Awesome Box was an Amazing Experiment. Thank you!

Awesome Box was a highly successful experiment that helped LIL explore new ways of enabling peer to peer reading recommendations in libraries.

The Awesome Box was a physical box that a library would sit next to the library’s regular returns box and if you thought the book was mind blowing, you dropped it in the Awesome Box instead of the regular returns box. The librarian then has the option to scan the book into the Awesome Box website to enable digital sharing of lists of awesome items. Or, the librarian can keep things no-tech and put the item on a shelf labelled Community Recommendations.

Annie Cain and I created the Awesome Box after hearing about a similar idea functioning in a European library. In 2013, we developed the web app, received a little grant funding from Harvard’s Library Lab and the Arcadia Fund, and started collaborating with libraries at Harvard, Somerville Public (first Awesome Box in the wild!!), Cambridge Public, and Brookline Public here in the Boston area.

Annie and I (with Annie doing the lion’s share) worked hard to develop the Awesome Box community by quickly replying with advice when emails arrived and talking about Awesome Box at several conferences and gatherings of librarians.

I learned a ton about product development and adoption with the Awesome Box, but two big things that stick out after much reflection — make the thing you’re building fit with the patterns of the folks that will use the thing (people are returning books anyway, they just need to choose a box), and you have to sell, sell, sell! Awesome Box is fun and free (as in open source and as in no money) and we still constantly talked it up and pushed it for three years. I’ve found that it’s hard to find success with a project if you just dump on the web and expect people to use it — you’ve got to wire people to your project.

Awesome Box is certainly one of the most successful projects I’ve been lucky enough to be part of. And, arguably, one of the most successful projects to roll out of LIL. Thank you so much to all the libraries that joined together to make Awesome Box so much fun! If you’re a library and you didn’t have a chance to export your Awesome items, please drop me an email and I’ll get your data to you.

Awesome Box was an experiment. It’s done and the servers have been powered down. During it’s glorious run, the Awesome Box supported 512 private, public, and academic libraries across the US. The members of those libraries dropped 104,715 items dropped in the Awesome Box from 2013 to 2016.

Thank you. Thank you. Thank you.

Physical Pitch Decks

I've been playing with physical pitch decks lately. Slides as printed cards.

PowerPoint, Deck.js, and the like are fantastic when sharing with large groups of people — from a classroom full of folks to a web full of folks. But, what if easy and broad sharing isn't a criteria for your pitch deck?

You might end up with physical cards like I did when I recently pitched Private Talking Spaces. The cards are surprisingly good!! Just like non-physical slides, they can provide outlines for talks and discussions, but they're so simple (just paper and ink), they won't get in the way when sharing ideas with small groups.

The operation of the cards is as plain as can be - just take the card off the top, flip it upside down, and put it to the side.

n cards = n screens in the world of physical pitch decks. I wish we had multiple projectors in rooms! In the photo above, I pin my agenda card up top.

I drew the slides in Adobe Illustrator. They're six inches square and printed on sturdy paper. If you'd like to make your own, here's my .ai file and here's a .pdf version.

It feels like there's something here. Some depth. If you've had success with physical pitch decks, please send me pointers. Thanks!!

pockets of people

we hosted a bunch of amazing visitors earlier this week (knight prototype workshop!) and we were fortunate enough to gather everyone for dinner. after drinks were served, i used my phone's camera and swooped into each booth aka pocket of people.

swooping into these pockets of people is surprisingly meaningful and rich — i very much get a distinct sense for the vibe/mood/energy at each table. this swoop in and pan pattern is deep.

what should i do with these clips? feels like there's some coolness here but i can't seem to grab it. ideas?

Nuremberg Tribunals Project Launches New Website

9019102

The Nuremberg Tribunals Project is excited to announce the launch of its new website.

