Development tasks, in order, and who is doing
for research in December
- ______(Judah) few line of code to pop up UI as per new procedure arriving
- ______(Judah) live mode testing of classifier persistence facility (add code in trainer class)
- ______(Judah) Attachment predictor class needs to read in classifier
- \SKIP(Jed, Oliver) no body event - Jed try html vs plain text email adjustment

- \SKIP(Chris/Oliver) successfully read Sensitivity from Pojo (awaiting help from Oliver, Chris)
- ___(Jed) harvesting issue "to" field not harvested correctly -> check IrisTracer log to see if To is populated in final email harvest
- ___API (JI) SATService.parseSparkleString(String s) using SAT (labelling working, feature detection not)
- ___API (JI) SATService.inferForgetting(IEmailMessagePojo p, boolean userIntendedTheAction) using SAT
- ___demonstrate fully integrated system (minus SAT)
- ______(Judah) (not necessary for demo or paper?) logic for switching from training mode to predicting mode
- \SKIP______(Judah) awareness of when emailFinal is sent to adjust training samples?
- \SKIP______(Judah) if hack still in place, correct the KB with final email, and persist training example
- \SKIP___(team) decide whether we will give learning feedback to the user
- ___(Chris) incremental reharvesting not working with outlook
- ______(Jed) schedule debugging session with Chris for later in week re -incremental harvesting problem
- ___IMPLEMENTATION (KJ) extend EmailContextInfo to include other entities and their
- \SKIP___(Oliver) no emailFinal event flowing (still looking into it as of 1118 mtg)
- ___(Oliver) batch harvesting in outlook
- debugging___API (JI) SATService.getLabelForEmail(IEmailMessagePojo p) using SAT
rough edges (12/09/08)
- ___(JI) consume live "new procedure event" from ITL. Got the updated api from ITL folks to work (99% sure) but never finished hooking it into AttachmentPredictor.init due to lack of time before viseo production needed to commence.
- ___(Jed) grant_all.policy adsded to build.xml so its copied to uil_plugin from uil_plugin_src - only done on venetie - not committed. So , never committed the ITL and OSU full integration to the point we got it to.
- ______(Jed) Tom guided me to not do dangerous last minute jar update commit of ITL dependency jars into uil_plugin\lib dir or in other places in iris
- ___(Jed) special hack jars for itl ui and ITLClient API (Jed note 12/09/08 - don't remember if this was still an issue - might have been tweaking jars during troubleshooting or this might just be referring to getting the references right for ITL.
- ______(Jed remove Jython files from uil_plugin - decided not to do this as need it for tests to translate sparkle into SharedSpark)
- DONE IMPLEMENTATION (KJ) implement UIL manager and learner component
- DONE___FEATURE (JI) generate SAT problem for forgetting inference
- DONE___(JI) fake forget detection for demo until SAT works
- DONE___(JI) fake email labeling for demo until SAT works# DONE FEATURE (JI) parse sparkl string into UILActionSequence
- DONE EVAL (KJ, JI) get Tom's partially labeled data into Thunderbird
- DONE EVAL (KJ) harvest that data in calo
- DONE API (JI) SATService.parseSparkleString(String s) faked
- DONE API (JI) SATService.getLabelForEmail(IEmailMessagePojo p) faked
- DONE API (JI) SATService.inferForgetting(IEmailMessagePojo p, boolean userIntendedTheAction) faked
- DONE/UT UTIL(JI) POJO value lookup for Strings
- DONE/UT UTIL(JI) POJO value lookup for Attachments
- DONE(JI/KJ) Generate email Data
- Done(JI) fake feature legailty for demo until SAT works
- DONE FEATURE (JI) adjust sparkl parsing to consume final format of sparkl procedure
- DONE___FEATURE (JI) SATService session persistence
- DONE___TASK (JI) Prep demo machine for calo
- DONE___TASK (KJ) pull iris, desktop onto demo machine
- DONE___FEATURE (JI) generate SAT problem for labelling
- DONE___FEATURE (JI) generate SAT problem for feature legality
- DONE___EVAL (KJ) perform experiments on Tom's data for attachment prediction and generate cross-validation results
