Dr. John van Wyhe has been on a mission for years now to collect and make available the text of Darwin’s work online. The web site has been up for at least a couple of years, but now apparently has attracted the attention of Nature, so there is now a bit of a buzz. One thing that looks new-ish on the site, or at least I overlooked it before, is the availability of 2nd through 5th editions of the Origin of Species. Something that has been a long-term idea of mine is to produce a variorum of OoS, and now there is at least a resource to get the text of all versions from. There is another site that has a list of Darwin’s correspondence, but last I checked, they simply allow a search on an index, and do not provide the actual text of the correspondence.
So far as I can tell, no one has published an algorithm for production of a variorum version of three or more source works. This is in itself a neat little problem to work out. Certainly, if I do fit that in somewhere, I’ll get in touch with Dr. Wyhe to contribute something back to his project.
CafePress Shop
October 19th, 2006 at 11:08 am
GNU diff3 will take an original text, two derived texts and generate a fourth text combining the changes in the two derived versions. That must be doing something like what you want. Modern distributed version control systems do multiway merging too. However, these things tend to be line-oriented (since they deal with code) rather than word/sentence/paragraph oriented. There are word-oriented versions of diff at http://www.gnu.org/software/wdiff/wdiff.html and http://os.ghalkes.nl/dwdiff.html .
October 19th, 2006 at 12:22 pm
I stand corrected. I will have to look into the diff3 algorithm to see whether it could be extended to arbitrary n derived texts.