Dear all, for today's survey result discussion, here is the latest status. We only got a few more answers since the last time. Nothing changed dramatically. Here is my own short summary: In sum, we got 27 answers. Some of them were incomplete. We have an equal number of answers from DRMAA implementers and DRMAA users. The major of people does not see a lack features or documentation preventing the usage of DRMAA in it's current version. TOP3 of DRMAA implementations: 1. Sun Grid Engine 2. Condor 3. LSF library (FedStage) TOP3 of DRMAA language bindings: 1. Python 2. Java 3. C We have balanced voting (important / not important) for most of the specific feature questions. More possible placeholders in job template attributes (issue 2837) are favored by a majority of people. The opinions about JSDL support are also balanced. Nobody wants JSDL and classic job templates both as mandatory parts of the spec. Finer-grained or fewer job states are rejected at all. Some people would like to see an extensible job state model as in OGSA-BES. Some specific issues are declared as important by most participants: More possible placeholders in job templates, monitoring of DRM resources, fetching the list of jobs running in the current session, persistent sessions, and job workflows. Backward compatibility is demanded by a small majority of people. Best regards, Peter.