[kepler-dev] Combined commandline and exec actors

Bertram Ludaescher ludaesch at sdsc.edu
Sun Jun 26 02:48:26 PDT 2005


Ilkay:

Nice that those have been merged (or rather their functionality).
BTW, what were the main differences of the two? (Sorry, I think that
might have been discussed before).  I guess merging made sense since
each actor had something useful that the other didn't have (in
addition to sparing the user to think about which of two very similar
actors to use).


All: 

I assume that once the new KSW files are in place, together with the
object manager (a better name than cache manager, right?) then
updating of workflows is no longer needed!? 
Legacy workflows can continue to use "old" actors ad infinitum and all
of this will be automatic, transparent, easy going..

Chad, Matt, Ilkay et al: 

Aproposa KSW: I realize that the extension KAR is already used. But
who cares. I think the advantages over KSW outweigh the (probably
non-critical) issue of "some other program uses that extension" (KAR
correctly connotates *AR*chival files such as TAR and JAR; also it's
pronouncable unlike KSW).

Bertram

Ilkay Altintas writes:
 > Dear All,
 > 
 > I've just checked in the changes for Bilsay's combined version of the 
 > old CommandLine and Exec actors that Christopher and I have 
 > concurrently implemented last year.
 > 
 > The new actor is named "CommandLineExec" to have a different name and 
 > avoid conflicts.
 > 
 > The old actors are kept so that none of the existing workflows that use 
 > them will crash. It would be great if you could check your existing 
 > workflows at your convenience to adopt the new actor.
 > 
 > We would like to adopt the new actor in the future workflows so the old 
 > actors don't show up in the actor library.
 > 
 > Hope this change is not going to cause any inconvenience for anybody. 
 > We appreciate if you can test it and send comments before Bilsay's 
 > internship ends in a few weeks.
 > 
 > Bilsay: I think we need to have an error port in the new interface to 
 > divert the standard error messages to. It is not an output but quite 
 > important for usability of the actor.
 > 
 > Thanks very much to all!
 > -ilkay
 > 
 > _______________________________________________
 > Kepler-dev mailing list
 > Kepler-dev at ecoinformatics.org
 > http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev


More information about the Kepler-dev mailing list