Size limit to changes

R. A. Harmon harmonra at
Thu Aug 13 22:31:35 UTC 1998

At 08:39 AM 8/13/98 PDT, you wrote:
>  Although I've never used Smalltalk professionally, I think I do serious
>  programming in it.  In my limited experience, RC seems a tacked on
>  after-thought...
>Revision control is better integrated for teams in the commercial
>Smalltalk products. Look for information on Team/V and Envy.

Thanks for the suggestion.

I think that nearly all programs that are used and maintained over time need
revision control either formal or ad hoc, not just those maintained by more
than one person.  Obviously I do not include those written in APL.  It is of
course a write-only language.

And again, these products, and other free versions, add very useful features
that are difficult to implement.  It seems to me the model of the SDE makes
it much more difficult than need be.  A (poor?) analogy might be programming
in C using one monolithic "code edit" file for all projects, then
partitioning out pieces into various .c and .h files to package them in some
kind C "program".  I do not have nice things to say about the way C
compilers assume source is stored, but it does seem in some ways easier to
manage.  I just think it might be worth stepping back for a moment to
consider whether there isn't a better SDE approach than the source and
changes files.

A future project -- I think it's 3,691th on my list -- is to directly store
the SDE code in RCS with the method being the smallest check in/out unit.
Using a declarative model of Smalltalk initialization code, pool, class and
method definitions would all be stored there.  Damn I'd really like to see
some that declarative model Smalltalk code so I'd know what it looks like.

I may have made a dumb suggestion -- I'm better at quantity than quality --
but it still seem like a good idea to me.

Richard A. Harmon          "The only good zombie is a dead zombie"
harmonra at           E. G. McCarthy

More information about the Squeak-dev mailing list