Scan2Run focuses on the digital preservation of computer heritage distributed in paper form (e.g. old magasines with BASIC programs). It may be the only available format Transforming such a listing in a running computer programs and sharing the experience requires quite a few steps: retyping the program, loading it into a vintage computer or emulator, and capturing some results either in textual, image or even video format. Our talk will illustrate our current approach and progress with a toolchain developed for the NAM-IP Computer Museum to help automating the scan of old listing (including learning and reusing profiles), then injecting the result into an emulator with MAME as primary target. Our talk will be illustrated with examples from the widespread Amstrad CPC and rare DAI In-DATA Imagination Machine. It will also be the opportunity to revive and illustrate some capabilities of those nice machines !
More technically, the scan2run project (available from github) supports * OCR with interactive learning phase, forked from JavaOCR, a ligthweight OCR engine * Lua scripts for MAME injection either on a by line or character per frame basis (depending on what the machine can accept as input) * complementary use of screen/video capture and post-processing Open Source tools for documentation and dissemination purposes (Gimp, OBS, OpenShot...)
Speakers: Christophe Ponsard