[Mediawiki-l] ramblings on MediaWiki interface design

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[Mediawiki-l] ramblings on MediaWiki interface design

David Gerard-2
Today I had an interesting experience: talking a fairly elderly
journalist (now mostly an occasional columnist) through creating a
Wikipedia account and creating an article.

Apart from en:wp's Byzantine policy considerations, it was an
interesting experience in finding out in real time just how horrible
the MediaWiki interface is in some ways. Trying to explain where to
find the thing that was on my screen and I *knew* was on his screen,
that sort of thing.

This is the sort of person it would be nice to create something for:
someone who knows a *lot*, can't work a computer and loves Wikipedia
as a reader.

A major part of excellence in technology design is to create something
that lets geeks go wild *and* is entirely usable by people who
basically can't work computers. MediaWiki is pretty good at this
already, I think - en:wp has quite a lot of contributors who can't
work computers but are excellent writers, researchers, editors and
even admins. But there's a long way to go.

Do we have any friendly organisations who can set up interface testing
labs with normal people in them? That relative whose computer you
really, really hate cleaning up for them.

Phil Sandifer posted about this to wikien-l today with regards to
en:wp's grossly newbie-hostile policy thicket:

    http://en.wikipedia.org/wiki/User:Phil_Sandifer/Susan

The text is:

Susan is a hypothetical Wikipedian, selected because she behaves in a
manner basically consistent with most of our editors.

Susan is a 40-year-old stay-at-home mother. She majored in English
many years ago, and still has a fondness for Jane Austen. She is idly
browsing the Internet, and happens by Pride and Prejudice. In five
minutes, her son gets off the school bus. She finds an error on the
page.

Wikipedia policy and process should be designed so that Susan can make
this change and have it not get reverted.

In fact, there ought not be any small task on Wikipedia that cannot be
completed by Susan.

This requires some things.

   1. There must not be any policies or processes that are
sufficiently complex that Susan would have to look them up before
doing anything. Everything should be both memorizable and of
sufficient simplicity that the remembered version will be trustworthy.
That is to say that Susan should be able to get by with the nutshell
versions of our policies.
   2. There must not be a bunch of code or formatting for what she
wants to do. If Susan has to go "Wait, what's the template for this?"
then she will have to get up and go meet her son instead of fixing the
problem.
   3. We must not require anything that Susan does not have fast
access to. No research projects, no scavenger hunts. Not even a Google
search or pulling a book off of a shelf. Susan should be able to
improve Wikipedia on her own.
   4. There must be a culture of good faith so that Susan's
contribution (which will probably come in as an IP contribution) will
not instantly be met with suspicion. Remember - if Susan goes back the
next day and her change has been reverted without explanation, she is
unlikely to edit again.

Think carefully about these issues when designing something for
Wikipedia. Susan is intelligent, well-meaning, and a valuable
resource. She will improve Wikipedia if we let her. And there are
thousands of Susans out there. Susan, or someone with Susan's
circumstances, is our average and most common editor.

Design for Susan.


- d.

_______________________________________________
MediaWiki-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/mediawiki-l
Reply | Threaded
Open this post in threaded view
|

Re: [Mediawiki-l] ramblings on MediaWiki interface design

George William Herbert
On 1/25/07, David Gerard <[hidden email]> wrote:

