I uploaded the pretty printer java program to the source auditor ftp server.  It’s a secure web server, so I apologize in advance if it’s a bit inconvenient to download.
 
The ftp server is at ftp.sourceauditor.com <ftp://ftp.sourceauditor.com>   You need to use explicit tls/ssl over port 21.  Logon with user spdx and password spdx1
 
The file SPDXPretty.zip contains the files mentioned in the previous email (copied below).
 
Let me know if you need more information or if you have any problems.
 
Gary

 

From: package-facts-bounces@... [mailto:package-facts-bounces@...] On Behalf Of Gary O'Neall
Sent: Sunday, August 08, 2010 11:32 PM
To: package-facts@...
Subject: Java Pretty Printer

I completed an “alpha” version of a Java based pretty printer.  It’s 10MB in binary form with its dependencies.  Is there a place on the Wiki I can upload this to?  I tried to add it to a page as an attachment to a new discussion page, but the .zip filetype was not allowed.  Please advise on the best method to get this to the group.
 
Attached is a modified zlib example (see notes below on what items were changed) and an example output.
 
Below is some information and discussion points related to the pretty printer development:
 
I’m sure there are a few improvements to be made before calling this a “release”, but it  does provide some formatting and works for the zlib example.  I would appreciate any feedback once you have access to the application.
 
To run the application, make sure you have a JRE 1.6 installed (JRE version 1.5 may work, but it untested).  Unzip the files in your favorite directory.  Execute the jar file with a single text parameter of a file path for the SPDX RDF Document.
 
On windows, this would be “java –jar SPDXPretty.jar examples\zlib-1.2.5.spdxv3.rdf (assuming you copied the attached example file into the same directory as the .jar file and your cd’d to that directory).
 
I made a few changes to the zlib example to bring it up to date to the draft 20100731.  It is in the zip file in the examples directory.
 
I run into a few questions/issues as I implemented this, outlined below:
 
·        Namespace and tags – I noticed in the example we have only one namespace for SPDX and the tags used in the example did not match the tags in the specification in all cases  - e.g. License in the file is tagged FileLicense in the example.  Do we want to have separate namespaces for File, License, and Document?  If not, do we want the tags to be unique (e.g. FileLicense and PackageLicense)?   Technically, the tags don’t need to be unique, but it may aid in humans reading the RDF/XML file.

·        I changed the tags in the example to match the document in cases where they were still unique (e.g. ShortDescription -> ShortDesc)

·        License Names and Pretty Printing – I was only able to extract the URL for the license (as a resource) from the SPDX document which doesn’t lead to a very pretty license name.  Should we add a property License Name?  Should I parse the URL and only print out the tag (e.g. after the #)?

·        Example use of hasFile – In the example, the object of the hasFile predicate for the package subject all have the same URI.  I believe these should be unique since they represent different file objects.  I changed the example to make these individual and unique.

·        The disjunctive licenses are implemented but not tested.

·        There has not been much testing (Unit or otherwise)

 
I would like to make the code available as an open source project.  It is written using Jena (http://jena.sourceforge.net/) and contains a Java class which is a model basically wrapping a Jenna model of the RDF document.  It would probably be useful for many of you who are writing tools.  
 
I could post the code to SPDX, but I would rather maintain it in a repository which supports svn.  I’m thinking Google code may be a good location.  Open to suggestions.
 
As far a licenses, it’s currently under a 3 clause BSD since it’s GPL compatible and simple.  I’m open to other licenses, so let me know if you have a preference – we could even create a nice complex set of license choices ;)  Do keep in mind this is dependent on Jena which is licensed under a 3 clause BSD and contains some Apache licensed code.
 
Appreciate any comments.
 
Best regards,
 
Gary O’Neall
Source Auditor Inc.