Identifying and inviting top contributors of other support channels


IMO the next logical step would be to find the people who contribute a lot to other support channels (#mediawiki, Support desk, wikitech-l, Stackoverflow…) and invite them to participate here (or find out why they don’t want to). What do you think?


Agreeing, but not sure how to gather stats (if it’s about specific names and not just a public call) for Support desk and Stackoverflow. For IRC and the mailing list, should help.


For StackOverflow (or Stack Exchange if we want to be thorugh) via the data explorer, for Support Desk just count the contributions in the DB (or find a tool that can handle Flow but I didn’t).


For onboarding part, we could consider sending personalized invites via the discourse’s admin interface. If there is an easy way to identify top contributors / maintainers of projects from within the volunteer community, then targeting them first could be a good first step. At WMF, it might be good first to reach out to team managers.


I never knew about! It’s cool.

Here (with username links) are the usernames of the people who answered the most questions tagged ‘mediawiki’, across all (non-Meta) StackExchange sites (with 20 or more answers):


@samwilson nice, thanks!

If you want to make it more generic, a decent list of Wikimedia-related tags is mediawiki, mediawiki-api, mediawiki-templates, mediawiki-extension, mediawiki-extensions, mediawiki-installation, pywikipedia, semantic-mediawiki, wikidata, wikidata-api, wikipedia, wikipedia-api, wikimedia-dumps, wikimedia-labs.


Flow doesn’t seem to be integrated to the history API so I wrote a script for scraping the history page:

tgr@tgr-ThinkPad:~$ cat 

while [[ -n $URL ]]; do
    >&2 echo $URL
    curl -s "$WIKI$URL" | xpath -q -e '//div[@class="flow-board-history"]/ul/li/a[@class="mw-userlink"]/bdi/text()' 2>/dev/null
    URL=`curl -s "$WIKI$URL" | xpath -q -e 'string(//a[@class="mw-nextlink"][1]/@href)' 2>/dev/null`
done | sort | uniq -c | sort -rn

tgr@tgr-ThinkPad:~$ ./ | awk '$1 >= 100' 
   4989 Ciencia Al Poder
   2681 MarkAHershberger
   1225 Bawolff
   1158 Johnywhy
    997 Florianschmidtwelzow
    878 Flow talk page manager
    623 Malyacko
    611 Subfader
    578 AKlapper (WMF)
    502 星耀晨曦
    388 Krinkle
    383 TheDJ
    338 Tropicalkitty
    332 Lieutenant S. Reznov
    316 MaxSem
    291 Mainframe98
    286 Jasper Deng
    259 MacFan4000
    257 Wargo
    248 Zoglun
    239 Emufarmers
    235 Reach Out to the Truth
    230 Fokebox
    194 Reception123
    150 Waanders
    150 Clump
    146 Jackmcbarn
    143 Stefahn
    135 Matiia
    133 Shirayuki
    133 Cavila
    131 Allen4names
    122 External Link to Interwiki (Bot)
    121 Ricordisamoa
    108 FlightTime
    106 This, that and the other


It’s doing any tag containing mediawiki (i.e. q.Tags LIKE '%mediawiki%') so should be getting all of them.

And nice work with the history scraping. :slight_smile:


@samwilson I was mainly thinking of the wikipedia-api tag which is fairly popular and covers the Wikipedia mashup developer population that AIUI is a target group of DevRel and this site alongside developers creating tools for Wikimedians to use.


Invites do not work with local logins disabled, and in most cases we don’t have the email address anway (and where we do we can just mail them). For MediaWiki users we can use sendBulkEmails.php. I guess we can just ping top IRC users manually. There is no official way to contect StackOverflow users - we could look at the user profiles and hope they have email addresses, or maybe buy an advertisement for the mediawiki tag.


(Aside: another cool place that’s hard to find is stackapps.)


Good point! I read too hastily. I’ve updated the table above to include mediawiki, wikipedia, and wikimedia tags, as well as links to people’s profiles — it looks like quite a few have some sort of contact info there.


For the records, in some cases, email addresses might be in the database behind that srishakatux and I can access.