Research Library Services

[bnewbold]

I think it would be helpful to provide local access to the "scholarly record": research papers, perhaps books, notable blog posts, etc. This would enable participants to check citations and dig in to knowledge a layer deeper than an encyclopedia. It also provides a convenient content set / corpus for people experimenting with distributed storage and communication systems on-site.

The most basic service to provide is a simple lookup when somebody already knows an identifier (DOI) or title/author, and return a PDF, though an HTTP web interface. More advanced services, which would be progressively harder to provide, are a dedicated lookup station (large monitor, seat, etc; would work if local network is down); print-out services (double-sided; color?); access over multiple protocols; accepting deposits and metadata edits (wiki-style); fulltext search and other rich "discovery" services (eg, "show me the best paper/book about carrot night vision"); additional media types (eg, XML, HTML); expanded coverage of books and web content.

I'm interested in trying to run such a service; it overlaps with the papers archiving work I do at the archive (https://fatcat.wiki). My basic plan would be to run:

• catalog of about 100 million papers, with fulltext copies for 10-20 million of thoseu
• a small wired front-end server (4 cores, 16+ GB RAM, 2TB+ SSD) to provide HTTP access to the catalog, with metadata (but not fulltext) search
• a fileserver with about 50 TByte of disk storage. I have a GNUBee device i'm curious to try, but a proper NAS/server, or even USB hub and a pile of external disks might be more practical
• laser printer with consumables to print about 200x 5 pages hard copies
• single-seat workstation (raspberry pi and a monitor?) attached to printer
• available over wired ethernet, bridging to whatever mesh/wireless or dweb tunneling folks are interested in
• be on-hand (a booth?) for some fraction of the time to help folks look things up and discuss distributed libraries

Happy to collaborate if others want to bite off some of the more ambitious options! Other brainstorm-y ideas:

• include datasets; not sure how to curate which ones would be interesting
• a place to "deposit" research works created at the site (datasets, pre-prints, etc)
• run a full-on book printer!
• library of 3D printing design files (probably a separate project on it's own)
• ensure ahead of time that notable/relevant works are in the library (whitepapers, protocol specs, books, etc)
• run/experiment with more dweb-style libraries

Some more thoughts:

• is the ability to "just look it up" actually desirable at an event like this? is it an intrusion of outside authority?
• does such a large body of work impede local creativity, or does it enhance it via remix/inspiration?

[benhylau]

When in conversations with people, I often get into these "ahh there's this article / paper" which I remember a partial title or the name of the author... except there's no Internet and I cannot recall the exact article. I think it'd be cool if there's a searchable offline library on site where I can say "why don't we walk over to the library and print this paper!" I think people will find this a useful resource.

I also wonder if the larger discussion of "local access to digitally archived content" can have some discussion space. Like where you mentioned:

be on-hand (a booth?) for some fraction of the time to help folks look things up and discuss distributed libraries

There may be some shared discussions with @mitra42's idea on #3

In terms of logistics, we are planning an indoor space that should have wired connections to the local mesh. Imagine a table with a couple seats in a large hall with wire drops where you can connect a NAS or a Raspberry Pi tied to an external HDD, where you can also put a laser printer, perhaps put up a schedule of some discussion activities...

is the ability to "just look it up" actually desirable at an event like this? is it an intrusion of outside authority?

I think this is very helpful and as long as the resource is pull-based (searched) rather than pushed onto people, it's not an intrusion at all!

[mitra42]

Right ... we'll have at least one mini-server (RPI or Rachel or could be something bigger) on site that can Crawl / Proxy / Serve content from the Archive.

[benhylau]

Is this and #3 the same project, or different? My original understanding is they are similar, but @bnewbold's is more specific to https://fatcat.wiki and papers, whereas #3 is another front-end and perhaps with multimedia content (e.g. 2018 recordings of DWeb Summit)?

[mitra42]

These are different projects, #3 is a generic IA server, the UI is a subset of the Archive UI, the content is standard UI collections, while if I understand @bnewbold 's its specialized/optimized to research papers.

In particular #3 is actually intended to be part of a decentralized internet archive, to complement the other servers (Wikipedia, Khan Academy, Open Street Maps etc) typically installed in especially disconnected or marginally connected community networks.

[benhylau]

This project is published onto https://dwebcamp.org/proposals/ thanks for submitting @bnewbold :)
Since we'll both be at IA in June / July let's work out logistics then.

[benhylau]

@bnewbold can we try this out at https://www.eventbrite.com/e/all-day-hackathon-for-dweb-camp-tickets-63718679285

[bnewbold]

Cross-referencing to hackmd.io:

• "Research Library Services" (DWeb Camp 2019)
• Written Content Requests (Dweb Camp 2019)