The workshop bought out many fascinating aspects of Libyan archaeology. As well as contributions from Libyan colleagues, presenters including Paul Bennett, Will Wootton from KCL, and Chris Blandford, who is preparing a World Heritage Management Plan for Cyrene. Also from KCL, Charlotte Roueche and Hafed Walda launched the Inscriptions of Roman Tripolitania and Inscriptions of Roman Cyrenaica websites. These a digital repositories of the corpora of inscriptions, gathered in Libya the 1950s by Joyce Reynolds, with the inscriptions marked up in EpiDoc XML What is most significant about these is that by using EpiDoc and describing the information within them using this standard, they provide the core for future work that others can build on in the digital sphere. It strikes me – a notion only reinforced by conversations with others here, including Charlotte, Hafed and Will – that the majority of the challenges will be non-technical. In the humanities we see publication as the final, finished thing (a view constantly reinforced by those in charge of pay, promotion and tenure), yet one of the most interesting aspects of the kind of work we do lies somewhere in between doing interpretive research and publishing it. How do we represent the interpretive process in the publication, while preserving its integrity and authority? This question is absolutely fundamental if we to get meaningful knowledge by applying computational approaches in the humanities. And it seems rather beautifully ironic that it should arise so prominently and urgently in the field of epigraphy – inscriptions would seem to be the ultimate ‘final publication’, but hey are not: walking round Leptis and Sabratha for the first time bought home the re-use and re-contextialization that an inscription can undergo before it comes to our attention, but before it was ‘published’ – it can be altered, erased, the block it is inscribed on used for something else entirely and so on. And the use of EpiDoc shows that we can continue building knowledge about them and their associations long after they have been deciphered, translated and pinpointed on the map.