Our initial thoughts on receiving author’s XML for this paper was that it would run smoothly through the production process. As we are introducing XML editing to reduce publication times and to generate multiple outputs (print, HTML), the first impression was that receiving XML instead of having to convert from a Word file would be a great advantage.
Since the authors had also provided a transformation of the article into the RSC’s own DTD, how could we lose?

Once we started to work on the file, we realised our options were somewhat more limited. The main problem we had was to preserve the integrity of the original source XML - which was required as supplementary data to demonstrate the possible transformations. If we had used the 'RSC' XML then we would have been unable to view it within Internet Explorer using the XSL stylesheet.

Although it would have been possible to develop a seamless transform of the paper from one schema to another and back, it would have taken a huge amount of effort. It became clear that the document design requirements of DTDs for authoring and production were quite different, and these would have made it difficult to transform the article easily from one schema to another. The authors quite sensibly use citation IDs that meant something; as article handlers we wanted incremental counts. Our different methods of adding non-ASCII characters and markup also showed the divergence of ideas.

We therefore used the original XML for editing and supplied an ASCII text file to the typesetters. Proof corrections were applied separately to the XML file and the typeset file.

It was an interesting exercise, but it wasn't easy to deal with by any means.

Richard Kidd
Karsten Harlow
15 March 2001