[HN Gopher] The Bookmarking Data Model Is Wrong for Highlighting
       ___________________________________________________________________
        
       The Bookmarking Data Model Is Wrong for Highlighting
        
       Author : bsnnkv
       Score  : 41 points
       Date   : 2023-05-16 16:40 UTC (6 hours ago)
        
 (HTM) web link (lgug2z.com)
 (TXT) w3m dump (lgug2z.com)
        
       | clueless wrote:
       | so essentially what amazd (1) could have been 10 years later if i
       | had kept working on it. I remained amazed (pun intended) that 10
       | years later (after I started working on this), we still don't
       | have a very popular platform for this sort of workflow. Full
       | disclaimer: I've come to use readwise.io more recently
       | 
       | (1) http://amazd.com/ahmad
        
         | bsnnkv wrote:
         | My friend, you were truly ahead of your time. If I had known
         | about this back in 2020 I probably would have bugged you to
         | restart working on it instead of building Notado! Would love to
         | chat with you sometime if you want to reach out on
         | Mastodon/Twitter etc.
        
       | tuukkah wrote:
       | > _the URL is an imperfect proxy for the article_
       | 
       | I find this part important but misleading. The URL is a fine way
       | to identify the article, _but_ it 's important to separately
       | identify and link to the multiple _discussions_ that can be had
       | about any given article. For example, an article URL doesn 't
       | unambiguously identify a HN discussion as the same URL is
       | submitted and discussed multiple times.
       | 
       | Rather, this may be the gist of OP: Because we post URLs without
       | top-level commentary here, the HN discussion (and HN URL) isn't a
       | single, meaningful conversation either. Only when we get to each
       | top-level comment (identified by the comment URL), we have
       | conversations that it may make sense to highlight.
        
       | CrampusDestrus wrote:
       | If you want to call it with an early-2000s-vabing name,
       | "webquotes"
        
       | bsnnkv wrote:
       | More context from a HN thread here.[1]
       | 
       | [1]: https://news.ycombinator.com/item?id=35942325
        
       | neilv wrote:
       | FWIW, lately I've been keeping my notes as markdown files in a
       | `~/wiki/` directory, and the most common links are now HN posts.
       | 
       | I find the HN discussion often adds a lot that will be relevant.
       | For example, in `~/wiki/machine-learning.md` the section on
       | specific LLM models is a nested bulleted list, with the top level
       | the names of specific models (sometimes as links to their
       | canonical landing package, and the bullets under each are usually
       | very relevant HN post links, and links to official other pages
       | for that model.
       | 
       | I only polish it when I really have to: the priority is to
       | capture the info, because 15 seconds now might save me days or an
       | entire endeavor later. But spending more time than that can
       | discourage capturing info, or later make me resistant to doing a
       | quick split of pages or sections that really should happen
       | immediately (because I don't want to spoil the polish that I
       | spent time on).
       | 
       | In Markdown, my current HN post links look like:
       | [HN, The Bookmarking Data Model Is Wrong for Highlighting
       | (lgug2z.com),
       | 2023-05-16](https://news.ycombinator.com/item?id=35964335)
       | 
       | (My Emacs mode hides the URL part, and makes it look like an
       | underlined blue hyperlink, with the remaining markup characters
       | de-emphasized.)
       | 
       | It's a variation on the familiar citation format, which I also
       | use:                   [Jane Doe, "Some Article Title", *Some
       | Journal*,         2023-05-16](https://journal.example/12345)
       | 
       | In the rare case I want to quote an entire HN comment in my wiki
       | (e.g.,, some super-useful insider scoop that might disappear),
       | I'll just click to its page, select it and copy&paste the test
       | with `>` characters, and then turn the header line into a
       | Markdown link to the comment page URL.
       | 
       | Someday, I'll make a browser plugin for the copy&paste parts,
       | while still keeping it very simple.
        
       | idlewords wrote:
       | This discussion conflates comments with highlights, which are two
       | different animals altogether.
       | 
       | The distinguishing feature of highlights is that they are
       | anchored to a specific passage in a (hyper)text, which you want
       | the bookmarking tool to make visibly distinct, searchable, and so
       | on. Ideally you want this feature to work seamlessly across
       | paginated texts as well as on PDFs.
       | 
       | Comments are an entirely different animal, and the way they do
       | (or don't) interact with highlights will depend on how people use
       | the site socially, which is not really within the control of the
       | site author.
       | 
       | Binning bookmarks, highlights, and comments together as "content"
       | is like calling wedding cake, bobcats and lighter fluid different
       | kinds of "matter"; it's reductive to the point of uselessness.
        
         | bsnnkv wrote:
         | Hard disagree with everything in your comment, but it's cool to
         | have elicited a reply from the Pinboard guy.
        
       ___________________________________________________________________
       (page generated 2023-05-16 23:01 UTC)