Working list of interface ideas

This is a place to collect and explore specific ideas for the WWP interface.

Widgets in the sidebar

I've been thinking about populating the WWO left sidebar with visualization devices that would orient the reader in what they were reading, or show connections between the current text and others in the WWP collection. It occurs to me that if a mandala browser were placed there in the left margin, with the main reading window (i.e. the main frame of the web page) full of text, one could imagine both using the mandala as a way of controlling what shows up in the main reading window, and also as a way of getting another view of that text. One could imagine having two mandalas: one looking "inward" as it were (showing a view just of the contents of the specific text) and one looking "outward" to the entire collection (showing the text's relation to all the others). The inward view would have as its magnets things like words (e.g. show the paragraphs that share a given term, show the poems that share a set of terms, that kind of thing). The outward viw would have as its magnets things like metadata terms (e.g. show the texts that are published on the same date/have the same author/share title words) or shared vocabulary (show the texts that have the same words in them/in the same contexts as the text I'm currently reading and the search terms I'm currently viewing).

Word cloud view of the text you're reading, or of parts of it: perhaps allow the reader to specify the context to be viewed (e.g. whole text, selected chapter/section, selected generic contexts) and then generate a word cloud for just the specified context. Perhaps this could be generated from the Philo word frequency stats for the text?

Exposing structure

Need to think about ways to make structure visible and useful to readers, going beyond the classic "context-sensitive searching"

"Find similar" based on structure: allow reader to select a text chunk (behind the scenes it's an element: a paragraph, a division, a poem, etc.) and identify other chunks that are similar, based on location in the tree ("I'm a paragraph in a novel") or based on contents ("I contain an epigraph followed by line groups"). We would need to think about how to prune the XPaths that we use to identify similarity, behind the scenes, so as to focus in on significant similarities and ignore insignificant ones. We could also allow readers to see and prune an XPath that describes the structural nature of what they're looking at ("I don't care whether it's in a novel or not, but I do care whether this paragraph is the first paragraph")
A simpler initial version of this would be to simply expose the XPath for any selected node; by default, this view should map element/attribute combinations onto human-readable terms, so <div type="chapter"> becomes "chapter", but readers should be allowed to toggle to a raw XML/XPath view.
Ideally, we'd like to treat various versions of the structural context as facets that readers can turn on and off (broadening and narrowing the scope of the "similar" found set), just as they might manipulate the found set through other facets such as publication date or genre.

Palettes

Think of a reading space with a set of accompanying palettes that carry different groups of tools; allow reader to show/hide different palettes based on what they're interested in, what they're doing at the moment

commenting tools
structural exploration (XPath builder/pruner)
word searching/text analysis/word frequencies: highlight a word and find all other instances, displayed in a sidebar; show/hide "similar" word data

The sidebar in which these palettes and their visual results are shown should be expandable to create a two-column workspace, where one column is the "inspector" showing the global/context view, and the other column is the "text" view showing a reading version of a specific selected text

Need to emphasize the fact that the "tool" or "search" view is just a view of the collection: it is affected by altering the parameters of investigation (words of interest, structural properties, etc.) and then seeing the results immediately reflected. No pipeline.

Contextual information

Need to make contextual information appropriately available

Biographical information about people
Reviews/reception items about a text
Exhibits, RWO essays
Geographical information
Comments/annotations

Some contextual information operates at the text level:

Author bio
Related exhibits
Reviews/receiption items
Statistics for this text: e.g. words that only appear in this text; gender balance of persnames; map of non-fictional places named

Some contextual information operates at the local level:

Biographical information on people named in the text (mouseover or reveal on click)
Place information for specific non-fictional placenames

Timelines

Offer a main timeline (preferably zoomable) that shows the selected texts/people/events; this could function as a results viewer for certain kinds of searches. The interface could offer a set of search options (e.g. date ranges, types of items to display, such as texts or people) and the results would be shown on the timeline. Individual items would then be clickable (click on a text to read it; click on a person's name to get information about them, including a list of their texts if relevant). The same interface could also be used to display word search results: the distribution of texts in which the word appears could be shown on the timeline, with height showing the number of hits at each date point. This kind of timeline could be displayed on the same page as the regular search results (e.g. at the top).

Timeline could also be located in the side bar of the main reading interface, showing the position of the text in time.

Search options

Find similar texts (e.g. texts that share a certain percentage of frequent or infrequent words)

Develop a faceted search/browse interface that uses not only WWP metadata but also text properties to limit the collection; it would be especially nice if we could display the proportion of texts represented by each facet graphically (e.g. a pie chart or some other method) so that readers can easily see the effect of choosing a given facet.

Complete pie in the sky: provide a facility to develop and save a facet based on a search (for instance, a word search) that divides the textbase into "texts that contain this word" and those that don't. Could also be texts that contain references to a certain person or set of persons or quotations from a specific text, etc.

Based on user feedback and also from Katie's work on names research (which often required her to find a name in one of the texts so she could see the context in which it appeared): in our current interface, any text that appears in a note can't be found by searching in a browser window. That's because we set the CSS "display" property to "none" until a note anchor is activated by hovering over it with the mouse cursor. It would be great to find a way to make this text searchable from within the browser window, not just via PhiloLogic.

Miscellaneous desiderata

Functional TOCs for texts; maybe redo the TOC encoding to make them more homogeneous, easier to format

Functional cross-references