incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dan Haywood <dkhayw...@gmail.com>
Subject Re: Apache Isis: project pre-proposal: looking for mentors and a sponsor
Date Fri, 30 Jul 2010 11:22:02 GMT
Hi Manos,

On 27/07/2010 13:57, Emmanouil Batsis (Manos) wrote:
> Hello Dan,
>
> On 07/21/2010 12:56 PM, Dan Haywood wrote:
>
>> So: the Apache Isis (?) project will provide the ability to ...
>
> Sounds cool. ... Anyway, I checked out the provided URLs and I'm 
> having a really hard time understanding the plumbing used. Perhaps 
> some examples would help people (and me!) figure out if they are 
> interested in participating.
>

Thanks for the feedback; which means that (a) we should have better info 
on the NOF/Isis website, and (b) I should include it in the incubator 
proposal.

But to answer your question right here:
* Naked Objects core (http://www.nakedobjects.org/core) has very few 
dependencies; indeed when originally written it depended only on crimson 
and on log4j, which is why I was able to cross-compile it to .NET as a 
J# project from which the commercial .NET version eventually emerged.
* right now the core depends on Apache commons, and I've recently 
started using google collections since it's so nice.  We've also got a 
discussion going on to start using Google Guice in order to remove some 
of the boilerplate, though a refinement of that will be to use 
JSR299/330 annotations.
* Another thing for the roadmap is to introduce instrumentation, 
probably via JMX.  Another possibility is to expose more clearly the 
transactional boundaries so that we can leverage JTA services (at the 
moment the transaction management stuff is pretty much home grown); this 
will allow for transactional domain services.

The plugins ((http://www.nakedobjects.org/plugins) and the sister 
projects (http://starobjects.org) do have more dependencies:
* the berkeleyDB persistor is a persistor for Oracle Berkely
* the JPA persistor uses Hiberate under the covers; one of the tasks 
during incubation will be to replace this with OpenJPA (since Hibernate 
is LGPL)
* the Wicket viewer uses Apache Wicket, obviously
* the RESTful viewer uses JBoss RestEasy; again I'll need to change this 
to an Apache-license compatible library (probably Jersey)

One of the things that's nice about the architecture is that it's easy 
to add in new viewers/persistors based on XYZ technology.  One of the 
commentators here (Ulrich) has mentioned an interest in a Tapestry 
viewer, which would be great; I'm also in discussion with someone 
offline about a Vaadin viewer.  I'm kinda hoping that during incubation 
we might find someone who's might like to write a persistor based on one 
of those trendy NoSQL databases such as CouchDB or MongoDB, or perhaps 
Apache Cassandra or something from the Hadoop family

Hope that's useful (and hope that inspires people/you) to get involved,
Cheers

Dan


> Cheers,
>
> Manos


Mime
View raw message