Sir

It is becoming increasingly common for articles in top-tier journals to be accompanied by online supplementary text, figures, graphs or other materials. These supplementary materials typically provide either additional evidence for an article's claim or a detailed description of the methods used to carry out the study. Unfortunately, the format in which this material is stored is not always optimal.

Nature's policy is essentially “anything goes”, accepting supplementary files in a wide array of formats, and publishing them as provided by the author (see http://www.nature.com/nature/submit/finalsubmission/SI/index.html). Nature's preferred format for text, tables and images is the Microsoft Word format (.doc), which seems to be the natural choice, given the widespread use of Microsoft Word by scientists.

Unfortunately, .doc is particularly ill-suited for archival and online-publishing purposes. Whether a particular .doc file can be opened and printed successfully depends on the exact version of Microsoft Word installed, the version of the operating system installed, the printer installed, and the fonts installed. Furthermore, the details of the .doc format are secret and change from version to version. As a result, some of Nature's readers will have problems opening and printing supplementary material. Moreover, we should expect that many of these documents will fail to open properly 10 to 20 years from now.

I believe that by allowing publication of .doc files alone, Nature does a disservice both to its authors and to its present and future readers. At the minimum, Nature should convert .doc files into portable document format (.pdf) — which has fully published specifications (see http://partners.adobe.com/asn/tech/pdf/specifications.jsp) and is suitable for online publishing and archiving — and publish the .pdf files alongside or instead of the .doc files. Similar procedures have been adopted by Science and by the Proceedings of the National Academy of Sciences.