Joel's work log

From OpenCog
Jump to: navigation, search

Both for interested parties, and so that I can keep track of ideas and what I've been working on, I'm going to start recording all my work including unfruitful avenues to avoid repeating the same mistakes if I have to leave something and come back to it.

TODO


Documentation Tasks

Log entries

July

  • Misc bug fixes
  • differentiated agents-list and agents-active commands
  • GSoC guidance of David Kilgore and others in IRC.
  • Began adding unit tests for the PLN BITNode classes.
  • Integrated Ubigraph client library into codebase (as it's open source and relatively small).

June

  • Applied (and cleaned up) various patches from contributors.
  • Playing with RelEx OC output and started collecting frame output.
  • Checking David Kilgore's python bindings work
  • Merge my RelEx changes with Linas's updates
  • Investigating the appropriate axioms required for syllogism Mortal Socrates
  • Discovered that "men" doesn't form a LemmaLink to "man". Linas fixed this, but it doesn't resolve the fact that some words have multiple lemmas. Thus RelEx may need to be altered to record multiple lemmas for a word. This will, among other things, require RelEx2Frame to apply the rule against each lemma.
  • Investigating why the UUID of words in RelEx interfers with Relex2Frame rule matching...
  • Couple of days trying to get R2F to do conversion to UUID words properly. Problem is that constants in R2F rules, as well as variables, need to be converted to appropriate UUIDed word instance.

May

  • Moved PLN from reasoning to opencog::pln namespace.
  • Integrated various task lists into Doxygen.
  • General code maintenance (trying to get code comments to be the right format for Doxygen).
  • Compiled link-grammar, RelEx and it's dependencies.
  • Got cogita-bot code running and interacting in opencog (named p-bot in #opencog - short for PLN-bot)
  • Added command line configuration for cogita-bot
  • Doxygen cleanup - worked out how to integrate README files into doxygen output (post to email list).
  • Investigate build types for CMake (discovered ccmake and cmake -i)
  • Investigated BigTable providers to guide GSoC student Jeremy. Decided on HyperTable.
  • Added script to run relex, cogserver, and chatbot in separate tabs of gnome terminal.
  • Clean up categories on the wiki.
  • Lots of PLN shell usability fixes.
  • Remove many global variable from haxx in PLN.
  • Trying to get Linas' ODBC persistence back end working. All the pieces in place, but it segfaults.

Old