Write something like T4778 but focused on the "What".
Initial text split off T4778, quoted from chad and epriestley:
Is upstream Phabricator for me?
===========
It's useful to remember that Phabricator is not targeted at consumers or the public at large. It is targeted at professionals in software fields and adjacent fields. Product and prioritization decisions we make in the upstream reflect that.
[[ https://secure.phabricator.com/w/projects/nuance/ | Nuance ]] is our future product intended to let installs interact with common folk (for issues) anywhere on the web.
Local patches and forking are what we're currently recommending to teams who feel the defaults or how Phabricator works for their team or use case are not correct. We are not offended by this, we ship what we call "Vanilla Phabricator" which is set up for a very agile, lean, "open", and likely smaller company and we understand it won't meet every installs needs. But there isn't any reasonable way we can take every installs wishes into the upstream and still be able to move forward/maintain it off of three people.
Tips for proposers
==========================
Even when your proposal certainly isn't unreasonable, you can make it more bite-sized and defined to improve its chances. Smaller proposals are easier to implement and less likely to cause problems, as follows.
Phabricator covers a lot of ground with many different applications. Obviously some things make sense, so they do get added if we feel it's something that many installs will benefit from. In any 'feature request' situation we weigh the likely usefulness across all installs.
Any feature incurs both development time upfront and support/documentation/upkeep after. We are fairly picky about what hits the upstream because it's our time on the line. The number of one-off requests for features/options we get is fairly large.