buku

y don't offer sufficient longevity, even should they meet my other requirements, which they don't. Assigning metadata is a significant pain point. Coming to some aggreement as to what metadata to assign is a signficant pain point. I'm coming to see librarians and library cataloguing as essential domain knowledge and experience. In all seriousness, I suggest any project looking to make use of categories and classification look to the US Library of Congress Classification System: it's extant, expert, unencumbered, comprehensive, hierarchical, extensible, has a change management process, and is applied to a store comprising 164 million works. https://mammouth.cafe/@dredmorbius/56485 http://www.loc.gov/catdir/cpso/lcco/ There's also a top-level reduction to 21 distinct categories, and the possibility of, say, coming up with a short-list of frequently-used classifications, as well as of assigning multiple classifications to works. The rationale for only storing bookmarks is ... generally not valid. There are a few types of online resources, generally: 1. Interactive or volatile pages. 2. Static pages. For the first, search engines, web apps, landing pages, etc., storing a static instance isn't tremendously useful (though it can be more useful than you'd think). For the second, a locally-stored version is almost always more useful than the online instance. And space for text is now beyond cheap. I'm looking at this problem in terms of desired outputs, workflow, various states of resources, how to (reasonably) uniquely and persistently identify a given document, managing media (images, audio, video, other interactive elements), etc. And yes, this starts to look rather much like Memex, for similar reasons.