[Wiki World] a metric for community
Brianna Laugher
I was thinking today that it would be rather good if there was a
standard measure for whether a user is part of a wiki community or
not. It is very hard to compare and even talk sensibly about growth
when we don't really know what we should be discussing.
Maybe this would be interesting for people from Wikimedia, AboutUs,
Wikihow, Wikitravel, Wikia, usemod, etc, to get together and figure
out a half-sensible definition.
My thoughts:
* Wiki community = editing community (not including readers)
* May be useful to distinguish between community and active community
(for latter, insert clause 'in last X months') - or maybe active
community is the only meaningful measure...
* Metric: number of edits? Size of edits? (bytes) interaction with
other users? Number of edits seems to be the most useful measure to
me.
* Should the metric be proportional to the size of the wiki? Is that sensible?
*Do differently-focused wikis have massively different editing
patterns anyway? (eg if you are in a books-type project, is the
average edit size much larger than an encyclopedia-style project or
dictionary-style project?)
Perhaps this is a sign of a community member: one who edits a
community discussion page or another user's talk page.Measurable
interaction with another person.
cheers,
Brianna
--
They've just been waiting in a mountain for the right moment:
http://modernthings.org/
_______________________________________________
wikiworld mailing list
wiki...@lists.ibiblio.org
http://lists.ibiblio.org/mailman/listinfo/wikiworld
Milos Rancic
This is the most important measure. However, if there are no edits, it
is hard to talk about "editing community" :)
> * May be useful to distinguish between community and active community
> (for latter, insert clause 'in last X months') - or maybe active
> community is the only meaningful measure...
Thanks to spammers, it is not so hard to find does some project have
at least people who are taking care about it. However, it doesn't say
if the project is active or not. For example, I am cleaning spam from
not so active language editions of Anarchopedia, but I don't know even
a word in, for example, Japanese.
> * Metric: number of edits? Size of edits? (bytes) interaction with
> other users? Number of edits seems to be the most useful measure to
> me.
There is one interesting implication of this questions: Maybe we
should work on some kind of "project standardization".
All of the projects should have database dumps (daily, weekly or
monthly; SQL or, better, XML). There should be a possibility to put
those dumps somewhere if the project doesn't have enough space for
backups (for example, Internet Archive). And programming methods for
such backups should be developed.
When we have such standardization, we would be able to measure project
activities. Also, all of the "standardized" projects would be properly
backuped; as well as it would be possible to make other kinds of
content analysis. And, of course, the list of backup places should be
listed somewhere.
> * Should the metric be proportional to the size of the wiki? Is that sensible?
I don't think that it is relevant. If English Wikipedia has 2M
articles and 100 edits per minute, it doesn't makes it equal with some
other project which has 20K of articles and 1 edit per minute.
However, maybe it would be possible to make some conclusions in the
future based on some kind of proportion between number of articles,
number of edits per some period of time, number of editors, etc. For
example, I would like to see some graphs of relation between number of
articles and number of edits per minute: is it a constant? If not,
what are the differences? I suppose that such proportion goes down as
number of articles grows (it is very logical consequence: one editor
may make the same number of edits and number of articles is
cumulative).
> *Do differently-focused wikis have massively different editing
> patterns anyway? (eg if you are in a books-type project, is the
> average edit size much larger than an encyclopedia-style project or
> dictionary-style project?)
>
> Perhaps this is a sign of a community member: one who edits a
> community discussion page or another user's talk page.Measurable
> interaction with another person.
I heard that a number of people are interested in at least some kind
of wikimetrics: from scientific researches to more sensationalist
(like wikiscanner). It would be good if it is possible to make some
kind of common resources for wikimetrics. And this is in the range
from making dumps of English Wikipedia more useful (for example, I
would like to see unzipped dumps in the sensible portions up to 1GB)
up to the sophisticated methods for analysis of dumps of different
wiki engines.
Angela
> My thoughts:
>
> * Wiki community = editing community (not including readers)
Although the readers aren't part of the community, they're still
important in many wikis. Although some wikis may be aimed purely at
the editing community, others are written with the intention of being
read, so the ratio of readers to editors is interesting in
highlighting how well the wiki is promoted beyond it's editing group
and how good it is at drawing in editors from that groups of readers.
> * Metric: number of edits? Size of edits? (bytes) interaction with
> other users? Number of edits seems to be the most useful measure to
> me.
I'd like to add "range of edits". Some wikis have very different
editing patterns where someone will regularly edit a single page or
small group of pages but never even look at the rest of the site. I'm
guessing AboutUs might be one example where many (most?) people keep
an eye on the page(s) about their own website but aren't part of the
community and have no involvement in the meta side of the project
(developing policies, reverting vandalism on pages that aren't
"theirs", interacting with other users, etc).
Angela
Brianna Laugher
> Thanks to spammers, it is not so hard to find does some project have
> at least people who are taking care about it. However, it doesn't say
> if the project is active or not. For example, I am cleaning spam from
> not so active language editions of Anarchopedia, but I don't know even
> a word in, for example, Japanese.
