> From: Lassi Kortela <xxxxxx@lassi.io>
> Date: Wednesday, July 10, 2019 5:26 PM

> > [...]

> > Potentially develop a *Persistence* Manager SRFI that breaks
> > absolutely no ground in low level implementation details, concept,
> > and primitives and code we ... borrow from the best.  But is easy
> > to grok and use.

Already renamed it to sound less imposing.  And instead of a RDBMS
mailing list/de facto working group at the level of Schemedoc and
Schemeweb, we could create a Schemepersist one, covering the current
OKVS work (SRFIs 167-8), RDBMSes, this persistence manager concept,
whatever less ambitious things we might try in fsync land, etc.

> This sounds a bit like "fast, safe, easy - pick any two". But major
> bragging rights to anyone who manages to pull it off for any scenario.

Heh.  But the idea is to provide a nice and well documented interface
to existing good things while inventing absolutely nothing underneath
it so the users can pick the trade-offs they need.

> > This has been a interest of mine since I was given a job to make a
> > custom database that just stored data reliable; I'll put it on my
> > long term TODO list, during/after tackling the SQL RDBMS issue.

> Not the easiest job in the world for one person (or even a team!)

Indeed.  Which means ruthlessly limiting the scope; I would say to a
single system, but that doesn't eliminate partition problems.

- Harold