Also I think that discussion around the use of a feature should take place
on the user list, instead of the dev list. Users are the intended audience
for that kind of question.

- Houston

On Mon, Aug 31, 2020 at 5:21 PM Houston Putman <[email protected]>
wrote:

> Thanks for bringing this up David.
>
> To start, I definitely agree that removal of contribs should have a
> discussion, then a vote before any action is taken.
>
> There have been a few bug fixes/improvements made this year, but there has
> not been a lot of development on it.
> If there is little use of the component, then I agree taking it outside of
> lucene-solr is probably a good idea. If it does have use in the community
> and stays as a contrib, I think it is still a good idea to turn it into a
> "package". That way it is easily loadable by users.
> But I guess that is a part of a larger discussion about how we should
> treat contrib modules, and whether they should be "first party" packages.
>
> As per your question around compatibility with JSON Facets, there is
> definitely overlap but differences in the options provided by each. I'm not
> an expert on JSON Facets, but there are certainly things that each module
> can do that the other can't.
> Ideally we could take the functionality unique to analytics and add it to
> JSON facets, that way there is little reason for the analytics component to
> remain. But that is a larger effort.
>
> - Houston
>
> On Mon, Aug 31, 2020 at 4:43 PM David Smiley <[email protected]> wrote:
>
>> I'd prefer discussion about removal of contribs (or major parts of Solr)
>> occur here first with a vote, and _then_ file a JIRA if the conclusion is
>> to remove it.  It's more visible to each other and the community.
>>
>> ~ David Smiley
>> Apache Lucene/Solr Search Developer
>> http://www.linkedin.com/in/davidwsmiley
>>
>>
>> On Mon, Aug 31, 2020 at 4:14 PM Ishan Chattopadhyaya <
>> [email protected]> wrote:
>>
>>> +1, please open a jira and I'll take care of it.
>>>
>>> Fyi, Marcus is helping me with migrating contrib modules to first party
>>> package, and this was on our radar. I am totally occupied in other things
>>> lately and hence slowed down on that effort. Will pick up pace end of this
>>> week.
>>>
>>> On Tue, 1 Sep, 2020, 12:32 am David Smiley, <[email protected]> wrote:
>>>
>>>> Our new package system allows for a viable/healthy ecosystem of 3rd
>>>> party packages that are easy to find & install.  I wonder if the Analytics
>>>> contrib module should be 3rd party?  I suspect it is very rarely used,
>>>> especially with the relative strength and growth of Solr's built-in JSON
>>>> facet module.  I've never used it.  I raised these points in this comment 2
>>>> years ago:
>>>>
>>>> https://issues.apache.org/jira/browse/SOLR-12045?focusedCommentId=16437754&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16437754
>>>> My biggest point there is the redundancy Solr has with competing ways
>>>> of doing facets.
>>>>
>>>> I ran "cloc" on a clean checkout of this dir:
>>>>
>>>> github.com/AlDanial/cloc v 1.82  T=0.72 s (416.9 files/s, 69593.9
>>>> lines/s)
>>>>
>>>> -------------------------------------------------------------------------------
>>>> Language                     files          blank        comment
>>>>     code
>>>>
>>>> -------------------------------------------------------------------------------
>>>> Java                           290           5950           8886
>>>>    35035
>>>> XML                              6             76            123
>>>>      186
>>>> HTML                             4              8             59
>>>>       38
>>>> Ant                              1              7             14
>>>>        7
>>>> Gradle                           1              3             16
>>>>        6
>>>>
>>>> -------------------------------------------------------------------------------
>>>> SUM:                           302           6044           9098
>>>>    35272
>>>>
>>>> -------------------------------------------------------------------------------
>>>>
>>>> 35k lines of code is a lot of code to maintain.  AFAIK there haven't
>>>> been bugs/features reported against it in a long time.  Maybe it would be
>>>> maintained better if it was a 3rd party package?  There is some maintenance
>>>> cost as refactors touch it.  Maybe not "a lot" but it's something.
>>>>
>>>> As a counterpoint to all this, if maintenance cost is low and it's
>>>> already where it is (thankfully not in Solr-core), and if it's not
>>>> redundant with JSON Faceting module (?), and if it's used (outside of
>>>> Bloomberg, who contributed it) (?), then let it be.
>>>>
>>>> ~ David Smiley
>>>> Apache Lucene/Solr Search Developer
>>>> http://www.linkedin.com/in/davidwsmiley
>>>>
>>>

Reply via email to