Our team at the Harvard Library Innovation Lab and the Harvard Law Library's Department of Historical and Special Collections has been working hard the past year to create a new, rich, flexible and visually appealing discovery and viewing experience for our 750,000-page Nuremberg Trials archive.  The archive materials comprise the full document record for all 13 Nuremberg Trials, held at Nuremberg 1945-49.  Of those, so far the project has been able to process the materials for 5 of the 13 trials and make them available online.  The documents include all trial exhibits, source materials from which most of the trial exhibits were selected, and the full day-to-day proceedings for each trial as recorded in the trial transcripts.  Also included are several hundred annotated photographs taken of the trial proceedings.

hlsl-nuremberg-document-4430-page-1

The Harvard Nuremberg collection is one of the few comprehensive document sets available for these trials.  While many institutions possess materials relating to the International Military Tribunal, relatively few have significant archives relating to any or all of the follow-on 12 National Military Tribunals adjudicated by the United States.  Aside from the U.S. National Archives, Harvard is unique in offering access to the full set of materials generated by all 13 Nuremberg trials.

The new website replaces our previous online presence dating from 2003.  Its beautiful, elegantly functional design was created by Frances Duncan and implemented by Emma Cushman, whose open-source project code is available at GitHub.  It offers through its new design deep, faceted document and photograph search and full-text keyword transcript search.  Document page images are viewable at a variety of zoom levels, and the transcripts are rendered as both plain text and scanned page images.  All document and transcript page images and transcript full text are downloadable and printable.  The website also offers rich introductory materials to the subject matter of the Trials and document archive as well as detailed introductions to each trial, supporting our goal of making the site useful for the general public as well as researchers in the field.

The Project and website are profiled in the Harvard Gazette.

The Web project is a multi-year, multi-phase, on-going initiative.  We have completed the digitization work of scanning all the materials into digital format, have completed the conversion of a third of the transcripts into full-text searchable format, and have finished the analytical work of describing in detail the documents for 5 of the 13 trials.  We are actively pursuing outside funding to support expanding the exposure of this unique collection to the world in a fully open and accessible way.

Ars Electronica Highlights 3

I'm sharing more highlights from this year's Ars Electronica Festival. See parts one and two for even more.

The future of the lab — I saw Ivan Poupyrev talk about about the future of labs. He said a whole bunch of interesting things, but the thing stuck most with me is his advice on staffing a lab. He said something like "Bring in people that are focused on solving a problem. If they have a project you want to support and grow, bring them in and give them space to build that thing." Totally sold on this idea. Projects often need the stamina and focus of a single person (two people feels good too) to jam it through to success.

Jllr by Benjamin Maus, Prokop Bartoníček — A beautiful, relaxing, rock sorting machine. An instrument floats over the top of a bed of rocks looking at each one. After examination, it picks up a rock and moves it to a place in a grid of rocks sorted by geological age.

Parasitic Symbiotic by Ann-Katrin Krenz — A machine that draws on trees. 😃🍃

Running Cola is Africa by Masao Kohmura, Fujino, Kouji, and Computer Technique Group — A classic piece of computer art from 1968. An algorithmic creation of frames starting with a running person, route through a bottle of cola, and end in the shape of the African continent.

Versioning in the Caselaw Access Project

We have a data management dilemma, and we hope that you — data-smart people of the world — can help us out. We need a versioning and change tracking system for around 50 million XML files, and no existing solutions seem to fit.

About The Project

The Caselaw Access Project or CAP, previously known as Free The Law, is making all U.S. case law freely accessible online. For more information, see our project page, and this New York Times article.

Our Tracking Task

Like most digitization projects, we generate many page images. The binary image files rarely change and are not difficult to track. However, in addition to images, we create rich XML files containing descriptive/structural metadata and OCR. As we uncover mistakes in the OCR, encounter metadata anomalies, and gather new data through CAP-facilitated research projects, we will need to update these files. Tracking those changes is going to be a bit more difficult.

