deg:fall-2015-minutes

December 7, 2015

Dr. Lonsdale: Not much to report – trying to survive the semester (life begins again after tomorrow)
Dr. Woodfield: Also working on surviving the semester; some work with Dave
Dr. Embley: Has been doing lots of programming: cleaning values, etc. Working on parsing of dates. Things are pretty well ready to go to the constraint enforcer. Ready to place notices in XML docs. Will work on messages for handlers. Then will need to split into forms.
Dr. Liddle: Did some work with Dr. Embley, also trying to survive the semester
Sidebars:
- How to place new messages into JSON. How do I code that in JavaScript?
- How to learn Android.

November 24, 2015

No DEG meeting this week

November 17, 2015

Dr. Liddle: not a lot to report
Dr. Lonsdale: ditto; back from trip yesterday
Dr. Woodfield: made a little progress working on plug-in directories, etc.; has ideas on general constraints, but won't pursue those yet; needs ability to identify connections and relationship sets in a manner that is externally stable, so assumptions:
- Object set names are unique
- Relationship set names are sufficiently unique (mostly identified by connected object sets; if not, RS name is unique)
- Relationship-set connections are uniquely identifiable (e.g. by specialization if needed)
Dr. Embley: ListReader bug that prevented running Ely pages is resolved; have been chasing the “no labeling” issue (where user has chosen not to label something ListReader asked to have labeled); working more on the merge
Sidebars:
- “No labeling” original specs
- End-of-line hyphens: what happens in COMET?
- Christopher: are we on same page?
- Logging issue: how do we not send pages to console?
- Discussion on messages to be sent to COMET with constraint checking

November 10, 2015

Christopher: working on updating PHP models for saving and loading pages, annotations, etc. with the new filesystem (COMET project); has a job interview next week in Florida
Dr. Embley: merge tool is working; has full thin line through the entire system again; working on ListReader and found a workaround for the bug, but not sure – build is broken but needs to merge changes in two branches
Dr. Liddle: met with Christopher to work on COMET restart; noted that we needed update to fe6 filesystem
Dr. Lonsdale: got the new functionality working that puts raw text in original text field of the annotations; unknown whether offsets are okay now; Dr. Embley to look for potential issues with offsets
Sidebars:
- /data/fe6 filesystem structure
- OntoSOAR from the command line?
- COMET design question (Liddle/Almquist)
- ListReader code merge (Liddle/Embley)

November 3, 2015

Dr. Woodfield: making some progress; blocker: can't get classForName() to load a class in Netbeans
Christopher: has gotten back into working on things; remembering past issues; may want to meet with Dr. Liddle (Friday 9am 784 TNRB)
Dr. Embley: has been having more fun than a barrel of monkeys getting the merge to work, but now it appears to be working (gen/spec); needs to get back to ListReader (bug deep in the code)
Dr. Liddle: spent some time with Dr. Embley working on debugging, etc.
Dr. Lonsdale: working on the same mini-project; making progress
Sidebars:
- classForName() loader
- Q4 goals
  1. Integrate OntoSOAR and ListReader into the ensemble, working satisfactorily
  2. Merge of data from extraction tools prepared for input to COMET
  3. Constraint enforcer designed and planned for COMET integration, coding under way
  4. Fill out the full-line pipeline so we handle corner cases in batch processing (e.g. complex and off-page annotations)
  5. GreenFIE rule generation working and integrated into ensemble
- Interface with constraint enforcer and COMET

October 27, 2015

Dr. Lonsdale: fixed the enhancement issue he's been working on for a while (also works on Dave's machine); now working on offsets and system-corrected text to forward original text
Dr. Woodfield: two blockers: 240/340
Christopher: he's back!
Dr. Embley: spent last week with George Nagy; has been working on merge program; not quite sure what to do with overlaps; should be ready for Dr. Woodfield to use OSMX as input
Tae Woo: baby is growing! Had issues with charger for computer, but back in action again. Now creates fully-built regular expressions, and is ready now to run experiments.
Dr. Liddle: Austria, Sweden, ER report
Sidebars:
- Overlaps
- Deleting bogus annotations and objects
- log4j warning

October 20, 2015

No DEG meeting

Schedule:

Dr. Embley in California 9/29 and on-call for two weeks, also gone 10/20 to New York w/Nagy
Dr. Liddle 10/7 through 10/23 in Austria and Sweden

October 13, 2015

Yes DEG meeting

October 6, 2015

Dr. Liddle: hier ist alles super; headed to Austria/Sweden 10/7 through 10/23
Dr. Lonsdale: working on a new project that has taken a lot of time until General Conference
Dr. Woodfield: found you can have any value for Object-Set Cardinality Constraint as long as it's “1”
Christopher: has been sick for the past two weeks; barely making it through school; should be better soon
Dr. Embley: enjoyed Disneyland with grandkids; has done coding for transferring ListReader output to common extraction ontology format; will now do a merge of all the information from the various tools; needs to do some parsing and convert dates into form the constraint checker can use
Sidebars:
- Constraint checker
- Access to DEG backlog Google doc

