Mailing-List: contact ctakes-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: ctakes-dev@incubator.apache.org
Received-SPF: pass (nike.apache.org: local policy)
From: "Kusenda, Brandyn J" <brandyn-kusenda@uiowa.edu>
To: "ctakes-dev@incubator.apache.org" <ctakes-dev@incubator.apache.org>
Subject: descriptors referencing by location vs name
Thread-Topic: descriptors referencing by location vs name
Thread-Index: AQHNmyeub7TD/zFcREaQYXomIbg+9w==
Date: Tue, 25 Sep 2012 14:49:57 +0000
Message-ID: <29A5CA3F695C3C429DF68C926D9BCEB0366FAFE5@ITSNT437.iowa.uiowa.edu>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

First off, thank you for sharing all your hard-work on this fantastic proje=
ct.  =0A=
=0A=
What is the reasoning behind choosing to specify descriptors by location ra=
ther than name?  In the context of Maven, using "name" seems to make creati=
ng a self-contained jar that executes a pipeline much simpler. I've been in=
vestigating running a UIMA pipeline on a hadoop cluster and referencing all=
 descriptors by name is one of the suggested steps. I personally have not e=
xperienced any need to recompile because of a descriptor change in my Eclip=
se environment when the descriptors are in the classpath.    =0A=
=0A=
Also, as someone who would like to use some or all of cTAKES as a module in=
 our own pipeline, having each component somewhat self-contained is really =
helpful. =0A=
=0A=
So in regards to your thread on where to keep resources, I would suggest ei=
ther:=0A=
- creating another artifact (example: dictionary-lookup-resources) for each=
 artifact's resources or =0A=
- including the resources in each artifact in the classpath (example: /src/=
main/resources).=0A=
=0A=
=0A=
Thanks,=0A=
Brandyn=0A=