Last updated: 17/12/05 VulSearch: A History
First steps
It's 1997. An insufferably pious lad with little experience of life, but a firm belief that God's native tongue is Latin, has his first chance to 'surf the net'. Of course, my first priority is to download a copy of the Vulgate (boys will be boys!). I remember being rather non-plussed to discover that several hours of downloading had delivered me a text devoid of punctuation and with hardly a capital letter in sight. My five years of Latin at school (which admittedly had ended eighteen months previously, during which time my knowledge had lain unused and decayed rapidly) had failed to prepare me to read things like "rursum summus sacerdos interrogabat eum et dixit ei tu es Christus filius Dei benedicti" without doing a double take. It wasn't long before I swallowed my pride and wrote myself a little program to display this bizarre Vulgate side-by-side with the Douay–Rheims translation—literal to a fault, inelegant and often incomprehensible, but Catholic. And traditional. I don't have a very clear memory of this early program: I don't think it did anything much beyond display the two texts. On the other hand, it must presumably have been the source of the ridiculous name, which has dogged the program ever since.
VulSearch goes public
The story now goes quiet for three years. I very rarely use VulSearch, and it lies forgotten on my hard disk. Finally, in early 2000 I suffer a major computer crash and lose everything on my hard disk. For a reason which now escapes me, I took this as a cue to start again, and re-write VulSearch from scratch. This version definitely did feature searching—the quite brilliant algorithm read in the entire Bible line by line at run-time, searching each line as it went. As you can imagine, this was tremendously effective on the 120 MHz machine I had at the time—on a good day, a search might take less than two minutes. By this time I was at Oxford, and, with a remarkable absence of shame, I decided that I might as well put VulSearch 2.0 onto my college web-site. I have no idea how many poor souls downloaded the thing, but the site had some 2,500 hits in two years.
The Clementine text project
Meanwhile, my worst excesses of naïvety were being slowly corrected by the loon community at Oxford (and sometimes very quickly by the hard knocks of real life). My love for the Roman Church's texts—above all the Missal and the Breviary, which are beyond any shadow of a doubt the pinnacle of pre-modern western culture—was completed by an intellectual engagement, which can only help to make one's emotional attachment less superficial. Somewhere in the midst of this I began to think about versions of the Bible, and the general dissatisfaction I felt with the artificial Stuttgart text used in VulSearch became unsupportable. For all the countless millions of pages of trash on the world wide web, I discovered that one couldn't for love or money find the Clementine Vulgate there (though one could find any number of people complaining that the Clementine Vulgate wasn't available on the internet). It seemed very clear to me that someone had to create an electronic version of the Clementine: it was just impossible that such a thing not exist! It was equally clear that nobody would do it: the task is at once skilled, demanding a basic knowledge of Latin, but also unskilled, since it largely consists of mechanically transferring text; and its scale is almost overwhelming. With a chilling sense of inevitability, I began on the New Testament in the summer of 2002.
A new home on Sourceforge
At the same time, I had just gone down for the last time. My college web-site's days were numbered, and I wanted to find a permanent home for VulSearch and the newly formed Clementine Vulgate project. I have no idea how I discovered Sourceforge, but doing so was a piece of very good fortune, as it has proved a stable and accommodating host. The site provides free services to Open Source projects, that is to say programs that are not only distributed without charge, but whose source code is also freely available so that others can re-use or modify it if they so wish. So my humiliation was complete, as my ill-planned code, full of kludges and inefficiencies, entered the public domain.
The first serious version of VulSearch
I worked on the New Testament through the summer of 2002, then did nothing more until after Christmas. I finally convinced myself that having come this far I may as well at least finish the New Testament, and I decided that the Clementine text deserved a better showcase than the rather embarrassing setting of VulSearch 2. So, VulSearch 3 was born. My number one priority was a decent search facility. I had always shrunk from the idea of writing an indexing program, and eventually it dawned on me that somewhere on Sourceforge there must be people providing such a program for others to use. This was indeed the case, and so VulSearch was married to a cute little thing called Swish‑e. I also needed to include a bookmark tree—I wanted this to collect together all the readings from the Missal and Breviary, though I often wonder whether other people use it at all. In a vague and probably misguided attempt to provide things that might be useful to others, I also added exceptionally primitive cross-referencing and annotation gadgets. With a modest fanfare, VulSearch 3 appeared at Easter 2003, complete with the Clementine text of the New Testament and Psalms.
