[Savannah-hackers] Evaluation: OJE

[Hugo Gayosso]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


This evaluation was requested by savannah-hackers, how should we
proceed?

Do we send this kind of evaluations to you (RMS) and Cc:
Savannah-hackers?

I guess that you decide if this becomes part of GNU, and then notify
savannah-hackers so they approve the creation of the project in
Savannah.

Is this the procedure we want to follow.


From: Colin Walters <address@hidden>
Date: 06 May 2001 13:40:53 -0400

I. Package description

 -Package name and version:
        Open Java Extensions (0.1)
          
 -Authors
        Andrew Selkirk <address@hidden>
        Nic Ferrier <address@hidden>
        
 -Should the authors(s) be contacted? (Yes/No):
        Yes
        
 -Home page:
        http://oje.sourceforge.net
        
 -Distribution site:
        http://oje.sourceforge.net
        
 -Describe in your own words what job or jobs this program does.
        The Open Java Extensions is a project to implement some of
        Sun's extensions to the Java core class libraries, such as
        javax.mail, javax.activation, and javax.media.

        See the description attached to the evaluation request for a
        slightly more in-depth description.
        
 -Binaries available:
        There are precompiled Java bytecode files for the released
        portions of the project.
        
 -GNU/Linux support?
        Yes, via any one of the free Java systems for GNU/Linux, such
        as GCJ or Kaffe.
        
 -License:
        LGPL

II. Package specifics

 -Dependencies:
        The OJE components in general depend on the classes in the
        core Java class library, of which GNU Classpath is an
        implementation.
        
        The IMAP implementation in javax.mail depends on GNU Kawa.
        
 -Configuration:
        Being implemented in Java, most platform dependencies are
        hidden by the core Java classes.
        
 -Compilation:
        There are no Makefiles provided, but the OJE should be able to
        be compiled by any Java compiler.
        
 -Usability/interface:
        Since this is a project to reimplement an already defined
        interface (created Sun), interface design is less relevant.
        
        As for usability, the OJE classes are in varying states of
        completion.  There are still a number of FIXME comments, and
        zero-size files (representing classes that need to be
        implemented).  But a lot of the code is very useful as it
        stands.
        
 -What languages is the program written in
        Java

 -Code clarity/maintenance:
        The code is generally pretty clear, and it is heavily
        commented.
 
 -Performance:
        Performance seems ok.
        
 -Documentation:
    -- Does the package include a good introduction or tutorial manual?
        Not at the moment.

    -- Does the package include a good reference manual?
        Well, the class libraries include documentation in Javadoc
        format.  However, the quality and completion of the
        documentation varies somewhat.  It's difficult to blame this
        on the authors, however; the documentation from Sun is
        generally poor, and the OJE authors are trying to reimplement
        the interface in a bug-compatible way, and to come up with new
        documentation at the same time.

 -Does the program recommend or encourage the use of any non-free
  software?
        No.

 -Does it have have certain capabilities that can only be used in
  conjunction with some non-free software package?
        No.

III. Evaluation summary

 -Does the program fit coherently within the GNU system?
        I would say very definitely yes.  The GNU system already
        includes Classpath, and the OJE project makes a good
        compliment to that package.

 -Does the program meet necessary requirements for being
  a GNU package?  If not, can changes be feasibly implemented
  by the author in order for the program to be acceptable?

        I believe so.

 -Are there any licensing issues that need to be resolved?
        No.

 -Is there a large overlap with some other GNU package?
        No.

 -Is the program free of gratuitous incompatibilities with other GNU
  packages?
        Yes.

IV. Notes and comments during the evaluation process
    (Attach the "Comments" section from eval-pending.texi)

        None.


<summary>

oje
===

Received    : 06 Apr 2001
Requested by: Savannah <address@hidden>

Author     :  <address@hidden>

Homepage   :  <http://oje.sourceforge.net>
Source     :

Evaluator   : Colin Walters <address@hidden>

Started     : 28 Apr 2001
Completed   : 06 May 2001
Filename    : oje

Description:

The OJE (Open Java Extensions) project develops (L)GPLed versions of
Sun's java extensions.

Sun has a 2 levels of Java code. The java library is "core java" and is
(roughly) what is delivered with Sun's Java platform download. This is
also what the Classpath project is developing a free version of.

The second layer is the javax layer. This includes many different,
unrelated libraries. Sun group some of these libraries into an
Enterprise Java product but that's a marketing distinction rather than a
technical one.

For example, one library is the javamail library. This is distributed
under the name javax.mail. and is an abstract API for representing
internet (and other) mail. Along with the javamail API Sun distribute
some implementations of the API that are based on SMTP, POP and IMAP
respectively.

The OJE project is in the business of developing free versions of as
many of these libraries as possible.

The project was started at sourceforge by Andrew Selkirk
<address@hidden> but Andrew has burned himself out a
bit and has asked me (a contributor to the project) if I would be
willing to suggest to GNU that it become part of this project.

The existing OJE project contains finished, or near finished
implementations of:

- - JAXP
- - Javamail
- - activation
- - infobus

You may ask: why should this not be in the Classpath project?  Here's
some reasons why not:

- - the amount of code in the OJE could easily double the size of the
Classpath project

- - Classpath requires copyright assignment and that might not be
appropriate for OJE

- - the projects are functionally seperate

- - a seperation is not problematic because subversions can be used to tie
the two CVS trees together if necessary

The API deliverables of the project will be LGPL licenced. The
implementation deliverables of the project will be GPL licenced as much
as possible.

Comments:
</summary>


- -- 
Hugo Gayosso
GNU Software Evaluators
Coordinator

http://www.gnu.org
http://wildebeest.myip.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE69zikMNObVRBZveYRArOkAJ4xI2/qkiQy/I1ALwE9BM3whKNEfACdGpTj
yZLqkOVKKBkfHvXaNgSKnM0=
=SRaV
-----END PGP SIGNATURE-----