I agree that a lot can be done to make cTAKES easier to use and there will always be room for improvement. However- I strongly disagree that we should delay releases for new features. We should release early and often. Folks can continue working off trunk; a tag/release does not stop anyone from to continue and work on enhancements.
Considering it’s been almost 2 years since the last release; There are a lot of fixes that should justify a release. Sub-task [CTAKES-124 <https://issues.apache.org/jira/browse/CTAKES-124>] - remove internal UIMA types from coreference Bug [CTAKES-76 <https://issues.apache.org/jira/browse/CTAKES-76>] - get third party dependencies into Maven Central [CTAKES-155 <https://issues.apache.org/jira/browse/CTAKES-155>] - SimpleSegmentWithTagsAnnotator assumes all section names are 5 characters [CTAKES-178 <https://issues.apache.org/jira/browse/CTAKES-178>] - parsing of medication strength does not verify a number was discovered (strength value includes both the dosage and strength value in some cases) [CTAKES-213 <https://issues.apache.org/jira/browse/CTAKES-213>] - ModifierExtractorAnnotator should produce XxxxModifier subtypes [CTAKES-275 <https://issues.apache.org/jira/browse/CTAKES-275>] - some of the older junit tests don't have the right Project name in the run configurations [CTAKES-302 <https://issues.apache.org/jira/browse/CTAKES-302>] - Element type "hibernate-mapping" must be followed by either attribute specifications, ">" or "/>". [CTAKES-320 <https://issues.apache.org/jira/browse/CTAKES-320>] - Methods used by getDefaultPipeline should be able to load reasonable defaults without expecting external files. [CTAKES-341 <https://issues.apache.org/jira/browse/CTAKES-341>] - FractionFSM annotates incorrect span [CTAKES-364 <https://issues.apache.org/jira/browse/CTAKES-364>] - Dictionary-Lookup-Fast will fail with URI is not Hierarchical [CTAKES-371 <https://issues.apache.org/jira/browse/CTAKES-371>] - update PTB tokenization logic in fast dictionary module [CTAKES-376 <https://issues.apache.org/jira/browse/CTAKES-376>] - ctakes-temporal throws NPE if LVG isn't run [CTAKES-379 <https://issues.apache.org/jira/browse/CTAKES-379>] - Placeholder - ISO Time Normalizer (wrapper) [CTAKES-389 <https://issues.apache.org/jira/browse/CTAKES-389>] - cTAKES dictionary lookup missed word starting string bug [CTAKES-401 <https://issues.apache.org/jira/browse/CTAKES-401>] - License headers are JavaDoc comments but should be block comments Improvement [CTAKES-77 <https://issues.apache.org/jira/browse/CTAKES-77>] - Update POSTagger Unit Tests [CTAKES-78 <https://issues.apache.org/jira/browse/CTAKES-78>] - Update Chunker unit tests [CTAKES-122 <https://issues.apache.org/jira/browse/CTAKES-122>] - include LVG with a future version of cTAKES? [CTAKES-172 <https://issues.apache.org/jira/browse/CTAKES-172>] - relation-extractor is using StatusAnnotator and NegationAnnotator instead of AssertionAnnotator [CTAKES-295 <https://issues.apache.org/jira/browse/CTAKES-295>] - Use UIMAFit-style configuration annotations [CTAKES-328 <https://issues.apache.org/jira/browse/CTAKES-328>] - Clean up XML Annotator Descriptors [CTAKES-366 <https://issues.apache.org/jira/browse/CTAKES-366>] - Add Cas Consumer that will save pretty human-readable annotations to file [CTAKES-367 <https://issues.apache.org/jira/browse/CTAKES-367>] - FileLocator should output $CLASSPATH if file isn't found [CTAKES-368 <https://issues.apache.org/jira/browse/CTAKES-368>] - Allow alternate CUI formats in fast dictionary lookup module [CTAKES-373 <https://issues.apache.org/jira/browse/CTAKES-373>] - MaxentParserWrapper can't handle section dividers: "==========" [CTAKES-380 <https://issues.apache.org/jira/browse/CTAKES-380>] - Placeholder - New Sentence Detector [CTAKES-381 <https://issues.apache.org/jira/browse/CTAKES-381>] - Improved Location Of Model [CTAKES-385 <https://issues.apache.org/jira/browse/CTAKES-385>] - URI is not hierarchical when attempting to obtain lvg.properties within JAR [CTAKES-386 <https://issues.apache.org/jira/browse/CTAKES-386>] - Remove all .classpath and.project files [CTAKES-387 <https://issues.apache.org/jira/browse/CTAKES-387>] - Upgrade gov.nih.nlm.nls lvg artifact across the