[Sigia-l] well defined IA activities?

[Eric Scheid]

In a recent blog posting, Gene Smith writes:

    Tasks that have well-defined inputs and outputs and can be
    completed remotely --like content inventories, card sorts,
    indexing, metadata and thesaurus development-- are obvious
    candidates for outsourcing.

    http://atomiq.org/archives/2004/12/outsourcing_iaux.html

I think he's got a point here, although I'd like to explore it some more.

Firstly, do we in the IA field have a clear idea on just which IA tasks have
well defined inputs/outputs, or for that matter well-defined processes? What
are some of the tasks/activities which are more, um, fuzzy ... based on
heuristics and general rules of thumb, or are really "flying by the seats of
your pants" opinion/designer/consulting efforts?

Secondly, what other factors are there that affect whether an activity could
be outsourced? Are activities which are self-contained more amenable than
those which are hooked in with a raft of other activities, and also
dependent on other practitioners (eg. usability, coders, etc)? Well, yes,
but to what extent?

I also think that it would be beneficial to our field to further develop and
refine the individual practices and techniques we employ, even without
considering the issue of outsourcing. Things like JJG's visvocab and similar
documentation have been very useful in practice, and also valuable in
furthering our credibility.

e.