From: Paul Lotz
Sent: Thursday, January 17, 2019 12:09 PM
To: Tiago Ribeiro <TRibeiro@lsst.org>
Cc: Andy Clements <AClements@lsst.org>
Subject: RE: other TPC issues we should go over
I am summarizing our discussion, in part to inform Andy.
TPC-148: OSL has embraced creating a UML model and has done a great job with this. It offers many advantages for us and we want to keep it. The only issue is that the script to map to XML in its present form does not handle units. This is probably not difficult to add, but has not had high priority, and OSL has been adding the units to the XML via other means. It would be very good to resolve this issue but it may not happen.
TPC-147 and TPC-69: These are overlapping issues and affect which data a late joiner sees. This implementation was promised way back when we were working on LSE-70 and is captured there. This has been implemented recently. I requested that Dave update and resolve the issues (which he has now done).
TPC-39 This is not strictly necessary, but all the LSST stakeholders canvassed so far agree this is a good idea to implement. There are two outstanding items for this:
1) modify the TMA interface to add this (Dave agreed we should do this, but understandably did not want to make a change request at that phase of the contract. This may be a change LSST makes after TMA delivery. That timing has its own challenges, though, and an appropriate strategy needs to be agreed.)
2) Determine which component will generate the trackID (TCS? Pointing component? OCS?) and create a change request to modify the requirement(s) and interface(s) accordingly. (Possible implementation: At DCT we concatenated a timestamp and a random number to create a trackID, so there was an increasing part but the values were guaranteed to be unique. The most important aspect is that the values are unique.)