The Files

We are scanning about 37,000 volumes. Each volume contains multiple pages (obviously) and multiple cases. Usually, a case takes up a few pages, but some cases are so small that several can fit on one page, so there's no direct parent/child relationship between them. Cases never span volumes.

If you're interested in checking out a case for yourself, you can grab a sample case with all the associated files here.

How we split these things up into files:

For each volume:

  • One METS XML file with all volume-level metadata (~ 1 MB avg)

For each page side:

  • One lossless jp2 (~2.5 MB avg)
  • One 1-bit tiff (~60 KB avg)
  • One ALTO v3 XML file (~75 KB avg)

For each case:

  • One METS XML file, which includes the text of each case body, and all case-level metadata (~75 KB avg)

The Scale

  • Roughly 37k volumes, so about 37,000 volume XML files
  • Roughly 40mil page-sides, so that many jp2s, tiffs, and ALTO XML files
  • A bit fewer than 10 million Cases, so that many Case METS XML files

Our key requirements:

Data Set Versioning

Ideally this could be done at the corpus or series level (described below.) This would be useful to researchers working with larger sets of data.

Sanitizable Change Tracking

As is the case with most change-tracking systems, when recording changes, we usually want to be able to ascertain the state of the data before the change, whether this is by recording the old version and the new version, or the delta between the two versions. However, with some change types, we do require the ability to either delete the delta or the old data state. Ideally, we would be able to do this without removing the entire change history for the file.

File Authentication

People should be able to check if the version of the file they have is, or ever has been in our repository.

Open Data Format

Even if the change/versioning data isn't natively stored in an easily human-readable format, it must at least be exportable into a useful open format. No strictly proprietary solutions.

Access Control

We have to be able to control access to this data.

Our Wish List

  • FOSS (Free Open Source Software) Based Solution
  • Diffing — allow downstream databases to fetch deltas between their current version and the latest
  • Minimal system management overhead
  • Ability to efficiently distribute change history with the data, ideally in a human-readable format
  • XML-aware change tracking, so changes can be applied to XML elements with the same identifiers and content, in different files
  • Will automatically detect replacement images

What we've considered, and their disadvantages

Git

  • Dataset is much too large to store in a single repository
  • Non-plain-text change history
  • Redacting a single file requires rewriting large portions of the tree

Media Wiki

  • Not geared to handle XML data
  • Would require storing in a different format/syncing
  • Non-plain-text change history
  • Provides sanitizable change tracking but no versioning of larger data sets

BitKeeper

  • Non-plain-text change history
  • Seems to not allow easy sanitization of change history

Dat

  • P2P Architecture doesn't give us enough access control for the first phase of the project.

Something we write ourselves

  • Reinvents the wheel, at least in part
  • Probably not as efficient as more mature tools

Should the data be restructured?

Currently, the repository is fairly flat with each volume in its own directory, but no other hierarchy.

Files could be partitioned by "series." A series is a numbered sequence of volumes from a particular court, such as the Massachusetts Reporter of Decisions. The largest series so far contains approximately 1k volumes, 750k pages, and 215k cases, but they are rather inconsistently sized, with the smallest containing only one volume, and the average containing 71. There are 635 series in total.

Many data consumers will want only case files, and not per-page or per-volume files. It may make sense to store case XML files and non-case-XML files in separate repositories.

What We Need From You

Ideas. We want to make sure that we get this right the first time. If you have insight into solving problems like this, we'd love to hear from you.

Next Steps

Please reach out to us at lil@law.harvard.edu.

Ars Electronica Highlights 2

I'm sharing more highlights from this year's Ars Electronica Festival. See part one for more highlights.

Animated restroom sign — The men's bathroom had the best sign ever! I've never been so delighted by a bathroom sign. Playful and fun use of a projector and an animation.

