Jim, thanks for looking at this again, and for making updates to the page. To respond to your four comments:

1. You are correct that, at the moment, there isn't really another place to discuss background modelling. Anyway, someone reading about photometric measurement would surely wonder how BG is dealt with. I removed the words that suggested sensor-level BG characterization is in any way unusual.

2. I did extract some of the verbiage re: sinc interpolation from the B&L paper. I can ask Steve to have a look.

3. I'm not completely sure I understand what you have in mind, but I moved "Adaptive Moments" to be under the sub-heading Galaxy Photometry. I suspect you are looking for a new section on "Shapes" (of which adaptive moments would be a sub-heading). I'd be glad to see this, but one sub-topic is a little thin for content. Unless of course you'd like to contribute a few paragraphs on the more general topic of shapes??

4. I deleted what I think is the offending text re: Multi-Fit. I didn't delete it entirely because the term seems to have entered the common vernacular when discussing LSST measurement. (And with an incorrect definition, as you explained.) So I'd like to leave the current text just to try to correct the record, even if there will be much more to say at a later time.

I added a table at the end about the image-level flags that affect photometry. Flags do appear in our source catalogs, so some measurement is called for. This table is drawn from the rendering of flags when calexp's are displayed in ds9.

Thanks again for your review.

I'm looking forward to seeing the code-level doc for meas_base, and I'll update this page in the context of a new ticket.

Dick, doesn't a lot or even all of this documentation belong with the code? I'd like the SourceMeasurementTask docs to contain references to all its algorithms, and the whole thing would be machine-linked from the source.

So your descriptions should be moved to the .py or .h files and some way (probably a doxygen group – maybe LSST_measurement_algorithms) invented to link them together. We can then add code to lsstDoxygen to organise the links better than the current page

Robert Lupton we should talk about this, probably in a wider forum. I understand your concern. I'd like to move to a consensus solution in the W15 timeframe.

See, for example, https://jira.lsstcorp.org/browse/DM-1016 (there should be an epic for these)