> But does it help in any way to recall a bookmark when I need it?
Not particularly, except that the same bookmark can be placed in multiple groupings, so the usual organizational structure used is typically more conceptual; e.g. I had a grouping for Anime Recommendation System, which shares a bunch of links with CS Whitepapers as well as Social Dynamics; and of course Evangelion gets plastered in all sorts of random topics. CS Whitepapers is a fairly useless category for recall but a nice place to dump things, whereas Anime Recommendation System is a more focused project — and if I’m looking for recommendation system papers, I’ll obviously remember the latter more clearly, and find the related.
So basically recall by duplication, and the organizational flexibility.
Recall by auto-suggestion I think is useful in a different fashion, but vaguely meets the above by accident. It mostly serves to:
1. Help you complete a URL you only partially remember
2. Show you bookmarks that are strictly related to what you’re doing, based on similar keywords + a neighboring expansion of them
3. Accidentally show you bookmarks completely unrelated to what you’re doing, but causes a mental refresh - this is where it mostly competes with are.na
In other words, the autosuggestion is googling, and are.na is acting like Wikipedia. The former is much faster, and better when you’re precise, but its similarity metrics are quite limited. Whereas the latter can cross much wider chasms of similarity
Not particularly, except that the same bookmark can be placed in multiple groupings, so the usual organizational structure used is typically more conceptual; e.g. I had a grouping for Anime Recommendation System, which shares a bunch of links with CS Whitepapers as well as Social Dynamics; and of course Evangelion gets plastered in all sorts of random topics. CS Whitepapers is a fairly useless category for recall but a nice place to dump things, whereas Anime Recommendation System is a more focused project — and if I’m looking for recommendation system papers, I’ll obviously remember the latter more clearly, and find the related.
So basically recall by duplication, and the organizational flexibility.
Recall by auto-suggestion I think is useful in a different fashion, but vaguely meets the above by accident. It mostly serves to:
1. Help you complete a URL you only partially remember
2. Show you bookmarks that are strictly related to what you’re doing, based on similar keywords + a neighboring expansion of them
3. Accidentally show you bookmarks completely unrelated to what you’re doing, but causes a mental refresh - this is where it mostly competes with are.na
In other words, the autosuggestion is googling, and are.na is acting like Wikipedia. The former is much faster, and better when you’re precise, but its similarity metrics are quite limited. Whereas the latter can cross much wider chasms of similarity