For eb.tbicl.org, the pattern I use for illustrations is minimal. I will give examples from http://eb.tbicl.org/architecture-greek . [illo src="/vol02/4/images/img380a.jpg"] Fig. 17.—Lycian Tomb of Telmessus. [/illo] All that is absolutely required is the filename and the caption - and not even the caption if there is none in the source. The default format is 75% width, centered, but this is configurable. If you want other widths or positions, you can add classes css-style: [illo class="w35 lfloat" src="/vol02/4/images/img379a.jpg"] Fig. 15.—Plan of the Temple of Poseidon at Paestum. [/illo] Notice that some illustrations have credit lines: [illo src="/vol02/4/images/img379b.jpg"] [credit]From a photo by Brogi.[/credit] Fig. 16.—Temple of Poseidon at Paestum. [/illo] Others have an extended key to items in the illustration. [illo class="w35 rfloat" src="/vol02/4/images/img378.jpg"] [credit]From Curtius and Adler’s <i>Olympia</i>, by permission of Behrend & Co.[/credit] Fig. 14.—Plan of the Heraeum. [key] A, Peristyle; B, Pronaos; C, Naos; D, Opisthodomus; E, Base of statue of Hermes. [/key] [/illo] I have enough different tags now that the only html I include is in-line tags for e.g. italics, which work fine in every context I've wanted to display so far. Here is the markup for the page numbers in the right column: [pgnum]380[/pgnum] Note that, for presentation in the browser context, this produces a link to the TIA image for the page. The resulting html is: <a href=" http://www.archive.org/stream/encyclopaediabri02chisrich/#page/379/mode/1up" class="pagenum" target="_blank">379</a> but that's all produced automatically. Footnotes are equally trivial. There's only one in this article, but it's a little tricky because it includes a link to another article for which I use the [qv] tag. Notice that the footnote is available as a popup as well as in a list at the end of the article. The earliest Greek temple of which remains have been discovered[ref]Except, possibly, the earliest of those at [qv]Sparta[/qv] (<i>q.v.</i>).—Ed.[/ref] is that of the Heraeum at Converting from DP output (which is optimized for a full-screen browser, and no separate pages for articles - you get a whole volume) is mostly a matter of removing markup, most of which is automated. I generally post the articles an hour or two after the PG etext is available. I have a control bar with buttons for most tags: http://eb.readingroo.ms/editor-box-with-control-bar/ The most tedious challenge is articles with maths in them because i convert them (the maths, not the whole articles) to latex. My main poiot is that all that's really needed in PG text is accurate content and only enough markup to designate and locate the structural elements, and formatting should be standardized and automated. Don