Shakespearean HackFest


On Saturday, I attended the Shakespearean HackFest organised by HiCor / TORCH at Oxford University.

I'd like to discuss the hack our team created and how the event differed (in a positive way) from other hack days I've been to.

The Event

On arrival at the Radcliffe Observatory Quarter, I was automatically assigned to a team. I think this is a great model for hackdays. To often, I find, anti-social nerds struggle to form teams - or they team up with people they came with - or non-hackers are left adrift without any technical help. The organisers had made sure that each team of four had at least one hacker, with the rest being made up of linguists, scholars, Shakespearean experts etc.

The first half of the day was a lecture! This isn't something I've experienced at a hackday before. Usually there an intro talk, someone thanks the sponsors, then everybody get cracking! Instead, we heard entertaining talks from Pip Willcox and Dr Emma Smith about the history of Shakespeare's First Folio - how it had arrived at Oxford, been lost, returned, and then digitised via crowd funding. Knowing the incredible provenance helped me feel closer to the data we would be working on.

We then had James Cummings talk us through the XML versions of the text, carefully explaining the somewhat esoteric TEI-C structure. He also gave us a huge data dump. We were given direct access to 2GB of high quality scans of the folio, a zip of all the XML they had available, and direct links to other useful bits of data. We were even encouraged to submit changes via GitHub. Much better organised than some hackdays where the advice is "just Google it."

Rather than getting lost in a vast sea of data, we were all told to work on a single play - Hamlet. Again, this was a sensible decision on the part of the organisers. I think we probably would have struggled if we were left to choose our own play.

Because of the layout of the rooms, there wasn't any real collaboration between the teams. With only a few hours in which to hack, this wasn't a serious problem, but did reduce those little sparks of creative cross-fertilisation you often get at these events.

Of particular interest to me was the gender ratio. Hackdays are usually male dominated affairs. This was decidedly the opposite with a F:M ratio of about 2:1! I don't know if the organisers did anything specific to achieve this, or whether it's a natural consequence of a "Humanities" event. Either way, it's a solid reminder to other event organisers that it can be done!

Our Hack - Too Many "To Be"s

Our team included... Sarah Ellis - the Head of Digital Development at Royal Shakespeare Company, Thomas Jo Johansen, Dr Erin Sullivan a Lecturer and Fellow at Birmingham's Shakespeare Institute, Michelle Morton also of the RSC, and me.

As with all good hacks, we went down several dead-ends before setting on something workable. Here are the ideas, in case you want to try them yourself!

  • Sentiment analysis of the text. Rejected because the corpus contains the original archaic spellings - e.g. "loue" for "love".
  • Multimedia collage. Can we extract tweets, videos, images, animated gifs, etc to match each line of the text? We couldn't find an automated way to do this, and our manual efforts were a bit disappointing.
  • Correcting transcription errors in the text. A bit dull - but the data are on GitHub if you want to do it!
  • How much time do characters stay on stage? The text has metadata to indicate stage directions, entrances, and exits - but it rarely indicates who has left. This could make for a nice animated visualisation and would also help improve the XML.
  • How does this 1st Folio scan differ from others? Can we average out the scans to see what unique features each one has?

Finally, we settled on a somewhat avant-garde sound piece. We were discussing placing links in the XML pointing directly at sound or video files - for example Richard Burton's Hamlet. The problem is, most recordings are taken from the modern text. And most modern products inevitably cut some of the text.

One thing we could be sure of, is that no-one in their right mind would cut "To Be Or Not To Be".

What are the different ways of performing that famous soliloquy? Is Richard Burton's performance notably different from, say, David Tennant's? Who is the best actor for the role? How many different ways can it be interpreted? What if you played every recording to "To Be" all at once?

The result? Noise! Glorious noise! Magnificent noise!

To Be website

We ripped the audio of Burton, Tennant, William Shatner, amateurs on YouTube, women singing with a ukulele, and a dozen more. We trimmed the audio in Audacity, and used HTML5's audio element to play them all simultaneously. In the browser we were able to mix them - changing the volume on the fly, fading in and out, starting and stopping the performances.

Listening to several speakers at once is incredibly disorienting. As you start picking out one or two voices, you begin to understand the wild variance with which people perform. Take a listen to Richard Burton, David Tennant, and Kenneth Brannagh performing at the same time.

Ok, that's three classically trained actors wringing every last drop of melancholy out of it. Even still, fascinating to hear how they approach it. Let's now hear it sung by Courtney Welbon and performed by Paul Scofield.

For students coming to Shakespeare for the first time, it can often feel like there is only one way to perform Shakespeare. That it's a fixed point from which there can be no deviation. Nothing could be further from the truth! We can play it as a rap, as a Greek Chorus, as a man, a woman, or any variation that pleases us.

It also calls into question our notion of "perfection". Many of the other hacks during the event were quantitative in nature - breaking down the text, extracting semantic meaning, adding metadata, analysing the work. Our is, I hope you'll agree, qualitative. Which is the "best" Hamlet? What makes the perfect performance? After listening to dozens of recording, I'm fairly confident I can tell a good performance from bad - but when they're mixed together it becomes less clear. Perhaps I like the sung "slings and arrows" and the spoken "outrageous fortune".

We can take this further. Splicing in recordings so that Burton's 1964 Hamlet is performing against Julia Stiles in 2000's film of Hamlet.

Taking It Further

Putting aside the tricky issue of copyright, everyone felt that it was a unique and interesting way to examine the works. For schools, it allows you to show that it isn't just dusty old men wandering around a stage. For actors, it's a resource to inspire you to take a performance in a different direction.

The Shakespearean HackFest was a hugely entertaining and a great success. Thanks to Kerri Russell and all the organisers for an amazing day and a fascinating set of data.

You can access the digital edition of the Bodleian's First Folio - and edit it on GitHub.

Leave a Reply

Your email address will not be published. Required fields are marked *