Hello,

I have been adding stuff to the wiki as I get feedback from the dev list. I
did not include all comments/requirements. I tried to keep the requirements
to the most basic level. It would be nice to find some hardware to use for
hosting the project and also discuss what platform to use for development

I know work has already been done using C#. Would it be feasible to use Java
and Spring for this project? What are your thoughts?

--Dan

Search Discussions

  • Itamar Syn-Hershko at Sep 24, 2010 at 9:10 am

    On 22/9/2010 9:23 PM, Dan Cardin wrote:
    Hello,

    I have been adding stuff to the wiki as I get feedback from the dev list. I
    did not include all comments/requirements. I tried to keep the requirements
    to the most basic level. It would be nice to find some hardware to use for
    hosting the project and also discuss what platform to use for development
    As I responded to Grant's e-mail, I think the first efforts should be
    made on creating a banal implementation that only allows to judge a
    document from a corpus against a given topic. This functionality alone
    can provide us with priceless data. Stuff like saving judgments of
    results per query etc is much more complex to do right, and should be
    left for phase 2, IMO.

    For that first phase, there isn't much that needs to be done.
    I know work has already been done using C#. Would it be feasible to use Java
    and Spring for this project? What are your thoughts?
    Orev was build on top of ASP.NET MVC 2.0, and NHibernate as an ORM. It
    should be fairly easy to port it to Java using Spring or Struts and use
    Hibernate for DB connection.

    But its really a question of availability and time. It only makes sense
    to do that only if you and the rest of the community can put the time to
    do the port and then maintain it. Otherwise, I think it makes sense to
    just complete the .NET implementation, and then expand it upon need.
    IIRC, Apache can host Orev in its .NET shape as well, so from that
    perspective it won't make any difference.

    I should be available to complete whatever left in about a month or so.

    Itamar.
  • Tommaso Teofili at Sep 24, 2010 at 9:53 am
    Hi all,
    I try to give some hints but I am pretty new to ORP so I may miss some
    points.

    2010/9/24 Itamar Syn-Hershko <itamar@code972.com>
    On 22/9/2010 9:23 PM, Dan Cardin wrote:

    Hello,

    I have been adding stuff to the wiki as I get feedback from the dev list.
    I
    did not include all comments/requirements. I tried to keep the
    requirements
    to the most basic level. It would be nice to find some hardware to use for
    hosting the project and also discuss what platform to use for development
    As I responded to Grant's e-mail, I think the first efforts should be made
    on creating a banal implementation that only allows to judge a document from
    a corpus against a given topic. This functionality alone can provide us with
    priceless data. Stuff like saving judgments of results per query etc is much
    more complex to do right, and should be left for phase 2, IMO.
    +1

    For that first phase, there isn't much that needs to be done.


    I know work has already been done using C#. Would it be feasible to use
    Java
    and Spring for this project? What are your thoughts?
    Orev was build on top of ASP.NET MVC 2.0, and NHibernate as an ORM. It
    should be fairly easy to port it to Java using Spring or Struts and use
    Hibernate for DB connection.
    But its really a question of availability and time. It only makes sense to
    do that only if you and the rest of the community can put the time to do the
    port and then maintain it. Otherwise, I think it makes sense to just
    complete the .NET implementation, and then expand it upon need. IIRC, Apache
    can host Orev in its .NET shape as well, so from that perspective it won't
    make any difference.
    From what I read on ORV requirements page [1] I understand it should run on
    most OSs so maybe .NET could not be the "ideal" architectural choice, even
    if Mono could trigger it on Unix-like platforms.

    Then I agree with Itamar that making a port (I don't know history about this
    already-in-place-.NET-impl) requires some time and community effort; however
    I think it would be better to have such a Java based port in terms of
    portability but also in terms of integration with many Lucene ecosystem
    projects that runs on top of a JVM. Moreover as I imagine UIMA [2] as a
    possible architectural framework candidate for processing corpora it could
    make more sense to have a Java impl.

    I should be available to complete whatever left in about a month or so.

    Itamar.
    Newbie question here: what is left to complete? The .NET ORV impl?
    Thanks and regards,
    Tommaso



    [1] : https://cwiki.apache.org/ORP/open-relevance-viewer.html
    [2] : http://uima.apache.org

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupopenrelevance-dev @
categorieslucene
postedSep 22, '10 at 7:24p
activeSep 24, '10 at 9:53a
posts3
users3
websitelucene.apache.org...

People

Translate

site design / logo © 2018 Grokbase