[Sigia-l] Elements of Documentation

[Anders Ramsay]

On 6/30/02 7:30 PM, Listera at listera at rcn.com wrote:
 
> If you could create functional prototypes in roughly the same timeframe
> would you bother with any of the other stuff?

I don't think a prototype by itself could replace documentation.  For any
system more complex than brochureware, e.g. anything containing elements
with more than one state, it could inherently not represent all potential
states of the system. E.g. it can not both represent the state of a list at
a maximum allowable length, and an empty list.  Even if someone were able to
finnagle it to represent all states and rules applying to the system,
actually revealing them would require that the user traverse every nook and
cranny of the prototype, which creates a very high risk of missing a
functional or content requirement (and is very labor-intensive).  And I
speak from experience in this case.  I was handed a project in which the IA
officially was complete, and the completed IA was in the form of a
functional prototype of the site.  Well, I can tell you that the IA was
pretty darn far from complete, and somewhat of a mess to boot.  For example,
the prototype did not/could not reveal how templated pages were to appear
depending on the multiple variations of content that might (or might not) be
present for optional  content slots.  It was in no way apparent what the
display rules were for these variations.  Or because the mock copy in text
pages in the prototype often was at the maximum expected length (which is
good on one hand), it made sense to have redundant navigation (such as
'next/previous' links) at the top and bottom of the page, but it would look
ridiculous if the copy were 1 or two sentences long, which was completely
possible.  The rules for when redundant links were non-existent. But
probably the most significant problem with having only a prototype is that
it increases the risk of what I think of as a cardinal sin  in the world of
system design--inconsistency, in part because it is completely unnecesary.
Default values for drop-down selects sometimes said 'select...' and
sometimes said 'choose...' or sometimes were not specified at all.  Page
titles were not generated consistently.  Sub-level navigational links used
inconsistent terminology.  And on and on.  Consistency requires some form of
global referencing, so that elements appearing throughout the system are
only specified in one place, which goes counter to a prototype.

So, I guess my point is that while I think prototypes are great for proof of
concept, exploring custom designs, testing, and presentation, I don't think
they are quite up to the task of replacing documentation.

-Anders