Our core activity is typesetting mathematical books and journals for academic publishers.
We restructure, enrich and copy edit author sources provided in different formats/variants of TEX or word processors, check style for math, references, clean up and/or generate graphics, etc. We deliver XML/MathML, PDF, rich PDF, HTML, ePub, images for math and whatever else the client requires. All deliverables are generated fully automatically from XML/MathML.
XML–first — A challenge
Publishers rightly demand both PDF and XML from suppliers — PDF for ‘now’, and XML for the future. A problem with XML is that it is very difficult to check for content, because it cannot be easily ‘proofread’ like a PDF file. So it is imperative that when we generate the XML, we can guarantee precise fidelity with the PDF.
Some ten years ago we decided that the only reliable way was to produce the PDF directly and automatically from the XML. This way, checking the PDF would be equivalent to checking the XML. We knew this would be hard, but it proved even harder than we thought, especially with heavy mathematics. But we did it, and now every single page of PDF that we produce, even for the first proof, is generated 100% automatically from the XML file a true ‘XML-first’ workflow.
Unlike XML first, some publishers insist for LATEX first work flow. This is mainly due to the presence of excessive math content which evades the fine typographic conventions of mathematical constructions if typeset from XML.
As LATEX is much more typographically sensitive and mathematicians prefer PDF generated from LATEX directly instead of going through the standard route of LATEX → XML → PDF.