OK, so another interesting thing is: how many *active community
members* are needed for an *active community*?
Is there more to an active community than a collection of active
community members?
> > * Should the metric be proportional to the size of the wiki? Is that sensible?
>
> I don't think that it is relevant. If English Wikipedia has 2M
> articles and 100 edits per minute, it doesn't makes it equal with some
> other project which has 20K of articles and 1 edit per minute.
That's not what I meant. What I meant was maybe on a tiny wiki, 5
edits makes you an active community member. On English Wikipedia,
maybe it is 100? Or is it still the same 5?
regards,
Brianna
--
They've just been waiting in a mountain for the right moment:
http://modernthings.org/
Brianna Laugher
> On 9/29/07, Brianna Laugher <brianna...@gmail.com> wrote:
> > My thoughts:
> >
> > * Wiki community = editing community (not including readers)
>
> Although the readers aren't part of the community, they're still
> important in many wikis. Although some wikis may be aimed purely at
> the editing community, others are written with the intention of being
> read, so the ratio of readers to editors is interesting in
> highlighting how well the wiki is promoted beyond it's editing group
> and how good it is at drawing in editors from that groups of readers.
Yes. OK, so we have a set of useful metrics, rather than a single measure. :)
> > * Metric: number of edits? Size of edits? (bytes) interaction with
> > other users? Number of edits seems to be the most useful measure to
> > me.
>
> I'd like to add "range of edits". Some wikis have very different
> editing patterns where someone will regularly edit a single page or
> small group of pages but never even look at the rest of the site. I'm
> guessing AboutUs might be one example where many (most?) people keep
> an eye on the page(s) about their own website but aren't part of the
> community and have no involvement in the meta side of the project
> (developing policies, reverting vandalism on pages that aren't
> "theirs", interacting with other users, etc).
Yes, I agree that seems likely. Compare to Wiktionary where I imagine
an editing range is likely to be much more broad.
Sites like Wikipedia are more in the middle, perhaps. Maybe people are
drawn in by one initial topic and discover wikicrack and turn to a
wider range of articles. (I know I have devoted a lot of time to
topics I don't actually care about simply because I was wikiaddicted.)
Another metric: random spread. If one successively hits special:random
a lot, how likely is it similar topics will be stumbled across? (so, a
documentation wiki, installing is not similar to customisation, for
example.) What does the graph of relatedness look like.
Possibly that's not the right question. A 'page' means different
things to different people. Wikibooks and Wikisource have pages as
part of larger works. They are intended to be read in context.
Wiktionary and Wikipedia and AboutUs are "dive in, start anywhere"
style projects.
random musing...
cheers,
Brianna
--
They've just been waiting in a mountain for the right moment:
http://modernthings.org/
Nick Jenkins
> members* are needed for an *active community*?
Perhaps:
One active member = a soliloquy
Two active members = a conversation
>= Three active members = a community
(i.e. maybe an early sign of a community is that someone says something, then
someone corrects them, and then someone else tells them that they're both wrong!)
When a community goes from being a community to an *active* community is
pretty subjective. Maybe it's when sub-communities start to appear? I.e. when
the original single community is sufficiently broad and noisy that people feel the
desire to break it out into different parts / interests.
> Is there more to an active community than a collection of active
> community members?
Yes - interaction between the members. A site with people working off on separate
areas and never interacting is not an active community. It's just a group of people
who happen to share a wiki.
> > > * Should the metric be proportional to the size of the wiki? Is that sensible?
> >
> > I don't think that it is relevant. If English Wikipedia has 2M
> > articles and 100 edits per minute, it doesn't makes it equal with some
> > other project which has 20K of articles and 1 edit per minute.
>
> That's not what I meant. What I meant was maybe on a tiny wiki, 5
> edits makes you an active community member. On English Wikipedia,
> maybe it is 100? Or is it still the same 5?
Isn't that partially also a metric for how "sticky", and how much interesting content,
and how old the site is? Also maybe people see themselves relative to other people
in terms of determining whether they're an active community member or not - so
5 edits on a very small wiki probably does make you an active member, and on EN
Wikipedia it probably is 100 edits - because "active" can be a relative measure -
i.e. active compared to whom - and I think the normal answer is "compared to most
other people using the wiki".
-- All the best,
Nick.
Ben Yates
by users who didn't register or come back to edit again.
--
Ben Yates
Wikipedia blog - http://wikip.blogspot.com
Brianna Laugher
Also maybe people see themselves relative to other people
> in terms of determining whether they're an active community member or not - so
> 5 edits on a very small wiki probably does make you an active member, and on EN
> Wikipedia it probably is 100 edits - because "active" can be a relative measure -
> i.e. active compared to whom - and I think the normal answer is "compared to most
> other people using the wiki".
Hmm...a relative measure may not be that useful. I was thinking
"maybe: the number of people who make more edits than 90% of the
total... oh wait, that's always going to be 10% of the total." :)
That never tells you if the active community as a proportion of the
total community is increasing or decreasing.
or maybe I am thinking about it slightly wrong.
Brianna
--
They've just been waiting in a mountain for the right moment:
http://modernthings.org/