in reply to Re^5: CPAN indexes *.pm as "Documentation"? (Memory leaks)
in thread CPAN indexes *.pm as "Documentation"?

References to objects outside of the direct inheritance hierarchy. (Nothing you need, I'd guess.) I don't want code to store the IDs in strings or something because they might be renumbered. If e.g. a part of an object definition are to be copied into some other place.
"Has-a" versus "Is-a" relationships? That's what's bugging me: nodes on trees (especially terminal nodes) ought to be crossreferenced with Operational Taxonomic Units (more sane people would call them "species"...), which in turn ought to be crossreferenced with comparative data (e.g. DNA sequences).

None of these relationships are through inheritance, but more data-oriented rather than object-oriented. This all works in that I have methods to match up the different types of information by species name, and build references between the objects - but now everything is one network of circular references that I'm pretty sure keeps floating somewhere in RAM space after it's gone out of scope (sidenote: how might you test that???).

I'm interested to see how you go about doing that, so will you post a message to announce your module?
  • Comment on Re^6: CPAN indexes *.pm as "Documentation"? (Memory leaks)