SQLite Forum

> ... just deploy an instance of say Discourse \[rather\] than build your own forum software...

SQLite is built using a lot of home-grown infrastructure

  *  The Lemon LALR(1) parser generator rather than Yacc/Bison
  *  Fossil rather than Git/Hg
  *  And now the new Forum rather than a prepackaged system like Discourse.

It is a recurring theme that this is bad and that I should be using
third-party "standard" components.  But in my experience, the use of 
home-grown components has been a great boon for SQLite.  Fossil in
particular has been enormously helpful in moving SQLite forward.

Fossil, in addition to being the VCS used by SQLite, is also the main
beta-test project for SQLite.  It is very important that I be closely
involved in the on-going development and support of Fossil because Fossil
uses SQLite extensively, and that puts me in the position of being an
SQLite user instead of an SQLite developer.  When I am in the position
of an SQLite user, I more easily appreciate the pain-points that users
experience, so that the next time I switch back over to the SQLite
developer role, I am more motivated to address those pain-points. Problems
are fixed and new features are added more quickly when they affect Fossil.

So, in other words, I am not wasting time working on the Forum software.
This new Forum is built around SQLite, and so working on it helps me to
stay in touch with the experience of users of SQLite, which is important
for keeping SQLite working smoothly.

A few people report that the transition from legacy mailing-list to a
forum system is annoying¹.  That is not the intent of the transaction,
but rather an undesirable side-effect.  The transition is designed to
help make SQLite a better product, and will (it is hoped) be a net
positive in the long-run.

¹In private communications, far more people have welcomed the
transition than have denegrated it.