mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lance Norskog <>
Subject Re: Use of JIRA was [jira] [Resolved] (MAHOUT-752) Semantic Vectors: generate and use vectors from User/Item Taste data models
Date Tue, 19 Jul 2011 00:54:39 GMT
Looking at the Solr and Hadoop JIRA's, I'm wary of "JIRA pollution".
Also, patches that don't get committed are left behind by code changes
and are just a pain in the neck.

Github has URLs for specific directories and files. I'm fine with
posting pointers to them. Also, I have my own tech blog
( specifically so that I can post pictures
for longer explanations.

On Mon, Jul 18, 2011 at 10:52 AM, Sean Owen <> wrote:
> I'm fine with that. No problem in putting stuff in JIRA per se -- it's
> just making sure we preserve JIRA's role as a workflow tool. That is,
> as you say, if such issues were marked early as WontFix or something
> similar that would solve it; we wouldn't have issues floating around
> that appear to want review and committing when they don't.
> On Mon, Jul 18, 2011 at 5:53 PM, Grant Ingersoll <> wrote:
>> I don't think this is quite right.  If he puts it up on Github, then we have to
go searching for it and it is not officially part of the project.  JIRA is a perfectly fine
place to explore ideas with real patches, which, at the end of the discussion may result in
a "won't fix", but it shouldn't be decided up front just because it is still being thought
through.  Naturally, a discussion on dev@ might be warranted first if no code is involved,
but by putting it in JIRA we have captured: the idea, the code, and the contributor's intent
to donate it.   Someone else or Lance may very well come along and improve on it such that
it does become commit worthy.  Iterating on ideas in JIRA should be encouraged and I think
Lance let everyone know right up front that they can choose to ignore this for now if they
are so inclined so as to not waste anyone's time.

Lance Norskog

View raw message