It was a nice group of people, and I was able to put some faces to names of people I’ve had in my feedreader for a good while. I had to miss the start, which was family dinner / putting our 5yo to bed time in our timezone, but was able to join three sessions afterwards. The conversations were interesting and gently paced. Thank you to Chris Aldrich for organising, and to all participants for their contributions to worthwile conversations.
The videos and notes are / will be linked to on the event page (see link in first sentence).
The sessions were:
- Ad hoc book discussion clubs and sessions (how to use personal libraries to facilitate ad-hoc discussion of a book with other current active readers of the same book?) Moderated by Maggie Appleton
- Decentralised bookshelves (What conventions do we need to reach a useful level of interoperability between the different ways people and platforms make such data available) Moderated by Manton Reece
- Book Identifiers (There are multiple identifiers in use for books, ISBN, OLID, ASIN, WorldCat etc. How can we interact with them, how are we supposed to, how is it useful?) Moderated by Tom MacWright.
While these were three distinct sessions, to me it felt like basically the same ongoing conversation, so my impressions aren’t tied to specific sessions as such.
Elements that stood out for me, or that I realised listening to the other participants:
- OpenLibrary is a good neutral way to link to a book (I avoid Amazon as well as Amazon’s GoodReads links, and publisher’s or author’s links aren’t always available), and they have an API. Missing books can be added, as OpenLibrary is a wiki. They also make a useful difference between the work (the book written) and the editions (the book version you’ve read) that they list for the work. I may want to check out their API and see if I can use that for my internal book notes and/or public book postings.
- Calibre, an app to manage e-book collections I use for my non-Kindle books, has an API as well. It makes me wonder the same about Delicious Library, of which I have a 2012 database somewhere, from the 500 or so physical books we did away with that year.
- Discovery to me requires both information about the book and about the readers, so I can navigate triangles, the key element in social media. To evaluate recommendations they require information about the person making the recommendation more than about the book being recommended.
- I’m not interested in pretend-social information around books that really are masked statistics. They may seem to provide what I seek in the previous point, but actually provide a meaningless regression to the mean. Things like ‘others who bought this book also bought’ don’t increase the space of discovery but will eventually always limit the space of discovery to the fat head of the long tail. The stuff far down the long tail sees too little interaction for such ‘also bought’ algorithms to aggregate.
- Whatever you want to do with book information, you need to first publish data about your reading in each case. So that is the focus. I’ve been collecting a list of some URLs where people share book lists they’ve read. They are all different it seems to me, but at least the data is out there to try and consume in a personally meaningful way.
- Whatever is consuming book related data or posts, needs to take on the burden of figuring out what identifiers are used, or what other meaning can be gleaned from it. A to me key remark made by Jacky Alciné. The flip side for me is, I only need to concern myself first with publishing information that is meaningful and useful to myself.
- It is possible to help others though by providing multiple formats. Specifically now that I have automated generating OPML from my booknotes for my booklists. Creating the same lists in CSV, in RSS, or JSON for instance is easy enough to do now. This I added to my list of small side projects.
- Book lists are basically just spreadsheets Drini Cami of the Open Library / Internet Archive remarked. This ties into the above. I also realised it’s true for pulling together the data about books I bought and read from the various platforms I use. So this morning I pulled the information about the 800 or so books I bought with Amazon over the past 15 years into a spreadsheet. This is a first step to backfilling my book notes and reading lists, as well as my anti-library, using a script.
- I’m not very interested in algorithms across the reading patterns of the general population of readers. This is another statistic basically, not a socially meaningful thing. I would be very interested in running algorithms or analysis across the information about other people reading, who are within my scope. E.g. the bloggers I follow in my feed reader, or the bloggers they follow. An algorithm that serves me, not describes or commodifies me, an algorithm as personal assistant. That way my own preferences can be its default.
- We talked about ad-hoc book clubs, both where a book is the key point, as well as where the socialising is more important. I’m on Bookwyrm.social which aims to be a federated GoodReads, something that is in itself not appealing to me as just a means of replicating the same data as on my site. But I have seen instance form around an established group or niche themes. That is a more appealing usage. Only afterwards I thought of how this ties into the Micro.blog Readers Republic we recently formed. A next meeting is in 2 weeks, I’ll think about how to feed some of these discussion back to that group.
- Amazon with their ASIN numbers is messing up the discovery value of ISBN numbers, by deliberately removing the relation between the two. That frustrates interoperability with other platforms and resources, which is their point I’m sure.
That’s a good list of take-aways from a few hours of conversation!