andrew said:

A brief explanation here of the historical edition numbering of PG texts. Every text was released initially in a version "10" (Think of that as 1.0) And then subsequent "editions" would be numbered 11, 12, etc. If you look hard enough in the pre-10,000 files, you can find a couple of exceptions, but that will cover most cases. Also note that the consensus that emerged was that a small number of minor corrections could be made without increasing the edition number.

it is this last "consensus that emerged" that is most troublesome. as long as the filename stays unique for each different version, at least people can depend on the name to identify the version. when you substitute in a different file without changing its name, you've introduced unnecessary ambiguity into the situation, and thus made it extremely difficult for people to keep track of things. and when you fail to provide changelogs on the entire process, the difficulty-factor starts to climb into the "impossible" range... apologies for spoiling your fun... -bowerbird