How do most publishers convert to XML?

[Steerpike]

I've been tasked with converting our journal's articles to NLM XML for archiving in PubMed Central, and Googling around I'm surprised at the lack of tools that seem to be available. Does anyone here happen to know how larger publishers usually go about converting their author submissions to NLM XML? Do they really go through and tag everything manually? Or do they all use proprietary systems developed in-house? Are there commercial products I've missed? Lemon8-XML seems to be the only thing out there that tries to do what I need, but it's still maturing, and I would prefer an offline solution.

[mj]

Hi Steerpike,

As someone who's undertaken conversion and submission of articles into Pubmed Central XML for two journals myself, let me start by wishing you good luck! You have quite an adventure ahead of you.

The lack of XML conversion tools you identify is precisely the ground from which Lemon8-XML originally sprang. I would love to be corrected, but my understanding is that even now, 4 years after my initial work that yielded L8X, the vast majority of XML conversion work is done manually, even among the larger publishers.

Adrian Stanley from the Charlesworth Group was at the first PKP conference, and described in some detail the custom in-house system they use, coupled with a small army of manual typesetters. Julian Fisher of Scholarly Exchange believes that it can be done reliably by outsourcing for a few dollars per page, but given my experience with PMC's stringent review standards, I'm yet to be convinced that the quality of this approach is sufficient. There are a few tools which have been discussed, including eXtyles (one of the most prominent, and among the more expensive, tools), and the recently announced Article Authoring Add-in; both of these integrate with Microsoft Word, and are designed to assist in manual XML markup.

Regarding an "offline" solution, L8X is fully intended to be downloaded and installed on a private server for those who wish to do so; that is, it can be just as much an "offline" solution as any other tool. I may be misunderstanding your rationale here, so please correct me if that's the case.

I should also add that a significant part of the strength of L8X comes from its history of development in collaboration with journals who have gone through the exact conversion and submission process that you're about to undertake; so, we would be more than happy to work hand-in-hand with you to help share our experience, as well as see if and how L8X can help your efforts -- and then feed those lessons back into development so that it can benefit others similarly.

Best,
MJ

[ConversionWhiz]

Besides asking Julian Fischer, you can also ask Peter Suber and the PubMed Central folks who they recommend.