On Fri, May 16, 2014 at 1:25 PM, Erik Moeller e...@wikimedia.org wrote:
On Fri, May 16, 2014 at 1:28 AM, Pete Forsyth petefors...@gmail.com
wrote:
I think it is much more likely that a Wikipedia reader would expect to
find
those images *used in Wikipedia articles* than a massive collection of
stuff that is somehow tangentially related to Wikipedia in a way that
they
don't fully understand.
So why on earth does the main multimedia search link on Wikipedia
automatically return unused results from Commons to begin with? Is that
really the right way to go?
I'm breaking out this question since it's a concrete technical
proposal; it should probably also be raised on the multimedia list.
But we should answer it from the perspective What's best for the
user, rather than have it be driven solely by the NSFW corner case
(which may also appear when searching images used on a project like
en.wp alone).
As a user, I might want to find images to add to an article. Having
results from the central repository presented locally makes it easier
to do so without visiting a separate site. (Consider this from the
perspective of smaller projects especially, where the local search
would be pretty much useless.) This is why VisualEditor presents
Commons search results, as well.
As a user, I might be interested in multimedia about a certain topic I
just read about in Wikipedia. Showing only the results already in the
Wikipedia article(s) about the topic would make it harder to find such
media. Simple example: Let's say I read an article about a city, and I
want to find other historic maps of that city. In many cases, these
maps do exist on Commons but not locally. Should we therefore force
people to visit Commons to find them?
I would argue that from the ordinary user's perspective, the
distinction between Wikipedia/Commons is less important than what they
have in common, i.e. being large repositories of useful educational
content (and hyperbole aside, 99% of Commons is pretty boring stuff).
We could default to displaying locally used media and offer to search
Commons with an extra click. From a usability perspective, you want to
minimize the steps a user has to take, so good UX design would likely
disclose results from Commons either a) always, clearly labeled or b)
when no local results are available.
There's no question that search UX, both on Commons and on Wikipedia,
can be improved. I'm just skeptical that an unbiased evaluation of the
user experience using standard UX heuristics would lead to a design
that hides explicit content from initial search results. Distinguish
different types of content more clearly and make it easier to find the
stuff you want - sure, that's doable.
With Cirrus (the New Seach beta feature) both marking stuff from commons
as from commons and adding a tick box to remove stuff from commons would
also be possible. Marking would be easier but both wouldn't be too hard.
We could even artificially push results that are on the local wiki higher
then those on commons.
I know its a horrible excuse, but we're spinning up a project to rework the
search page's user interface. It hasn't something like that since tables
in table in tables was normal. We've been talking about this. The project
is still in the those a pretty mock ups stage and it has been moving
slowly.
Nik
___
Wikimedia-l mailing list, guidelines at:
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
mailto:wikimedia-l-requ...@lists.wikimedia.org?subject=unsubscribe