Pulling in the reins on my imagination

Over the past week I've been investigating various technologies that could be applied to scholarly literature and primary source texts. These range from techniques of summarization to information extraction, machine translation to computer understanding, and use every possible linguistic resource, from gigantic tagged corpora for statistical study to meticulously-constructed hierarchies of semantic information for more rule-based grammatical analysis.

At the end of this week, I'm starting to realize that this is the kind of stuff in which professionals and graduate students get lost in for years, only to produce long monographs that conclude with statements along the lines of, "We haven't gotten any closer to what we want to do, but we've learned so much along the way as we have tried to do it."

So I am cutting back on how far my imagination will stretch for the new site. It will still be at earlywritings.com (as opposed to the sourceb.org name that I selected for something even more grandiose), it will at first cover no more than three distinct domains (Christian, Jewish, and Classic Latin), and it will branch out beyond its ranges of date and provenance only slowly, as supported by careful handmade information that expands the frontier and ensures it to be as high quality as what already is covered.

So I will be pulling in the information on the Early Latin Writings from Chris Weimer and combining it with the existing information on Early Jewish Writings and Early Christian Writings, and then adding a dynamic website engine (possibly built on Drupal or another existing CMS technology) with some features that I think are important, and then adding material and features in response to testing and feedback of the new site gradually. I will not try to build a masterpiece from the get-go; that's fine for engineering in the material world, but in the space of information science it is usually a formula for pushed deadlines and fundamental mistakes of chasing rabbits down holes, from which it is hard to recover. Rapid prototyping wins in cyberspace.