correction: PCA issue number is MAHOUT-817.
On Mon, Oct 24, 2011 at 2:30 PM, Dmitriy Lyubimov <[email protected]> wrote: > ok, per this direction,i removed my issues from 0.6 roadmap. > > If i finish any of them (797 50-50, 814 or whatever this pca issue > was, less likely) I will re-insert them to 0.6 roadmap. > > -Dmitriy > > On Mon, Oct 24, 2011 at 1:43 PM, Sean Owen <[email protected]> wrote: >> A closed JIRA isn't gone. It's still there and searchable. Marking it >> WontFix with a note that it's open for reopening seems pretty clear to >> future readers. I suppose we wouldn't know, but, I don't have a sense >> that anyone has ever found a closed JIRA, wanted to work on it, but >> given up because it was closed and they didn't read further. But I can >> point to a hundred cases of the opposite. >> >> If we're just talking about what to call these states, that's good. >> >> The only thing I truly don't like is a false "open" state, the "I'd >> like to think someone else will look at this" state. It seems like >> it's pro-community and some type of useful work, but I think it's the >> opposite. It's the kind of thing that discourages me personally, FWIW. >> >> Well, just leave the "Unversioned" tag as the bucket for everything >> else. That's pretty good. I won't molest it; I might suggest we push >> some things there. >> >> >> Obviously the more important thing is to action some of the important >> changes *that really should happen in a next release*, 0.6. Then file >> some JIRAs for additional things that can and should be done in the >> next month or so. >> >> >> On Mon, Oct 24, 2011 at 7:40 PM, Grant Ingersoll <[email protected]> wrote: >>> My first thought was what's the difference between open/unversioned, but >>> then I think it does require an explicit move which means we've indicated >>> we've looked at it. I do think this is a nice middle ground. >>> >>> >>> On Oct 24, 2011, at 2:12 PM, Dmitriy Lyubimov wrote: >>> >>>> I am really voting for a backlog target. most probably i won't >>>> implement pca idea by end of december but it doesn't mean i am not >>>> committed to see it thru. There probably will be some progress there >>>> if only in form of working notes and some math and discussions. I need >>>> this stuff to be peer reviewed. Why not have a 'backlog' target and >>>> let it live there? >>>> >>>> On Mon, Oct 24, 2011 at 9:59 AM, Jake Mannix <[email protected]> wrote: >>>>> On Mon, Oct 24, 2011 at 5:25 AM, Grant Ingersoll >>>>> <[email protected]>wrote: >>>>>> >>>>>> > - Anything that isn't fixed by December is WontFix and we release 0.6. >>>>>>> >>>>>>> I realize it's drastic, but it's a coherent position. >>>>>> >>>>>> Not at all drastic and perfectly sane. >>>>> >>>>> >>>>> So regarding JIRA management. I see that Benson and Sean come from >>>>> a viewpoint that long-lived open JIRA tickets are a bad sign, while people >>>>> like Grant, myself, and to some degree Ted, are used to seeing open >>>>> tickets >>>>> in an unresolved state that are used as placeholders which tell the >>>>> outside >>>>> observer what has been suggested in the past and what discussions have >>>>> gone on around it, and maybe even has a (currently outdated) patch of >>>>> a proposed solution. >>>>> >>>>> I'm really of the mind that WontFix is meant for "this idea does not fit >>>>> at >>>>> all / >>>>> won't work / and we never intend to do this". Good ideas which we don't >>>>> have the bandwidth for are instead unversioned and left open. I think >>>>> WontFix on an "old ticket" sends a message to the person who opened it >>>>> that we're not interested in their contribution, or if it's a bugfix, that >>>>> we're >>>>> arrogant and don't think they are correct in stating it's an important >>>>> bug. >>>>> >>>>> I'd much rather we find an acceptable unresolved state than always push >>>>> for "0 open JIRA tickets". The Hadoop community also has very long lived >>>>> open tickets with slow progress, it's not just Lucene. I think this is >>>>> healthy >>>>> and a nice way to keep track of what people have thought about in the >>>>> past. >>>>> >>>>> -jake >>>>> >>> >>> -------------------------------------------- >>> Grant Ingersoll >>> http://www.lucidimagination.com >>> >>> >>> >>> >> >
