On Wed, Dec 15, 2010 at 5:55 PM, Yakov <[email protected]> wrote: >> Yes, this one keeps eating away at me. There are powerful attractions >> for keeping tiddlers in working storage in a more structured tree >> format that makes it easier to do this kind of thing. (Another example >> of something that's hard to do with the current architecture is a >> checkbox that is typed [ ] and gets changed dynamically to [X] when it >> is clicked). > > I didn't really get this. What's the connection between rebinding > (which should be based on the "references" list) and .. what the > "working storage" really is? Neither I understant the issue with a > checkbox.. I mean it's connection with the problem.
Basically, because tiddlers are just stored as blobs of text, there is no easy way to retrieve all the references to a tiddler. So, internally, TiddlyWiki scans each tiddler to look for links within it, and then caches the outgoing links against each tiddler. This allows the core to (relatively) quickly produce things like Missing, Orphans and References. However, there a lot of problems. The link scanning needs to be quicker than a full scale wikification of the text, and there are subtle mismatches between what the link scanning algorithm sees as outgoing links, and what the wikification process sees. One approach would be to apply more consistent pre-processing to each tiddler, and maintain them in a more richly structured store that could support these link-based scenarios more quickly. Going the whole hog, we could store tiddlers internally in a tree structure (kind of like a parse tree). If we went that far, it would make it much more feasible to offer a feature that people have asked for since the early days - the ability to type [ ] and have it rendered as a check box that gets flipped to [*] when it is clicked. At the moment that's pretty hard to do because there is no relationship you can navigate between the checkbox in the DOM and the fragment of the raw tiddler text that it corresponds to. > I'd note that the references list should be formed a bit more > carefully: as for now, links placed in comments (which can be sections > that are shown with sliders) are not counted as references; as I > remember, in some tests transclusion macros also didn't cause a > creation of reference; but I'm not sure. That's the problem I'm referring to: false negatives and positives with the link detection. Cheers Jeremy -- Jeremy Ruston mailto:[email protected] http://www.tiddlywiki.com -- You received this message because you are subscribed to the Google Groups "TiddlyWikiDev" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/tiddlywikidev?hl=en.
