In this, my third post about the Celtic Poets in North America Digital Humanities project, I’ll describe some of the functionality I’ve designed for the system, and speculate on how it might be extended. Any comments or suggestions would be welcome.
In my last post, I described the characteristics of the three kinds of data. One of the most basic kinds of functions that users will do will be to browse through data after they have set search filters that narrow the data set down to what they are interested in.
For example, show me just:
- Defined locations in Wales
- Male poets who were born before 1800
- Female poets who died after 1900 in Canada
- Poems in any language dealing with war composed by men
- Poems in Irish composed before 1800 in Canada dealing with religion
So, you can see how helpful this kind of system would be, not only for finding information but for analyzing patterns and asking research questions. For example, are there clusters of poems about death that form around well-known war events? How much discussion of war are there in inter-war periods? Leading up to major wars? How does gender and language correlate with the topic of war?
The most basic form of browsing through the results of these search filters will be via textual lists. But lists are always not the most useful form of results, so I’m providing support for Google Maps and Simile Timelines. I’m also hoping to support graphs.
So, to return to the kind of examples given above, imagine being able to ask the system:
- List all of the poems in Breton composed between 1939 and 1944
- Show me a timeline of all of the female poets born in Canada
- Show me a map of all of the poems in Welsh about morality
- Show me a graph which compares the number of poems written in each Celtic language that deal with nature
Only the bare-bones of the system are working at the moment, and I’m not even ready to publish screen shots, but I’d like to anticipate how it might be extended in the future. Such future functionality sometimes must be allowed for in data and software design.
The most obvious extension would be to allow for prose material as well as poetry: expository prose, memorates, historical narratives, folktales, etc. This would necessitate adding some kind of genre tag to items and extending the topic tags, but it probably wouldn’t have too much of a design impact. As there are a great many prose items, though, it would be a much larger commitment of data entry.
Another kind of extension would be dealing with the primary sources themselves. This project only handles the metadata: information which describes the primary texts. To date software support for Celtic language primary texts in the form of parsers, tagged texts, and natural language processing in general seems almost non-existent, and such features of Celtic languages as mutation and noun case make primary texts much more complex than those in English. Until supporting technologies appear in all of the Celtic languages so it would be beyond the capacity of this system to extend into the primary texts themselves.
It has occurred to me that it might be useful to represent and incorporate social organizations which supported literary activity and their events. For example, there have been many Scottish Gaelic organizations around North America, some of which have organized Mòds (annual musical and literary competitions) since 1893. Although the Mòd was essentially imported from Scotland, it was largely modelled on the Welsh Eisteddfod, and these Welsh events were plentiful in the United States and certainly impressed Scottish Gaels in North America (as I’ve explained in an article in eKeltoi). The main complication with this idea is the primary research on these activities is still largely undone and unwritten.
Another important aspect of literary activity that could be modelled and incorporated into the system would be books and journals printed in North America – media that contained and facilitated literary activity… Perhaps this would be the best candidate for extending the system. Just having a list and database of such items would be useful.
Are there other ways in which the system should and/or could be extended?