search

dollabs / pamela

Probabalistic Advanced Modeling and Execution Learning Architecture
Apache License 2.0
233 stars 13 forks source link

Add HTN/TPN support for ask, tell, and assert Pamela statements (Issue 94) #95

Closed dcerys closed 7 years ago

dcerys commented 7 years ago

On Apr 27, 2017, at 9:25 PM, Tom Marble <notifications@github.com mailto:notifications@github.com> wrote:

@tmarble requested changes on this pull request.

Is there a reason the condition at 7e1ea1d#diff-89334453bea000451d6370c9a254cd9bR391 https://github.com/dollabs/pamela/pull/95/commits/7e1ea1de4daab466b24b16a9da1dc85d2402fbbd#diff-89334453bea000451d6370c9a254cd9bR391 doesn't want to be written using reserved-conditional-method-name? ? As above, also at line 1866 reserved-conditional-method-name? deals with symbols, and (1) and (2) deal with keywords. Fortunately, the names of the symbols and keywords are the same, so I can do some symbol/keyword coercion to localize the enumeration to one place. Good suggestion.

The concern at 1871 needs to be traced through the various "resolve arguments" call flow. It actually works. The argument resolution gleefully passes the map through to the HTN and TPN. That’s why I added to-pamela to convert the IR snippet to Pamela, for display purposes. So the concern doesn’t actually apply; I’ll delete the concern.

4, Could we get a new pamela source file in test/pamela (and corresponding IR in test/pamela/IR) demonstrating the new functionality? If it's not quite possible yet could the source file live in test/pamela/pending until necessary other bits are fleshed out? Yes, I’ll add one of my test cases.

Thanks for the feedback.