SQL is a nice formal definition for these metrics.  Even though it may
become out of date (the prod database doesn't change in backwards
incompatible ways that often), it's still a big win for the metrics pages
IMO.

-Aaron

On Wed, Oct 14, 2015 at 2:55 PM, Nuria Ruiz <[email protected]> wrote:

> >Keep in mind that, when I say "metrics documentation", I'm not referring
> to documentation about Hive
> <https://wikitech.wikimedia.org/wiki/Analytics/Cluster/Hive>, the webrequest
> logs <https://wikitech.wikimedia.org/wiki/Analytics/Data/Webrequest>, or
> EventLogging <https://wikitech.wikimedia.org/wiki/Analytics/EventLogging>.
> >To my mind, those are infrastructure topics that are relevant mainly to
> Wikimedia (not MediaWiki) engineers, and so belong on Wikitech.
> Agreed
>
> >I'm talking about documentation relevant to analysts, researchers, and
> end users of metrics: "this is how we define an *edit*", "this is why we
> use 5 edits as the cutoff for an active editor", "this is sample SQL for
> counting surviving new active editors",
> Agreed for all but last one, the sql will change. I will keep meta only
> about definitions. If you start adding "examples" , it is likely they soon
> will be out of date.
>
>
>
>
> On Wed, Oct 14, 2015 at 11:30 AM, Leila Zia <[email protected]> wrote:
>
>> Makes sense to me.
>>
>> On Wed, Oct 14, 2015 at 11:27 AM, Neil P. Quinn <[email protected]>
>> wrote:
>>
>>> Keep in mind that, when I say "metrics documentation", I'm not referring
>>> to documentation about Hive
>>> <https://wikitech.wikimedia.org/wiki/Analytics/Cluster/Hive>, the webrequest
>>> logs <https://wikitech.wikimedia.org/wiki/Analytics/Data/Webrequest>,
>>> or EventLogging
>>> <https://wikitech.wikimedia.org/wiki/Analytics/EventLogging>. To my
>>> mind, those are infrastructure topics that are relevant mainly to Wikimedia
>>> (not MediaWiki) engineers, and so belong on Wikitech.
>>>
>>> I'm talking about documentation relevant to analysts, researchers, and
>>> end users of metrics: "this is how we define an *edit*", "this is why
>>> we use 5 edits as the cutoff for an active editor", "this is sample SQL for
>>> counting surviving new active editors", and so on. I think that kind of
>>> information belongs on Meta (and not on mediawiki.org, which was the
>>> original thrust of my question).
>>>
>>> Does that seem like a sensible split to people, or am I just agreeing
>>> with one side of the debate?
>>>
>>> On Wed, Oct 14, 2015 at 8:32 AM, Aaron Halfaker <[email protected]
>>> > wrote:
>>>
>>>> I think having documentation in more than one place is an awful
>>>>> experience for newcomers.
>>>>
>>>>
>>>> Which newcomers are you referring to?  Newcomers to the WMF engineering
>>>> staff or newcomers to research/analytics of Wikimedia projects?
>>>>
>>>> It's OK to not understand the different purposes of our Wikis right
>>>> away, but I don't think that is a good reason to undermine their purposes.
>>>> I certainly don't see why wikitech is a desirable hub for this kind of
>>>> information.  From my point of view Wikitech is the *worst* potential hub
>>>> of information that is not specific to engineering.
>>>>
>>>> What, exactly, is the trouble with having metrics documentation on
>>>> Meta?   How would moving *some of the the documentation* to wikitech help
>>>> that?  (Because you're not going to move research project documentation
>>>> without even stronger disagreement from the locals.)
>>>>
>>>> -Aaron
>>>>
>>>> On Wed, Oct 14, 2015 at 11:05 AM, Dan Andreescu <
>>>> [email protected]> wrote:
>>>>
>>>>> Strongly oppose moving the Research namespace hosted metrics
>>>>>> documentation off Meta. It's s'posed to be broadly accessible. Wikitech 
>>>>>> is
>>>>>> on few peoples' radar. Mediawiki.org is for software documentation. Meta 
>>>>>> is
>>>>>> the central wiki for the movement (however imperfect). - J
>>>>>>
>>>>>
>>>>> I respect the fact that these kinds of distinctions make sense to
>>>>> people who are already familiar with the movement and research /
>>>>> analytics.  But to someone relatively new, and to me for the first year at
>>>>> the foundation, those distinctions made zero sense.
>>>>>
>>>>> I'm not saying it's easy, but I think having documentation in more
>>>>> than one place is an awful experience for newcomers.  We'll continue to
>>>>> move things to wikitech and leave nice high level landing pages on the
>>>>> other wikis.  Others are welcome to act differently if they so see fit.  I
>>>>> know a lot of research stuff is on meta, so maybe in your case it makes
>>>>> sense to standardize on meta and point to it from the other wikis.
>>>>>
>>>>> You're of course welcome to disagree with me but I'd suggest first
>>>>> trying to come up with examples of newcomers who understand the purpose of
>>>>> our different wikis perfectly right away.
>>>>>
>>>>> _______________________________________________
>>>>> Analytics mailing list
>>>>> [email protected]
>>>>> https://lists.wikimedia.org/mailman/listinfo/analytics
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Analytics mailing list
>>>> [email protected]
>>>> https://lists.wikimedia.org/mailman/listinfo/analytics
>>>>
>>>>
>>>
>>>
>>> --
>>> Neil P. Quinn <https://meta.wikimedia.org/wiki/User:Neil_P._Quinn-WMF>,
>>> product analyst
>>> Wikimedia Foundation
>>>
>>> _______________________________________________
>>> Analytics mailing list
>>> [email protected]
>>> https://lists.wikimedia.org/mailman/listinfo/analytics
>>>
>>>
>>
>> _______________________________________________
>> Analytics mailing list
>> [email protected]
>> https://lists.wikimedia.org/mailman/listinfo/analytics
>>
>>
>
> _______________________________________________
> Analytics mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/analytics
>
>
_______________________________________________
Analytics mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/analytics

Reply via email to