uima-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marshall Schor <...@schor.com>
Subject Re: UIMA Quarterly Board report is due
Date Mon, 12 Sep 2011 17:20:59 GMT
Hi everyone, here's the proposed board report.  Please correct /
augment as needed.  It may be a bit long, so I may shorten it slightly...

Board report for Apache UIMA, for September 2011.

Apache UIMA's mission: the creation and maintenance of open-source
software related to the analysis of unstructured data, guided by the
UIMA Oasis Standard.

Since last report, the Addons package for UIMA was released
(http://uima.apache.org/news.html#29 August 2011).  The addons package
contains 2 new annotators:
  Solrcas (for storing CAS objects into an Apache Solr instance), and
  AlchemyAPIAnnotator (wraps alchemyapi.com services).

A new contributor, Peter Kl├╝gl contributed a UIMA tool to the sandbox,
called TextMarker.

A user has set up a French language portal to all things UIMA, and
contributed a French language models for the Hidden Markov Model (HMM)
Tagger annotator, and generally improved that annotator.
The French language models are awaiting getting some additional permissions
before they are put into SVN.

Some attempts to package UIMA Annotators as OSGi bundles led to renewed
investigations toward this, and some progress was made in identifying
approaches and tools, including Maven integration / support.

There continue to be lots of incremental Cas Editor fixes, mostly driven
by user feedback and the development of a new Cas Editor based plugins
at the Apache Incubator OpenNLP project.

UIMA-AS had a few bug fixes, and some new features, including exposing
per-component statistics (for tuning) from UIMA aggregates for each CAS.

No changes.
Issues: No Board level issues at this time

  Previous work to add TM started (at some point in time) to fail to display.
  With infra's help, traced this to the fact that Apache Web sites
  are now being displayed using UTF-8.  Our website, generated by
  Anakia, was in ISO-8859-1, indicated on the charset= attribute of
  the META tag for the html page.  This was being dynamically changed
  by the Apache website to UTF-8. 
  We fixed this by inserting a replace character step in our site
  generation ant script that replaces the TM in iso-8859 with the one
  for UTF-8.

 Branding checklist:
   no change from previous report:
  Project Website Basics - done
  Website Navigation Links - done
  Trademark Attributions - done
  Logos and Graphics - not done
  Project Metadata - done
  Read PMC Branding Responsibilities - partially done.  Need to get confirmation
    that all PMC members have read this.

On 8/31/2011 9:28 PM, Marshall Schor wrote:
> Where did the summer go?  It's time for the next quarterly board report.
> Please reply to this message with items for the report.
> I will insert a bit about the addons release.
> -Marshall

View raw message