The overall conclusion of the workshop was that 2 separate proposals will be developed. One will focus on the language elements required for the actual "model composition" (but brings also additional, initially unforeseen, functionality - such as encoding parameter sets and multiple related simulations). The other will focus on the mechanics of simply assembling SBML from bits and pieces (linking in multiple files and/or fragments thereof). It is expected that the first proposal will be a bonafide L3 extension, wehereas the second will be (part of) L3 core enhancement(s). Links to the proposals are providede below, followed by the original materials/details/agenda, etc. of the workshop.