A long, hard slog
People were excessively generous with their compliments about VulSearch 3, but it is always pleasing to find that one's efforts are of some use and interest. I needed all the encouragement I could get as I braced myself to start the remaining 70% of the Bible. Of course, the text itself is often encouragement enough—discovering a gem like the wonderfully evocative "tenebræ…tam densæ ut palpari queant" can keep one going through days of genealogies. At about this time, there was a marked increase in the correspondence I received on Vulgate-related topics, and a small number of people wrote offering their assistance with proof-reading or other aspects of the project. I even received my first financial donation. In total, three generous users have made donations large and small, to whom my sincere thanks. I merely observe that this might not universally be considered an impressive ratio from 4000 downloads. Perhaps that kind-hearted millionaire with a love of the Vulgate is lurking just around the corner.
Dispiritingly few people persisted with supplying lists of corrections: in almost all cases correspondence faded away within a couple of weeks. Nonetheless, there have been a couple of stalwarts who have very faithfully kept up their support for the project. It was one of these who lobbied for a French version of VulSearch, which led to my producing a version in a noxious Franglais patois, whose numerous errors she very graciously corrected. She even served a brief stint as scribe, creating the electronic versions of Wisdom and part of Job.
VulSearch 4
The autumn of 2003 was a busy time for me, and so a fallow period for the project. Reinvigorated by the temporary pause, I managed to proof-read a fifth of the Bible over Christmas, and started the new year with the determination to create a version of VulSearch that would feel a bit more like a piece of professional modern software. Unfortunately the source code had by this point become completely unmanageable and unmaintainable, so I bit the bullet and started afresh. Everything about the new code is better than what went before, but it's still not the picture of clarity that I'd hoped for. I'm afraid I'm irredeemably undisciplined in these things. The two major new things to add from a programming point-of-view were the ability to work with a variety of Bible texts rather than having two hard coded, and integration with Whitaker's Words (a free Latin dictionary program that I cannot praise highly enough). The first of these was accomplished by a more logical organization of the classes in the program; the second had long been a dream of mine and required delving into the mysteries of Ada to make Words produce more amenable output. I think the purely aesthetic improvements were also important—VulSearch 3 now looks old and tired, and something a bit fresher should be more pleasant to work with. By the same token, I also tried to make the web-site crisper and clearer. The first beta version of VulSearch 4 was released early in 2004, and was quite favourably received. It included over two thirds of the Clementine text.
The end in sight
With the beta released, I could focus my complete attention on the text. In turn, as the text neared completion, proof-reading came more and more to the fore, so that by the end the proof-reading was running almost neck-and-neck with the creation of the text. On the down-side, my youthful optimism was once again dealt a cruel blow when it became obvious that a single round of proof-reading would not suffice to create a text with few enough errors that I could put my name to it without shame, so I began the heavy task of proof reading everything a second time. I'd like to believe that very few errors indeed survive this second reading, but the truth of human nature is that one could reread a text until the trump of doom and still leave errors slip through. I trust to the good will of those who use the texts to forgive any remaining errors, and above all to notify me by email of any mistakes they spot.
By September 2004 a real momentum was building up as the text neared completion: a very industrious new proof-reader came on board, and a couple of dormant checkers returned to the grind-stone along with the loyal stalwarts. With the entire text complete and proof-read, I cleaned up a few loose ends and released the final version of VulSearch 4, as well as the definitive version 3 for the benefit of those without the ridiculously powerful computers needed to run .NET applications at more than a snail's pace.
After a final burst, the second round of proof-reading was finally finished and the first good text (the "Quasimodo text") released on Low Sunday, 3rd April 2005. This was a sombre time to pick, as the Pope had died the previous evening, but I quietly said a Te Deum, and began the wait for reports of remaining errors in the text.
little.mouth@soon.com