[Foundation-l] The problem with Incubator: An interactive journey

[M. Williamson]

To anyone who's interested... on a request to close a Wiki on meta, we were
just debating the problems that exist with Incubator, when one of the
bureaucrats from Incubator asked something along the lines of "wait, there's
a problem with Incubator?" So I went on a little quest to prove how
difficult it is for users, particularly those inexperienced with Wikimedia
projects (or even with computers and the internet, as the case may be), to
even *find* the test-Wiki in their language.

So if you'll indulge me briefly, and follow along at the site if you'd like,
I'd like to go on a brief journey through just how many clicks it takes to
find a particular test Wiki (in this case I chose Central Morocco Tamazight
Wikipedia because it has quite a few speakers). Also I'd like to note that
this is just my commentary and perspective, maybe you'll actually find it
quite easy and disagree with me. So let's begin step by step:

1. You go to http://incubator.wikimedia.org/ (what are the chances a person
would even know to go to the incubator site in the first place? we don't
exactly advertise it; it's not linked to from any high-visibility page)

2. You scroll down. You look for awhile. Unless you're looking for Kichwa,
Ingush, Yucatec Maya or Tashelhit Wikipedias, you get confused and you look
some more. Then you _finally_ notice that there is a link to "incubating
Wikis" (what does that even mean? yuck. poorest choice of link label ever!)

3. You scroll through walls of text looking for your language. We're on at
least the 3rd click by now. You must look through a long list of languages
you probably don't even know are languages... how do you know you're even in
the right place? Also, rather than sorting by project popularity, or putting
the projects on separate pages, Wikibooks is first. THEN Wikinews. Then
finally, our most popular project (by FAR), Wikipedia.

4. It's probably been about 5 minutes already (if you're a novice user, that
is), you're getting bored of this searching. OK, whatever. Then you
''finally'' find your language. Let's pick the Central Morocco Tamazight
test-WP, which has lots of speakers but is near the bottom of a VERY long
list of test-Wikipedias.

5. What's this? There are 5 different links to click. Chances are, you
either didn't notice or you no longer remember what the column labels said,
so you have no idea why there are 5, or what the difference is. So at this
point, your chance of clicking the right link rather than just giving up is
1/5. Also, the most intuitive link to click, the language name + Wikipedia
("Central Morocco Tamazight Wikipedia"), leads to the PROPOSAL on meta,
*NOT* the test wiki!

6. Let's say somehow, you figure out which link is the right one to click.
Even for the tech-savvy, this takes 6 clicks already and is somewhat of a
usability nightmare. Then, once you click it, where do you land? Is it the
main page of the test wiki? Please? NOPE! It is a stupid, pointless, totally
useless "splash page" that serves no purpose to users, only to incubator
admins and nobody else.

7: If you made it to this point as a casual user without this guide, you've
probably used Incubator before when it was _slightly_ more user friendly
(they recently redesigned it to be even LESS usable, I didn't think that was
possible but they surprised me!), or you're just very determined. Anyhow,
thanks to certain editors (like one that creates nonsense or poor-quality
pages in many Mexican indigenous languages he doesn't speak at all), there's
a good chance that the page you do end up at if you find your language will
be totally useless and worthless, possibly filled with fake writing that
isn't actually in your language, or is supposedly in your language but makes
no sense at all. So even if you made it to this point, there's still a good
chance you'll get discouraged and leave!

Now add to this the confusing process for starting a new test-wiki, the
confusing and non-transparent process of adding new pages successfully,
etc., and it becomes clear the system is broken. Not that this means much,
but I'd like to note that I was the one who originally suggested the idea of
"test Wikis" way back in Febuary 2005, and have closely watched developments
since then with great disappointment.

-m.