stack architecture

[Sadhunathan Nadesan]

/ FROM:  Robert Brenstein <rjb at rz.uni-potsdam.de>, Dec 12  1:30 2001
| ABOUT: Re: stack architecture
|
| >Problem is that 2 has to be a substack of 3 to inherit its algorithms
| >and this causes the files to be saved as each other.  I have them
| >horribly mixed together and delete doesn't remove the algorithm code
| >from the data file, or the data from the algorithm file.
| 
| No, 2 does not have to be a substack. As a matter of fact, to achieve 
| the distribution goals, each of your stacks must be alone in its own 
| file.
| 
| There are several different mechanisms that you can use to get your 
| data and algorithms together. A simple one is to use fixed names and 
| complete references to objects in other stacks. For example,
| 
| get fld "data1" of cd "sampledata" of stack "sadjudata"
| 
| You just need to ensure that the shared items have unique names.
| 
| Another mechanism is "stacksInUse" which allows you to make stack 
| scripts available to any other stack.
| 
| Yet another mechanism is using frontscripts/backscripts feature of Metacard.
| 
| Robert
|
\ END: Robert Brenstein


Wow, Robert, I think you saved my sanity, thank you.  Your email
mentioning 'stacksInUse' led me to find the "start using" command which
seems to solve my problem.

I have now ended up with 2 stacks, a startup stack and a data stack.
I hope when I compile the startup stack into a standalone the start
using will still work.  That will be a next test.

Meanwhile, during the time I was struggling with the effort to keep
the data and the algorithms in separate files, but thinking the only
way I could share the algorithms was by making the algorithm stack the
mainstack of the data stack .. I seem to have gotten my data stack to
be a substack of itself.  It's own name is listed as a substack in
the components dialogue.  I can't seem to get rid of it from there.
Any ideas?

Anyway, thank you very much and for such a quick response!

Aloha,
Sadhu