Hi Folks:

I've been wondering about something lately...how are people on this list
deploying Squeak applications?

The way I see it you can do it one of two ways:

1) An image per application, with all the changesets for the application
pre-loaded into the image.  If you've got a lot of Squeak apps though,
this can get rather space-hungry (given the size of the image+changeset).
You can go through the shrinking routine, and perhaps modules will help
shrink things down as well.

2) A single image.  You have shell scripts/batch files written up for
each application, that fires up the Squeak VM with the stock image, and
loads a pre-made initialization script into the image on startup (which
then loads whatever is needed for the application).  The obvious problem
with this is if you've got an application that uses a LOT of changesets,
it could take a while to file them all in.

3) A compromise between 1 and 2.   You have a single image.  When you have
a new application, you pre-load the changesets into the image.  Then you
have a set of run-time scripts that activate the right objects for your
application at startup.

Is there a better way to do this?