Interface I by Ralf Baecker — A red horizontal line is adjusted up and down the vertical axis to make a fluid line graph irl. The horizontal line seems to be controlled by hundreds of little motors moving thin clear cable up and down. Gorgeously lit and placed in a large dark room.

Single Stroke Structures by Takahiro Hasegawa and Yasuaki Kakehi — make temporary structures (maybe even phone booths!?!) out of strategically crimped, inflated plastic tubing. How amazing would it be to keep a shed in your backpack?

img_6718

img_6717

Highlight by Jussi Ängeslavä, Michael Burk, Iohanna Nicenboim — direct light using a lampshade. 3D printing allows for the matching of shade with the room — direct light where you want it.

Photosynth printing — print on plants. Mask the leaf and expose the rest to light that causes alters the photosynth process.

Ars Electronica Festival Highlights

I'm so insanely happy to be at the 2016 edition of the Ars Electronica Festival. I've wanted to attend for a long time and this year things came together. The festival is as good as I expected.

The scale is large — seemingly endless talks, workshops, and exhibits spread throughout the city of Linz (Austria). I won't attempt any type of comprehensive overview but will share my personal highlights.

IMG_6421

Portrait on the fly by Christa Sommerer and Laurent Mignonneau is an interactive piece — stand in front of it (a monitor with a camera on top) and you see an outline of yourself. You quickly realize the outline is made of buzzing flies. The monitor/camera is surrounded by fantastic printed portraits generated by the installation.

IMG_6504

Body Pressure (this is a placeholder name until I track down the correct name) — lay down on a large deflated bag and feel it lift you toward another inflating bag. The two inflated forms come together gently squeezing you between. This piece is beautiful.

Face Cartography by Daniel Boschung — a robot moves a camera around to 600 different vantage points of a subject's face. The photos are stitched together into a tremendously high resolution photo. The shoot takes 20 minutes to produce one portrait — is this a snapshot?

IMG_6726

BitterCoin by Martin Nadal and Cesar Escudero Andaluz — an incredibly slow but deeply charming bitcoin miner made from an old calculator.

Loopers by Yasuaki Kakehi and Michinari Kono — 12 magnetic worms crawl back and forth to create rhythmic clicks.

Private Talking Spaces Progress

We've been working hard on our Private Talking Spaces effort. Lots of thinking about how to create equity in shared spaces in libraries. Lots of thinking about how to increase focus for folks talking on phones (and maintain the focus of those around them not talking on phones). Lots of thinking about where private talking spaces might be located in libraries. So much thinking!!

Much of the labor has been contributed by our collaborator, Nic Schumann, at Work-Shop. Part of the Private Talking Spaces team — Anastasia Aizman, Matt Phillips, Ben Steinberg, and Tiff Tseng — visited Nic in Providence this week.

We started in Work-Shop's building and had the opportunity to see what remained of the recent [Uncommissioned show].(http://uncommissioned.thedesignoffice.org/)

IMG_5137

We then played with Work-Shop's phone booth. Fantastic!!

IMG_5163

A quick walk down the street led us to RISD's Fleet Library. I'm still filled with inspiration a day later.

IMG_5144

So many thoughtful spaces in the Fleet Library including these little study cubbies.

IMG_5217

We wrapped up the day by taking the scenic route to the train station. ❤️ Providence.

Summer Fellows Share, Join Us

LIL fellows are wrapping up their terms this week! Please join us for and learn from our Fellows as they present their research involving ways we can explore and utilize technology to preserve, prepare, and present information for the common good.

Over 12 weeks, the Fellows produced everything from book chapters, web applications, and board games ­ and, ultimately, an immeasurable amount of inspiration that extends far beyond the walls of Langdell.  They explored subjects such as text data modeling, web archiving, opening legal data, makerspaces, and preserving local memory in places disrupted by disaster.

Please RSVP to Gail Harris

Our fellows will be sharing their work these fascinating topics on Wednesday, August 24 from 1:00-3:00 in Casperon Room.

fellows-poster