September 29, 2015

No DEG meeting

September 22, 2015

Dr. Embley: working on the finish-up of ListReader; it's moving along
Dr. Liddle: worked with Dr. Embley on ensemble; we found an interesting bug with UTF-8 characters in the input files
Dr. Lonsdale: was making some fixes to binary relationship sets; test showed a few remaining issues; things run with Windows 10 without having to do much
Dr. Woodfield: has The Big Nerd Ranch Guid to Android Programming (w00t!); Dr. Clement has a place for Christopher in his lab (at least screen and keyboard, maybe a machine); working on constraints; thinking about how to integrate the checker with COMET (possibly separate file containing constraint violations); dynamic handler-based architecture
Christopher: had a few minor issues with BYU systems; thinks those are fixed now; working on getting pages to load in COMET; have a minor issue with getting ListReader mode working
Sidebars:
- Constraint checker integration with COMET
- ListReader mode issue

September 15, 2015

Christopher: helped Tae Woo; having a hard time to find a place to work on campus; can finally work again and will begin being more productive soon
Dr. Embley: did some ListReader coding; decided we need to use filenames etc. as Thomas set them up
Dr. Liddle: helped Dr. Embley with some refactoring on the Fe6 Ensemble and ListReader process
Dr. Lonsdale: has been working on nulls problem and made progress (still more to do); got Soar 9.4 on Dr. Embley's FamilySearch computer
Tae Woo: continues to build up full regex; now needs to turn digit extraction into regex, changing field values to capture groups, and do some generalizations
Dr. Woodfield: has a blocker: in the XSD, co-occurrence constraints can be stored in relationship sets, but the Ontology Editor stores them only at the global level, so there are ambiguities; never use Undo – always use Delete
Sidebars:
- How to run tools over multiple pages (e.g. OntoSoar, FROntIER)
- Refactoring ListReader and Fe6 Ensemble
- Co-occurrence constraints in the XSD vs. how they are created in the Ontology Editor

September 8, 2015

Dr. Woodfield: Nothing to report
Christopher: Working on ListReader; fixed some PHP issues; campus WiFi working again, so hopes to get more done this week
Dr. Embley: Wrote code to do “short-circuiting”
Dr. Liddle: Worked with Christopher and then Dr. Embley on ListReader things
Dr. Lonsdale: Working on getting SOAR 9.4 running on Dr. Embley's machine; need to check DLLs stored in Subversion; still long list to work on
Tae Woo: Working on FormRenderer; ran into bug of continuous loading; commented out changes but still couldn't get it to work; changed to Christopher's latest version and it now works; has questions about text version of PDF pages
Sidebars:
- Missionary joining the group as coder
- ListReader (Liddle/Embley/Almquist)
  - Christopher's question
  - Interface issues: short-circuit path
- Embley needs to run PDFIndexer (Liddle/Embley)
- SOAR 9.4 Subversion Audit for SOAR DLLs
- AJAX retrieve text page
- Embley/Lonsdale at 4pm

September 1, 2015

Tae Woo: has the correct offsets and text and can create a regular expression; still exploring ways to create regex (e.g. how many context characters should he use); no blockers
Christopher: between moving and computer issues, hasn't accomplished as much as he would have liked; looked over new ListReader specs, has an idea of how to implement; has some questions (meet with Dr. Liddle Friday @ 9am)
Dr. Embley: made some progress on ListReader, creating input according to specs; wants to work on short-circuit of COMET (sidebar); HIP and ICDAR were great
Dr. Liddle: happy to have ER 2015 proceedings in press with Springer; involved in conference prep; CET managing director resigned, meaning I'll be more occupied
Dr. Lonsdale: ended up not going on trip because father-in-law passed away; building OntoSoar on Unix is something to do in not-too-soon future
Sidebars:
- HIP/ICDAR report: flight to Frankfurt was canceled, so had to swap presentation times with someone else; ICDAR was quite interesting: three of four keynote speakers addressed impact of document analysis efforts (large knowledge systems, e.g. along the lines of CMU's NELL); Faraday in 1800's couldn't keep up with scientific knowledge of the day – what would he think today? Big Data clearly is an issue here. There is a desire to make a greater impact with our document understanding abilities. Reading or extracting with respect to an ontology is something we could emphasize with this community. Perhaps we could augment FROntIER paper with respect to reading, talk about ensemble and target DAS 2016.
- ListReader: specs, bypassing COMET
  - Q. What response should we give to ListReader?
  - A. COMET will need to send everything that corresponds to the annotated form.
- Combining output from various tools into a form