Hi Folks, Jerome and I have been thinking a bit about the whole issue of "static" NutchConf, versus removing it and making it a constructor parameter, etc. I personally think that a lot of this issue stems from the fact that the actual source code for nutch, and the what I would call "source distribution", is in the same location at the actual "deployment" distribution. For example, we have a directory structure like:
$NUTCH_HOME: src/ build/ lib/ bin/ ... Which is where you run Nutch, and also where you build Nutch. I think that this would be drastically improved by defining the notion of a "Nutch Deployment". For example, in a lot of my projects at JPL, we check out the source code of our projects from CM, then we construct a "build" of our project. This build becomes what we then "deploy" to a particular deployment environment, or location, and that's where the system is run from. A simple example would be: I have project A, here is A's source code structure: /path/to/A/src/java/my/package/Test.java /path/to/A/build.xml /path/to/A/.... Then, when I type: ant deploy, the following structure is created: /path/to/A/build/distribution/lib/<all lib files needed by the deployment> /path/to/A/build/distribution/bin/<all scripts needed by the deployment> /path/to/A/build/distribution/LICENSE.txt /path/to/A/build/distribution/conf/<any conf files needed by the deployment> ...and so on Then, a user could take the /path/to/A/build/distribution folder, and then copy it to a "deployment" directory, and then, that's the deployment of the system, which is separate from source code, thereby untying the source distribution and the deployment distribution. If we had this concept currently in Nutch, I think a lot of the static Nutch conf issues dissapear, correct, because we have the concept of separate deployments, instead of just relying on the same deployment to run a whole bunch of distributed processes out of. I may be misunderstanding this whole conversation, but if I'm right, then I would propose that we formalize a notion of a "deployment" of Nutch versus the actual "source distribution", instead of co-mingling them. Thoughts? Thanks! Cheers, Chris ______________________________________________ Chris A. Mattmann [EMAIL PROTECTED] Staff Member Modeling and Data Management Systems Section (387) Data Management Systems and Technologies Group _________________________________________________ Jet Propulsion Laboratory Pasadena, CA Office: 171-266B Mailstop: 171-246 _______________________________________________________ Disclaimer: The opinions presented within are my own and do not reflect those of either NASA, JPL, or the California Institute of Technology.
