Minutes of the discussion led by Anita Richards on "Units description and usage", Cambridge, 27-Sep-07, A. Micol Ger van Diepe: Unit conversion should not encompass frame conversion which is a completly different thing. Tom McClean: start simple: I thought the aim was just to build a list of units. Jonathan McDowell: just first solve linear scaling (no wavelength to frequency, etc.) Anita: To keep an eye onto the NIST effort: http://unitsml.nist.gov/ Sebastian Derrier: should consider only quantity transformations that do not imply a change in dimensions. Ger van Diepe: the query language could agree on a specific set of standard/default coordinates Anita: check what has been used already so far. Anita: let's start with use cases. Alex Szalay: pragmatic path: major vo holdings to agree on exchanging metadata using common units, start with few small easy things, and grow later. Francois Ochsenbein: Let's look in the Derrier's comparison list (IAU vs CDs vs NIST), we'll learn a lot. Mark Taylor: applications should do conversions, not DMs: the easy bit is so easy that it is not worth a model, the difficult part would make the model extremly complex. Francois Ochsenbein: SpectrumDM is using some units, VOTable suggest a different set. Main aim is unifying them for interop.