As the letter points out: we need to assess the foreground/systematic effect issues, even when it comes to discuss parameter estimation.
In this respect, in my opinion the best studies up to now have been done in the context of Planck data analysis preparation.
Some of the foreground subtraction techniques developed there have already been applied to WMAP as well. Planck people started to think about these problems more than 10 years ago, and independent groups developed different strategies, which are now being compared.

Furthermore, issues on non-gaussianities, how residual foregrounds may affect the power spectra and which impact they may have on parameter
estimation are under discussion now in that context.
In the lack of an "already done" Planck mission, we can only leverage on "methods developed for Planck and applied to WMAP and simulations".
This is the most solid ground we have at the moment.
Charles Lawrence will certainly know how much of the expertise developed for Planck and not already published in papers may be put in that document.

Similarly, tools to investigate different scanning strategies have been developed within Planck (and I am quite sure also within WMAP).
Let's take advantage of those.
Planck has a whole simulation pipeline for foregrounds at different frequencies that took years and years to develop. Several "experts of each subfield" contributed to create them. Let's take advantage of them as much as possible.

These are just general comments, I am happy to discuss this issue further and help out in the writing. I also anticipate to dedicate more time to this issue during the upcoming year with postdoc and students. 

Best regards,

Elena