Re: persistent data feature

[Lars Ingebrigtsen]

Richard Stallman <rms@gnu.org> writes:

> If we provide a feature for storing persistent Lisp data, I contend
> that the default option should be to store it in a purely textual
> file.
>
> Many users will not store large amounts of data, and a textual file
> would be much less risk-prone than any other method.

There will, of course, be large amounts of data, and that will be slow.
(People already do this with a number things, and the results are
predictably sluggish.)

> If we decide also to implement the feature of sqlite3 database access,
> we could offer the option of storing persistent Lisp data in such a
> database.  Users should be able to choose one storage format or the
> other for each database.

You'll get a slower and messier interface.  If you're (optionally)
storing the data in different data files, which ones has precedence?
Will all these backends have the same semantics wrt. concurrent updates
from different Emacs instances?  Will they offer the same introspection
opportunities (like "delete all stored data that refer to 'foo'"?)

Offering a number of backend storage options is just needlessly
complicating things, and we'll end up with a big non-performant mess, as
usual in Emacs.

For this feature to be attractive, there's a bunch of performance issues
and features that have to be present, and they will be in my
implementation, based on storing using sqlite3.  I won't be adding any
other backends for this.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no