it's a lattice in which case the query language (potentially) gets more complex.
I assume that some sort of relational query language would give the most power and flexibility but might be hard to cram into a URL.
2 comments:
Composing stuff from smaller pieces definitely lurks in the back of my mind when I think about TiddlyWeb. It's not in the forefront at the moment because I'm trying to get TiddlyWeb to hit its original use case (composing TiddlyWikis). Once that's nailed it should be possible to expand to more stuff.
I think Ian's idea is a good one, especially if you are working in a WSGI stack it is dead easy.
Where I find the challenge with these sorts of things is how to communicate (between client and server and server and storage) the notion of the graph that represents the document or partial document involved. I've tried all kinds of things and it has all been rather complex. Which seems a bad smell.
I guess you have to make a decision one way or another whether your document is a tree, in which case a path to the sub-tree or filter criteria is sufficient, or whether it's a lattice in which case the query language (potentially) gets more complex.
I assume that some sort of relational query language would give the most power and flexibility but might be hard to cram into a URL.
Post a Comment