One of the key points about I.Sicily is that it will be continuously updated – it is not a one-time publication, but an ongoing project. When something new is discovered about a text, or we manage to study it directly, the edition will be revised and updated. But that presents a very particular problem: how to publish or cite something that is not stable?

You can always cite the I.Sicily number and go to the online edition at its webpage (e.g. http://sicily.classics.ox.ac.uk/inscription/ISic003380), but that will always present you with the latest version. That’s fine if all you want is unambiguously to reference that inscription. But what if you were making an argument which relied on the reading of the text, and in between your first and second visits to the I.Sicily edition’s webpage we visited the stone and decided the reading is different? If you simply cited the I.Sicily page URL, your future readers would think you had made a mistake and be unable to work out why.

There is an important conceptual difference here, between the identifier for a text (e.g. the I.Sicily or Trismegistos number, intended simply as a unique identifier for the individual inscription in the abstract – and in the case of I.Sicily resolving to the latest edition) and a specific edition of the text (think of any paper publication). This distinction is already quite blurred in epigraphy, and scholars frequently use edition numbers, such as the inscription’s publication in CIL or the report of a text in SEG, more as an abstract identifier than to make any specific assertion about that particular edition. The entirely understandable habit of saying, e.g., CIL 10.7133 = AE 1989.342g = ISic000413 already begins to do this, since what this is saying is that all three of these refer to the same inscription; what it does not (should not) mean is that all three of these are actually equal or identical as editions or publications of that inscription. This distinction has been further blurred by the development of the major text databases such as EDH, EDCS or EDR, which, just like I.Sicily or Trismegistos, or indeed any other epigraphic publication, assign the inscription their own unique identifier. However, because these large text databases are primarily aggregators of existing editions, it has become increasingly convenient simply to cite one or other of these database numbers as a proxy for the inscription; but at the same time, in many cases, what is actually being cited is the text, using the edition reported by that database but without actually considering which edition that might be, let alone acknowledging it. Even more worryingly, perhaps, citation of the texts in these databases is done without ever normally acknowledging the database’s creators, or the individual(s) who prepared that particular dataset. Mind you, that already happens most of the time in our citation of texts from CIL or IG – how often does anyone explicitly acknowledge the author of the corpus? (This touches on a whole separate problem, for another day – the attribution of proper credit for digital publications.)

In those dim and distant days before Coronavirus, in February 2020, I.Sicily attended the epigraphy.info IV meeting in Hamburg. We gave a brief presentation there of a solution we are adopting for this problem. Then the pandemic happened, and it’s taken time actually to implement it – but thanks to James Chartrand at Open Sky Solutions, we’re now fully operational.

Commonly, when citing a web-page, one includes the date of access. That offers a very limited ‘defence’ against the problem of change, post-citation, which we noted above. A more robust method (but not universally endorsed, because of debates around copyright and intellectual property) is to archive a copy of the page for one’s self, prior to citation, using resources such as the Wayback Machine or the WebCite service (the latter however is no longer accepting new deposits). I.Sicily instead now does this for you, generating citable, stable and permanent copies, with a DOI, in the Zenodo open access repository. Moreover, the copy of the EpiDoc XML file in Zenodo is accompanied by a human-readable PDF copy (including an image where available), making this a fully usable copy for any researcher. To return to the example of ISic003380 with which we began, the current edition is archived at: http://doi.org/10.5281/zenodo.4358476.

The DOI is in turn embedded in the I.Sicily file, and displayed on the edition’s I.Sicily webpage, to make citation easy:

A new copy, with a new DOI, will be uploaded into Zenodo each time the file is significantly revised. The list of DOIs and dates of upload is recorded within each I.Sicily file, so it is always possible also to trace back the previous editions.

We do not suggest that you should always cite an I.Sicily edition via the Zenodo DOI (although if you do, we would ask you to cite it, as suggested above, as a full publication, recognising the authors involved). But, if you are citing the I.Sicily page as an edition, rather than simply to identify the inscription, we would strongly recommend that you do so.