<choice> is the xml element we use to encode alternative spelling in our transcriptions of Blake’s writings. It’s what makes the Blake Archive’s search function forgiving. Say someone searches for all instances of the word “Tiger” in the Blake Archive. A choice tag is what would lead them to instances of the word “Tyger.”
But the more common search tags are the really mundane ones:
- abbreviations (“William” for “Willm”, “January” for “Jany”)
- non-standardized spellings (Blake has a habit of writing “recieved” and not “received”)
- words that are divided by line breaks (the word “accompa / nied” is written over two lines in one letter)
In the letter of 7 June 1825, Blake mentions the “D of C”, which we know to be an abbreviation for the “Dean of Canterbury.” Without a choice tag, users searching “Canterbury” would never be directed to that letter.
- Should we add a choice tag for words and names that have been abbreviated? (Yes.)
- Should we add a choice tag for regularized spelling? (Yes.)
- Okay, but should we regularize spelling to American English or British English if there’s an option? (If this comes up, I’d add a choice tag for both, just in case.) Should we use a choice tag to Americanize British spelling? (No, in part because then we’d have to Britishize American spelling to make things even for all imaginable users, whatever their nationality or dictionary of preference. We’d need choice tags in every line, and we’d need our interns and project assistants to manually add all of them.)
We do have set rules for when and how to use choice tags. But the main principle–the logic behind our rules–is that we have to anticipate the kinds of searches that users might perform, even if that search seems unlikely. We have to anticipate, and correct for, users’ errors. This means we have to think outside any narrow ideas of what needs to be searchable and what doesn’t.