codebase [CTAKES-388 <https://issues.apache.org/jira/browse/CTAKES-388>] - URI is not hierarchical when attempting to obtain LookupDesc_Db.xml within JAR [CTAKES-399 <https://issues.apache.org/jira/browse/CTAKES-399>] - Add profile to pom.xml which permits build with JDK1.8 New Feature [CTAKES-314 <https://issues.apache.org/jira/browse/CTAKES-314>] - BigTop/Hadoop cTAKES integration [CTAKES-378 <https://issues.apache.org/jira/browse/CTAKES-378>] - Fast Dictionary should be able to load custom codifications from db [CTAKES-382 <https://issues.apache.org/jira/browse/CTAKES-382>] - Add ability to easily add extension of UmlsConcept Type to jcas via dictionary lookup Task [CTAKES-74 <https://issues.apache.org/jira/browse/CTAKES-74>] - Tokenizer PennTreeBank breaks with certain apostrophes in tokens. [CTAKES-138 <https://issues.apache.org/jira/browse/CTAKES-138>] - Remove 3rd party jars from our SVN [CTAKES-232 <https://issues.apache.org/jira/browse/CTAKES-232>] - change concept type <>On Dec 6, 2016, at 11:20 AM, Jeff Headley <jeffun...@gmail.com> wrote: > > I realize I’m not a committer and maybe I shouldn’t express an opinion. > Apologies in advance if this is inappropriate. However as someone who has > gone through the pain of trying to install, learn, and use ctakes; I strongly > agree with Sean. I don’t inject myself into the situation lightly or to > “vent”. I have been in software development since 1996 and a lot of that time > in medical projects and using various open source frameworks like Spring, > Seam, Hibernate, etc. Sean is right. > > Jeff > > On Dec 6, 2016, 10:00 AM -0500, Pei Chen <pei.c...@wiredinformatics.com>, > wrote: >> Considering the amount of time since the release was created, we should not >> let any pending Jira’s or features hold up a release. >> I suggest just we mark anything that hasn’t been fixed in Jira into the next >> release and push forward- I’ll volunteer to do that right now. >> In the past, the documentation on the website also shouldn’t hold up a >> release either. >> >>> On Dec 6, 2016, at 9:20 AM, Finan, Sean <sean.fi...@childrens.harvard.edu> >>> wrote: >>> >>> Hi Murali, >>> >>> Before we make an rc, we must go through the list of currently open tars >>> and requests. SOP. A list needs to be compiled of what should be closed as >>> fixed or n/a plus another list of outstanding bugs that need to be dealt >>> with and an estimate of effort. Then we should try to gather volunteers to >>> handle said bugs. Can you take care of compiling those lists? I did this >>> many months ago when rc 3.2.3 came up, and there were items on which no >>> movement was made. If you can find my email that might be one place to >>> start. >>> >>> The primary takeaways from the hackathon were, not surprisingly: >>> 1. Installation of cTAKES is not as straightforward as we believe, and >>> 2. Getting started with cTAKES is extremely difficult (no good starting >>> point) and scares off a large percentage of people who try. >>> 3. Customization is next to impossible without diving into the code, which >>> is more time consuming than anyone can stand. >>> >>> All can be handled best by short and simple GUI tools and some "cTAKES for >>> Beginners" documentation. We have some documentation that was used for the >>> Hackathon that needs to be modified a bit, then posted on the main cTAKES >>> website. >>> >>> In my opinion these items should be worked upon before creating another >>> release, otherwise the release is not as useful as it could be. I have >>> started work on a simple pipeline builder gui that creates simple html or >>> text output. I will check it into trunk soon, but as new functionality >>> community testing will be required before a release. >>> >>> Sean >>> >>> -----Original Message----- >>> From: Murali Minnah [mailto:mmin...@gmail.com] >>> Sent: Monday, December 05, 2016 1:26 PM >>> To: dev@ctakes.apache.org >>> Subject: cTAKES - 3.2.3 release >>> >>> I wanted to check to see if there are objections to creating a 3.2.3 tag of >>> trunk now to prepare for a 3.2.3-rc1? >>> >>> Any comments from the participants/organizers on the success/lessons learnt >>> from the "hackathon" that the community can benefit from? >>> >>> Best, >>> Murali >>
