Variable Transcripts
A common aim of projects is to provide more than one version of a source to readers. It has become common to have linked diplomatic and normalised transcriptions of texts, but on-line publication provides the possibility of greater flexibility than this. You could create versions with:
- original spelling, unexpanded abbreviations and original layout for scholars;
- original spelling and layout, but expanded abbreviations for undergraduates;
- modern spelling and additional punctuation for use in schools.
In these circumstances you would want to ensure:
- only one base text was required;
- all versions were created automatically from the base text;
- it was always clear which version was being used or referenced.
We have created two examples of ways of meeting these aims. Both use the same XML encoded source texts and process them using the SAX Application Programming Interface.
The first example uses HTML documents generated off-line using a program written in the open source scripting language Python. It is fast, but the variants are predetermined.
The second example uses a server-side program written in PHP to generate HTML documents according to the reader's choices. It is slower, but provides greater flexibility.
Each HTML document seen by the user produced by these examples includes a reference, recording the source text, application used and version of output generated. In order to ensure that a particular HTML document could be recreated in the future, it would be necessary to preserve:
- a copy of the XML
- a copy of the source code (Python/PHP)
Since Python and PHP are well-documented, general scripting languages, it would not be necessary for a project to preserve the interpreters that run the code - although it would be good practice to record the versions used.