Hi.
I don't know REPLServer but if you want to implement it I will take part 
in the project.

Regards
Marco

Stephen Pair wrote:

>I don't really have any comment on the documentation itself; however the
>docs got me thinking.
>
>It would be interesting to combine Sprevaylor with REPLServer and a
>minimal image.  I could see deploying an application in this way: the
>deployment would consist of two running images...one to host the
>persistent/evolvable domain, and the other to host the transient/static
>application that manipulates the domain (this might be a morphic gui or
>web UI).
>
>The domain hosting image would be based on some minimal image (to reduce
>memory footprint) and accessed via REPLServer.  It would run in headless
>mode.  Every command that comes through the REPLServer would be logged
>(for the purposes of recovery).  On startup, any pending log entries
>would get replayed (this is the normal Sprevaylor stuff here).
>
>REPLServer could even be modified such that instead of responding with a
>normal print string, it could instead use something like an #xmlPrintOn:
>method.  The application image could then parse the XML formatted
>responses.
>
>- Stephen
>