Mathematical typesetting gets XML makeover
May 18, 2009
Siân Harris talks to Barry MacKichan the founder of MacKichan Software about why the company is bringing XML-based processing to its traditional LaTeX-based software tools
Since its creation in 1985, LaTeX has been an extremely popular typesetting system amongst physicists, mathematicians, economists and anyone else who has to deal with complicated mathematical formulae in their papers. “A lot of authors are very keen on LaTeX,” observed Barry MacKichan, founder of MacKichan Software and a former mathematics professor at Duke University. “They often try to do things in [Microsoft] Word, for example, and then get very frustrated and go back to LaTeX.”
LaTeX itself is open source but specialist software companies like MacKichan Software help researchers to use it. For the past 28 years the company in various forms has been developing word-processing tools and since 1988 it has been working on tools based on LaTeX.
Things are changing though: later this year the company hopes to begin the final testing of a new suite of products. These tools will still read LaTeX files and output LaTeX files if the users want them to. Behind the scenes though, the internal processing will be based on XML (extensible markup language), a flexible text format for transporting and storing data on the Web and in other applications.
This is a major move for the small software company; its four programmers have been working on the migration for the past four years and it was a research project for a couple of years before that. MacKichan insists that it is an important step though.
“We needed a more flexible architecture,” he explained. “Our current products were developed in the early nineties and much has changed since then, especially with the centrality of the Internet. Many of the decisions we made initially were because of limitations in the technology such as PCs with just 1 MB of memory. Such initial design constraints were holding us back in developing our software.”
The move to XML should make the products more user friendly, according to MacKichan. “In the existing products, users could start documents very easily without knowing any LaTeX but if they wanted to do more complex things such as changing the margins they would need to know some LaTeX,” he said. In contrast, the new XML-based products will have a WYSISYG (‘what you see is what you get’) interface and users will have much more control over the formatting of their documents without needing to know any LaTeX.
The move to XML also opens up the possibilities for new and enhanced features such as greater “searchability”. In addition, the new MacKichan tools are based on the open-source Mozilla code base, which underpins a wide range of internet applications such as the Firefox Web browser, the Thunderbird email client, both from the Mozilla Foundation. Mozilla open-source code is also used within popular social networking sites like Facebook and Second Life.
There are challenges with this approach though: “The Mozilla code base is huge. By building our product on top of that we get a lot of functionality but there is a lot to learn too,” said MacKichan. Nonetheless, he said that the company has now completed most of the difficult XML features and has proved that the concept works. Now, he believes, it is simply a case of filling in the product details, testing it and ironing out any bugs.
The development will affect all three of the company’s products. Scientific WorkPlace provides mathematical word processing, LaTeX typesetting and computer algebra and this will be the first package to get the XML make-over. This will be followed by MacKichan’s other two products: are Scientific Word, which has the mathematical word processing and LaTeX typesetting; and Scientific Notebook, which includes the mathematical word processing and computer algebra. As these two products are subsets of WorkPlace, the XML migration will simply involve picking the appropriate parts of the parent product, according to MacKichan.
MacKichan reports positive initial reactions from users. “Generally people are interested and supportive. Everybody seems quite eager to see it.”
And once this suite of products is out, what next? “We take a big breath,” joked MacKichan. After that though, there is plenty more to do. “It is very easy to add new features now. We will see what people want us to add the most. This might be exact diagrams, much better support for bibliographies or better integration with other products. We can take advantage of other products based on the Mozilla code base. We will also be producing different extensions for physicists or economists, for example,” he said. “There are many possibilities that we are mostly ignoring in the first release.”
Traditionally, the main applications of the LaTeX system and tools such as those from MacKichan Software have been in preparing scientific papers. LaTeX is still the preferred submission format for many journals with high levels of mathematical content, although XML is increasingly making its mark in this area too.
MacKichan would also like to see his company’s tools used more in teaching too though; people are already using the existing product to write books and lecture material. Users can include interactive features such as multiple choice tests that look different to each student and MacKichan would like to extend such uses. “We’ll have a scripting language built into the product and my hope is that people will use this as a teaching platform.”
About the author