[DGD] lambda operator re-re-visited?

[Dread Quixadhal]

If closures use a unique syntax, which they do in other LPC drivers, you
know, at the point of compilation of the object they are defined in, that
they are closures.

Is it impossible to mark them as such and put the expression source in the
object data?  For that matter, I don't see why the full source of all
objects can't be stored inside their object data.

The crux of the issue seems to be trying to support systems where source
code doesn't actually exist anywhere on disk, so that once you've compiled
an object it's "lost" forever.  If that's the case, it seems like a pretty
solid reason to store the original source that generated the object inside
the object itself.

Maybe this would be a good driver option, since if you are still running in
a small footprint environment, you may not want to waste the RAM to do
this.  Or, if you have the idea of distributing state dumps without source,
you might also not want this... and in those cases, perhaps closures
wouldn't be an option you can use.

As a side effect of having source in each object, you'd also be able to
easily modify any object's code without needing an underlying disk file,
and each object could report errors with full context, regardless of the
disk file being altered, or absent.

I would like to hear from people who use state dump based systems though,
as while I see it as a benefit, perhaps they would have other issues.