> Today I had an interesting experience: talking a fairly elderly
> journalist (now mostly an occasional columnist) through creating a
> Wikipedia account and creating an article.
>
> Apart from en:wp's Byzantine policy considerations, it was an
> interesting experience in finding out in real time just how horrible
> the MediaWiki interface is in some ways. Trying to explain where to
> find the thing that was on my screen and I *knew* was on his screen,
> that sort of thing.
>
> This is the sort of person it would be nice to create something for:
> someone who knows a *lot*, can't work a computer and loves Wikipedia
> as a reader.
>
> A major part of excellence in technology design is to create something
> that lets geeks go wild *and* is entirely usable by people who
> basically can't work computers. MediaWiki is pretty good at this
> already, I think - en:wp has quite a lot of contributors who can't
> work computers but are excellent writers, researchers, editors and
> even admins. But there's a long way to go.
>
> Do we have any friendly organisations who can set up interface testing
> labs with normal people in them? That relative whose computer you
> really, really hate cleaning up for them.
>
> Phil Sandifer posted about this to wikien-l today with regards to
> en:wp's grossly newbie-hostile policy thicket:
>
>     http://en.wikipedia.org/wiki/User:Phil_Sandifer/Susan
>
> The text is:
>
> Susan is a hypothetical Wikipedian, selected because she behaves in a
> manner basically consistent with most of our editors.
>
> Susan is a 40-year-old stay-at-home mother. She majored in English
> many years ago, and still has a fondness for Jane Austen. She is idly
> browsing the Internet, and happens by Pride and Prejudice. In five
> minutes, her son gets off the school bus. She finds an error on the
> page.
>
> Wikipedia policy and process should be designed so that Susan can make
> this change and have it not get reverted.
>
> In fact, there ought not be any small task on Wikipedia that cannot be
> completed by Susan.
>
> This requires some things.
>
>    1. There must not be any policies or processes that are
> sufficiently complex that Susan would have to look them up before
> doing anything. Everything should be both memorizable and of
> sufficient simplicity that the remembered version will be trustworthy.
> That is to say that Susan should be able to get by with the nutshell
> versions of our policies.
>    2. There must not be a bunch of code or formatting for what she
> wants to do. If Susan has to go "Wait, what's the template for this?"
> then she will have to get up and go meet her son instead of fixing the
> problem.
>    3. We must not require anything that Susan does not have fast
> access to. No research projects, no scavenger hunts. Not even a Google
> search or pulling a book off of a shelf. Susan should be able to
> improve Wikipedia on her own.
>    4. There must be a culture of good faith so that Susan's
> contribution (which will probably come in as an IP contribution) will
> not instantly be met with suspicion. Remember - if Susan goes back the
> next day and her change has been reverted without explanation, she is
> unlikely to edit again.
>
> Think carefully about these issues when designing something for
> Wikipedia. Susan is intelligent, well-meaning, and a valuable
> resource. She will improve Wikipedia if we let her. And there are
> thousands of Susans out there. Susan, or someone with Susan's
> circumstances, is our average and most common editor.
>
> Design for Susan.


I also left some comments on the talk page for Phil's concept, but...

1) We should work with a trained Information Architect to identify and
document what the key sets of information are which are required to go
from non-editor to making an "appropriate" edit (non-controversial
with experienced editors).  No professional website design would go
any distance without having IAs look at the layout and user access
paths.

2) Second the call for a usability lab review.  I know some usability
lab people but I'm not sure if WP could use those; I will see.



--
-george william herbert
[hidden email]

_______________________________________________
MediaWiki-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/mediawiki-l
Reply | Threaded
Open this post in threaded view
|

Re: [Mediawiki-l] ramblings on MediaWiki interface design

Bugzilla from sy1234@gmail.com
In reply to this post by David Gerard-2
On 1/25/07, David Gerard <[hidden email]> wrote:
> Susan is a 40-year-old stay-at-home mother. She majored in English
> many years ago, and still has a fondness for Jane Austen. She is idly
> browsing the Internet, and happens by Pride and Prejudice. In five
> minutes, her son gets off the school bus. She finds an error on the
> page.

Step one: Click 'edit'
Step two: Type text
Step three: Click 'save'

I think your example is lacking.  You're describing a simple wikignome
type task and not a creation or significant rework of an article.

Most of the policies are "obvious" and are there for reference for
when people stomp all over an article or a user and pretend like it's
alright.

Some of the policies have some subtleties, but everyday editors won't
need to worry about them.

The example being used seems to call for a cleaner interface, in-line
help - maybe also in-line help with tips and policies.

I mean.. if this person edits a page.. people aren't going to go "oh
my god, she fixed my grammar, revert revert!!"


I must conclude this comment with "Is there a real problem which needs
to be solved?"  This thread sounds like great fodder to tie up a
committe and keep people employed for a long time.  It does not sound
useful enough to boil down into a set of concise bug reports or
feature requests.

So the problem needs to be clarified..

Can we find some of these hypothetical editors, and give them a simple
questionnaire maybe?

_______________________________________________
MediaWiki-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/mediawiki-l