State of Transcriptions 12 3 2015

##Current state of transcription viewing in Mirador: AoR fork

[http://bookwheel.org/demo/]

This site as the most up to date version of our customization of Mirador.

###Usage (Once the user navigates to the desired page):

• Left side bar will be displayed by default. Under both 'Indices' and 'Annotations' tabs, the rather cryptic 'manifest' label is shown.
• Clicking on this 'manifest' label will toggle a panel on the right side, which contains the annotation data.
• The right side panel with the data can be manipulated in several ways:
  • Its position can be toggled between the right and bottom positions, such that the panel will fill either the full height along the right side, or fill the full width along the bottom. Every time the position is toggled, the panel size is reset.
  • It can be closed/hidden either by clicking the 'manifest' label again, or with the close button
  • The panel can be resized by clicking and dragging the panel border
• Clicking on an annotation in the right panel will highlight it with a red border, but will have no other effect.
• In the top left, there is an icon that looks like a table (spreadsheet, not furniture). Using this button, another window can be opened in Mirador. Any other book can be viewed and manipulated independently from the old workspace.

The right panel is currently floats on top of the image, meaning that it will block part of the image from view.

In the left side panel, there is a list of annotations called 'manifest'. This is the list that contains all of the AoR data. The reason behind this name comes from the origin of the code and original intention of this side panel, which was to display annotations separated according to their source. The name 'manifest' means that the annotations comes from the book data, as opposed to an external annotation server or some other source (in IIIF, the book is represented in an object called the manifest).

###Annotation XML to HTML Transformation

This roughly shows how the XML annotations are translated into HTML to be displayed in Mirador:

• Symbol: symbol_name (LOCATION)
• Marginalia:

bold(transcription as paragraph)

italic([ transcription ])

People: (list of names)

Books: (list of book titles)

Locations: (list of locations)

Code Origin

This transcription work on Mirador is based heavily on the work of Shaun Ellis from Princeton. This past summer, he was tasked with extending Mirador to handle transcriptions. While he progressed fairly far with the side panel display, he was ultimately forced to stop development of transcription handling in Mirador due to other looming deadlines.

Our modifications to Shaun's work consisted mostly of some bug fixes and CSS changes. We added the resize capability, position toggle, close button.

###Some changes that can be made:

• Change the appearance of transcription data, or change how the transcription data appears in Mirador (unbold text, display data in books list above people list, for example)
• Add the rest of the marginalia data to the transcription display (X-refs, internal refs)
• Change label in left side bar from 'manifest' to something more meaningful
• Have main image viewing component resize with the transcription display panel, making the transcription panel a true side panel instead of an overlay. This would mean transcription display would no longer block the image.
• It should be possible, for example, to group our annotations into separate tabs according annotation type or some other criteria. This could be done in either the right or left panels.

###IIIF and Mirador communities There has been some development in dealing with transcriptions in Mirador fairly recently. Shaun Ellis' transcription work was perhaps on of the first attempts at handling transcriptions. His work has been sitting fairly untouched since about August, barring our modifications, as he has been unable to work on it. One of the main developers is working on robust annotation capabilities for Mirador, though the focus is on annotation creation and editing. Still, this work looks promising (We saw a brief demo during one of the Mirador tech calls we attended while we were up at Tufts). This kind of work should be easy to extend to handle our data in the way we want, similar to our modifications to Shaun's work.

There has been several different people all working on different "side panel" widgets for Mirador. There are widgets for annotation viewing, searching, etc. In its current state, there is no standard way of adding these widgets to Mirador, so reconciling these different widgets can be challenging. There has been talk about refactoring parts of the Mirador codebase to make the integration of custom widgets simpler, which would simplify our implementations of the transcription viewing and search widgets. This, however, is still speculative at best with not time lines.