Hi All, I just wanted to apologize for not responding to all of your feedback. I have read it. I rushed to get this proposal done before I leave this week, despite unexpectedly working late a few nights at my new job. I'm leaving for Boston tomorrow and will return in a week. I'll go over all the messages, test them against the proposal and update it with all of the great ideas. (I may have limited access to the Internet, hence the delay)
When I get back I'll submit two things: * The revised proposal and * A first iteration implementation plan I'm a firm believer in iterative development, and the feature set as I've outlined it is probably more then I'd want to try and do in one swoop. Moreover, I don't want to do any more of this myself then I have to ;-) so its great if I can put stubs out there and have others submit the rest. As far as keeping Lucene an API. That is my every intention. This is new stuff to be added to Lucene as part of the project but in seperate packages. As for creating another project for this..that sounds counter-productive...secondly, the bar on jakarta for creating new projects is pretty dern high and I don't think splitting the Lucene community into really smart people like Doug and people like me who just want to add high level stuff to add to usability is productive. What I'm going to propose in the plan for our first iteration will be a limited set of classes and interfaces for starting this. These will be limited in scope and will keep the full feature set in mind, but not implement it. The whole idea behind this is that it lets us have several crawlers, etc not just one so everyone can contribute to the effort. In the first iteration implementation plan (some of this may be just in-lined patches), I'll propose an ant target or two, some packages, etc. In revising the existing proposal, there is no need to wait for me. I put the XML sources for the proposal at the same url as the html generated version. Just submit patches to it (diff -u original.xml new.xml) to this list. (http://www.trilug.org/~acoliver/luceneplan.xml) I realize the xml is slightly incompatible to the Lucene doc build (it uses Cocoon instead of anakia). I'll be happy to convert it (mostly make the <link> tags into <a> tags) when I get back (or if someone else wants to thats fine) and we can add it to the Lucene docs provided that's agreeable to everyone. I've also included a tarball version of my personal website which is nothing more than an ant-based docbuild that we use for POI and the three webpages (including the lucene plan) with some supporting images. Its about 5.6mb but for those inclined you can download it at (http://www.trilug.org/~acoliver/personal.tar.bz2). The targets are cleandocs and docs. (you can use clean but it will make generation take longer) Once again, thanks for all of the feedback. I'll submit the revised proposal when I get back next Saturday and we'll round us out a nice set of new features. I highly recommend everyone tear at htDig. They've done a great job of providing general functionality. They also DO in fact have an API as part of the project in addition to the "application". It's a nice piece of software...albeit a pain to install and I think Lucene will exceed it all around once we add these features. I look forward to working with you all on this functionality. Thanks, Andy -- www.superlinksoftware.com www.sourceforge.net/projects/poi - port of Excel format to java http://developer.java.sun.com/developer/bugParade/bugs/4487555.html - fix java generics! The avalanche has already started. It is too late for the pebbles to vote. -Ambassador Kosh -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
