A short history of Anastasia
The origins of Anastasia lie in the work done by Peter Robinson, Norman Blake and Elizabeth Solopova in 1993, in preparing the first publications of the Canterbury Tales Project. At that time, the major SGML (as we then used) publishing tool was Electronic Book Technologies DynaText system. This was a remarkable system, but we quickly ran into significant problems with it. Especially:
We came to realize that these problems were expressions of a single underlying difficulty. DynaText would not let us start processing at any one point of the text, carry on to any other point of the text, and understand and process all the encoding it found between the two points.
This need became the fundamental principle driving the development of Anastasia. The first fumbling experiments towards Anastasia were carried out by student computer programmers at De Montfort University in 1998 and 1999. These were mostly useful for establishing directions we should not follow. The core of the first version of Anastasia itself, which established the basic processing model, was written in three weeks by Peter Robinson in early 2000. With some refinement, this was used in Estelle Stubbs' edition of the Hengwrt Chaucer Digital Facsimile, published later that year.
Andrew West came to work, originally as a student programmer, on Anastasia in June 2000, and wrote most of the scripts for the Hengwrt Chaucer Digital Facsimile. The experience of this publication, and developing knowledge of the program and how it worked, led us to a complete rewrite of the whole program, commencing in early 2001. This took some thirty months of intense effort (compared to the three weeks required for the first version), with Andrew responsible for almost all the final code. In particular, the key memory handling routines, and the interactions with the Apache server environment, were completely rewritten
Anastasia was at first sold by Scholarly Digital Editions. Our thinking was that the people we thought might want to use it, would be prepared to pay a moderate amount for the software, in the expectation that this payment would help us give them support in using it. We now realize that people prefer to pay for support if they need it, and that there are great practical advantages in going open source. We do not have the time to do all that we would like to extend Anastasia (e.g.: prepare versions for other platforms; integrate with Perl, with PHP; run on IIP). If she goes open source, others may do these very desirable things. Also, in the massive rewriting of Anastasia we carried out, we became more and more reliant on open source tools and code. Anastasia relies on open source, especially on the Apache and SGREP source: so it should itself be open source.
Above all, we think she is a wonderfully capable tool, and (usually!) a joy to work with. She has allowed us to make the kinds of publication we used only to dream of; we would like others to have the same fun.