4 ms·

Regarding point 1, are.na effectively does this for bookmarking, by showing you the categories that others have stored the same bookmark to (so really showing y

by setr 4y ago

Regarding point 1, are.na effectively does this for bookmarking, by showing you the categories that others have stored the same bookmark to (so really showing you what they thought was related, according to whatever relationship they were interested in) which I think is a much more reliable/same way to approach the problem — its much more precise, as the other user was literally on the same page as you at some point, rather than whatever tangential topical relationship identified by keyword matching. And it’s more broad, as the relationships you can find are not simply the literal topic itself (e.g. hexing the technical interview might be stored by one user alongside programming taos, and by another with closure hackery, and another with programmers-as-wizards).

Abishek_Muthian 4y ago
Very interesting, Thanks for introducing me to are.na. I like that they've essentially outsourced the computation required to classify the content to real humans. I understand the value in adding context to the bookmarks, But does it help in any way to recall a bookmark when I need it? Because not using the bookmark is a real problem[1] but I find the suggestions from bookmarks in Firefox address bar quite useful; Integration of in other writing tools which knows the content of the bookmarks could be useful. [1] https://needgap.com/problems/57-i-forget-my-web-bookmarks-quite-often.-internet-organizing https://needgap.com/problems/57-i-forget-my-web-bookmarks-qu...
- setr 4y ago
  > 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

by setr 4y ago

Abishek_Muthian 4y ago

Very interesting, Thanks for introducing me to are.na. I like that they've essentially outsourced the computation required to classify the content to real humans. I understand the value in adding context to the bookmarks, But does it help in any way to recall a bookmark when I need it? Because not using the bookmark is a real problem[1] but I find the suggestions from bookmarks in Firefox address bar quite useful; Integration of in other writing tools which knows the content of the bookmarks could be useful. [1] https://needgap.com/problems/57-i-forget-my-web-bookmarks-quite-often.-internet-organizing https://needgap.com/problems/57-i-forget-my-web-bookmarks-qu...

setr 4y ago
> 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