Cache option of compile?

["Jon Paris" <Jon.Paris@xxxxxxxxxxxxxx>]

  >> Jon, the cache algorithm is very simple. ....... So, it will only cache
what that particular verify action requires.

I'll take your word for it Phil but it takes so long I have always felt that
it must be doing everything.  Guess it is just slow.  It is interesting that
if you have a really fast box (even a high-end 270) you can run a batch
compile and get the results faster than a Verify with refresh cache.  There
is something wrong with that picture.

As to your other comments.  I'm delighted to hear that the spec will be
open.  I've often said in the past that if IBM will give us the cache spec
we can do it ourselves.  As individuals or a group we can produce something
with far more compromises than IBM can.  For instance - Craig has been
making the point that he does iterative development of (say) a display file
along with the program that uses it.  It would be great for him to simply
use a pull down in the DDS editor to say "Update cache with changes".  If
IBM were to develop that functionality it would _have_ to work with
reference files (as one example).  If we build it ourselves, we can simply
accept that such support is not possible in the first release etc.


Jon Paris
Partner400