[WikiEN-l] errors versus errors (was: Don't remove a WP:OFFICE tag put there by Danny)
Delirium
delirium at hackish.org
Sun Mar 12 07:55:28 UTC 2006
Delirium wrote:
>The more such ancillary concerns influence things, the worse for the content, in my opinion.
>
>
I thought I'd elaborate a bit on this, because I think various people
are operating under fundamentally different assumptions of how Wikipedia
should operate, and how its operation affects things outside itself.
From my perspective, the sole goal of Wikipedia should be to produce an
encyclopedia that is as complete and accurate as possible. Any policy
changes that contribute to that are welcome. For example, we could
decide that Wikipedia is getting big enough to warrant being more
judicious about what we write: require more stringent referencing, and
remove (at least temporarily) unreferenced facts, especially ones that
seem questionable. That's starting to happen already. We could also
better label how good specific articles (and revisions) are. That would
allow us to have good articles interspersed with work-in-progress
articles, while keeping errors in the work-in-progress versions from
being too damaging to the overall trustworthiness of our information
(because those articles would be clearly marked as "in progress"). We
have tags that do that to some extent (dispute tags, references-needed
tags, etc.), and the long-anticipated sifter project (or whatever it's
called now) would do it to a greater extent.
There is another viewpoint: that we should take into account concerns
other than the overall accuracy of the encyclopedia. On this account,
inaccurate negative information about living people is worse than other
types of inaccurate information, and so should be treated specially. I
disagree with that. It's certainly a direct case where harm could come
about, but there are many other cases where much more harm could come
about. Yes, if it's inaccurate, someone could read our article on
[[Jack Thompson (lawyer)]] and get unwarranted negative views of him.
But if we have inaccurate articles, someone could also read our articles
on the Israeli-Arab conflict and come away with unwarranted views of
*that*. The latter sort of inaccuracy has the potential to have a much
greater negative effect on many more people than defaming Jack Thompson
ever could.
There are a lot more examples, but the main point is that there are a
*lot* of potential errors that could cause real-world problems if people
read them and believe them. Defamatory information on specific
individuals is neither the only nor the most problematic type of error.
Therefore, trying to make fine-grained decisions about which errors are
worse than others, and setting up special processes to deal with the
ones deemed to be worse, is entirely the wrong approach, and likely to
lead *neither* to a more accurate encyclopedia on the whole, *nor* to a
reduction in the negative effects inaccuracies in Wikipedia have on real
people. A better approach, in my view, is to attack the problem directly.
We don't want errors at all. Referencing standards and the like attack
this problem, and perhaps there are additional things we could change.
Since a work-in-progress is rarely error-free, we also want to better
mark how trustworthy particular revisions should be considered, which
will reduce the negative impact of any inaccuracies in the versions
clearly marked "hey this might not be right!". Various tags and a
future sifter project address this issue. And, finally, we want a
streamlined way to vet facts. A number of things address this problem:
a more intuitive referencing system (with references attached to
specific bits of information); community culture about referencing and
dealing with controversies; facilitation by the Foundation to pass along
inaccuracies reported by people who don't edit Wikipedia; and so on.
Of course, legal issues may intervene in some cases; if the Foundation
is the target of legal threats or lawsuits, then they can do what their
lawyers advise them. The main issue at hand, though, is what we should
do in other cases.
-Mark
More information about the WikiEN-l
mailing list