TEI Hackathon: Report by Elli Bleeker

Name: Elli Bleeker

Twitter: @ellibleeker

Email: elli.bleeker@uantwerpen.be

 

Working group project: ODD visualization

 

Contribution to group project: in comparison with the other participants, my experience with and knowledge of ODD and JSON was limited. Since it is very useful for anyone working with TEI XML to have at least a basic understanding of these concepts, I was happy to join this working group. At first, my contribution was mainly asking -perhaps obvious- questions about ODD. I’d like to think however that these questions helped to narrow down our goal. At least it was necessary to clearly describe the present state-of-the-art, what is missing, and why an ODD visualizer would be a welcome addition. During the workshop, I assisted with the creation of test data (several project-specific ODD, their transformation to flat ODD and subsequently to JSON); contributed on the design of the treemap and participated in the group discussions.

 

Methodology

The general idea and necessity of an ODD visualization was quickly explained, so the group focused on what exactly needs to be visualized and especially, how. The current ODD visualizer of Byzantium does not show the amount of customization. Moreover, it is useful to see on which specific areas the ODD focuses. We were quick in dividing into three smaller groups, each with specific assigned roles depending on our pre-existing knowledge and capabilities.

We decided upon a D3 zoomable treemap for the visualization of the amount of customization. The treemap shows how elements are grouped as well as –by their size or color- to what extent the elements are customized. The zoomable option could also allow to ‘zoom in’ on the customized elements to see the exact changes made.

 

Final Product

For all the deliverables (Google doc, Github, demo…), see the reports of the other group members.

 

Learning Outcomes

It was a very useful experience for me to work on this project. Although from time to time intimidating, the immediate jump from theory to practice was the best way to learn the concept (ODD, flat ODD, TEI schema, JSON, etc…) and to understand the workflow of the different processes. Having to think about the visualization of something that I was not familiar with helped me a great deal in understanding it. Whether the resulting visualization is as clear and useful for others remains to be tested. Nevertheless it is clear that this could be a necessary tool for users of TEI, regardless of their level of experience.

TEI Hackathon: Report by Raffaele Viglianti

Name: Raffaele Viglianti
twitter: @raffazizzi
The project
We wanted to visualize some aspects of a given ODD specification to aid the creation process, or to help the understanding of someone else’s ODD.
There are some precedent attempts at doing this: Byzantium (http://tei.it.ox.ac.uk/Byzantium/) already visualizes basic information about what your ODD includes or excludes; Gregor Middell’s visualization (http://gregor.middell.net/roma/) already shows relationships between classes and elements in TEI.
We attempted to quantify *how much* an ODD has been customized and show graphically which parts have been curated more and which less.
My contribution
I worked on the architecture of this small app and programmed a scoring system that we previously sketched as a group.
First, the app imports a JSON version of the full TEI P5 ODD; then it imports the compiled ODD of a customization (in JSON).
Second, the ODDs are compared and scoring is calculated for each TEI element. These are the changes affecting the score (in parenthesis there are changes discussed, but not implemented):
* changed element description
* added attribute
* added attribute *value* (e.g. by restricting values for @type) <— this is by far the most common change
* (added elements)
* (added constraint / schematron rule)
* (number of examples provided via <exemplum>)
Finally, the resulting object is passed on to a d3.js component for visualization (Nick worked on that part)
Deliverables
ODD being visualized (prepared by James Cummings): https://github.com/raffazizzi/Hackathon/blob/master/odd/test-odds/james.odd
Findings / goal
ODD is difficult and associate with advanced TEI users. But it needs to be the gateway to TEI and a constant companion to the life of a TEI project. We’ve been trying to figure out whether visualization can make ODD clearer and less daunting. Hard to tell with our simple visualization, but I think we had a consensus in the group that this is the way forward.
Not neglecting the power users, we think that visualizing ODDs can also help understanding other people’s usage of TEI, particularly when operating on a corpus created by someone else (cfr. Syd Bauman’s concept of “blind interchange“).
What did I learn
I learned that there is willingness to make ODD more accessible across the board; that the official JSON conversion from ODD needs improvement; and that underscore.js has some nice functional methods to deal with collections.

tei xsl styles 7.18

Dear colleagues,

I have made a new release of the TEI Stylesheets in the usual places;
if you use Debian packages or the TEI framework updater in oXygen,
you’ll see them come over.

amongst other things, this release

* implements the Pure ODD markup more thoroughly
* fixes many small problems in Excel and Word to TEI (you’ve no idea how devious a docx file can be…), thanks to Magdalena Turska
* improves Markdown <-> TEI
* improves WordPress -> TEI (similar unimaginable horrors)

there remains an unresolved error in some ODD processing
(https://github.com/TEIC/Stylesheets/issues/21)

https://github.com/TEIC/Stylesheets is the place to report
errors, and to send in fixes (as some people are doing,
which is great).

Sebastian Rahtz
Director (Research) of Academic IT
University of Oxford IT Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

Não sou nada.
Nunca serei nada.
Não posso querer ser nada.
À parte isso, tenho em mim todos os sonhos do mundo.