In a special guest post, Antony Gordon, webmaster of IAML (UK & Irl), shares his experiences of a recent indexing project for the branch’s long-running journal, Brio. Antony writes:
Having recently compiled name and title indexes to the content of Brio, it was suggested that I might outline the process for the blog. Thus, what will surely be the most technically boring blog of all.
I’ve long wondered whether it would be a useful idea to make Brio content more easily accessible to members, but the time involved in creating any kind of index was always rather daunting. More recently I decided just to give it a go and see how things progressed, with no firm idea of any final result, if indeed there would even be one.
Step one involved scraping the entirety of the Brio list of contents from the website page and pasting the result into a text editor (BBEdit). At this early stage I had to decide whether or not to represent every single entry in the indexes or to concentrate only on articles – including obituaries, however short. The first task then was to search for and selectively remove short miscellanea, entries that referred to reviews, and date-dependent entries.
Step two relied heavily on the use of regular expressions (regex in short) to enable me to manipulate blocks of text into a format of one line per article with appended issue number.
Still relying heavily on regex, steps three onward had to be replicated for each of two now separate indexes. For the author index, first the names needed to be inverted, then brought to the front of the string. For the title index I decided it would be easiest to strip initial articles after bringing the entire title to the front of the string. A brief attempt at title-inversion foundered on a multiplicity of subtitles that added too much unnecessary complication. The time had now arrived for each index to undergo the first of many sorts. The initial results were – not surprisingly – decidedly messy.
From this point on each index went through a process of refinement, of regularizing names that didn’t quite match, as well as fixing titles with subtitles that had messed up removal of articles.
At some point it became clear that the resulting indexes could more easily be viewed in spreadsheet format where mismatches of position and other irregularities tend to stand out. Problems identified were fixed in the text index(es) and the files(s) were re-sorted. Then back to spreadsheet format for another iteration.
What seemed at the time like an almost endless cycle of editing entries to fix inconsistencies, then re-sorting, finally reached its end. The indexes were converted to website table format and added to the Branch website.
From this point on, each index will need to be manually updated whenever new issues are published and their contents lists are added to the website.
Author and title indexes to Brio can be accessed here by logged-in members of the IAML (UK & Irl) Branch.