More complex magic procedures

Subject: More complex magic procedures
From: geoff-h -at- MTL -dot- FERIC -dot- CA
Date: Tue, 20 Jan 1998 22:17:57 -0600

Michael Wing keeps right on upping the ante on the problem
of providing context: <<Flowcharts, cartoons, decision
trees, and so forth aid in learning the sequence of what to
choose and in which order to choose them but still
does not satisfy the need for the user to understand the
ramifications of combining choices.>>

This is getting progressively more interesting... and no,
I'm not being facetious. Mike, you're posing us an
increasingly complex problem in information design, and I'd
be surprised if we can't crack it for you with a few more
details. Let's keep going on this!

<<If two parameters each have ten settings, there are a
hundred possible outputs. Now, illustrate the outputs for
10 parameters with a selection of settings ranging from 2
to 25.>>

You can undoubtedly simplify the problem considerably if
you define things in terms of user tasks rather than simply
permuting all the possible parameters. I doubt, for
example, that users will pick a line width and then fart
around with scales and azimuths until the find one that
works with that line width. Typically, users will want to
fit a map on a given paper size, with legible lines, at a
predefined scale (e.g., the 1:10 000 airphoto standard in
Ontario).

For example, pick the main parameter that characterizes a
specific task, and try a restricted example to see whether
it leads us to a general principle. [Caveat: I'm not
proposing a final solution here, and I haven't double
checked my logic/math/etc. It's just an example.] Let's say
the user's task is to produce a map that fits the Ontario
scale (a specific example of the general task of producing
a map with a specific scale). You might get away with
something like this:

***Map scale***
Once you've defined the area that will be included in
your map, setting the map's scale determines the size of
the final printout, the size of the features visible on
that printout, and the amount of RAM required to process
the image.

Printout size: For a given area, increasing the scale
increases the size of the printout; for example, doubling
the scale doubles the printout size. To avoid having to
tile your output, we recommend using a pin-fed plotter.

Size of map features: If you reduce the image size, you'll
need to increase the thickness of the lines to compensate.
Conversely, if you increase the image size, you'll need to
decrease the thickness of the lines so they don't become
unacceptably thick.

Memory allocation: The status line displays the size of the
printout based on the scale and area you've chosen. The
larger the printout, the more memory you'll need to handle
the map. We recommend 1 Megabyte of RAM per square inch.

And so on. The text and the example are admittedly trivial,
but the essential point isn't: Start with the task, and
identify the implications of each parameter/setting for that
task. If you understand what your users will do with the
software, you can prioritize the main settings or recommend
a sequence of settings. In the previous example, the
sequence would be define your area, pick a scale, and adjust
the output parameters to match. Continue from there!

<<Often, these types of applications use successive
approximations in which the user gets within range of the
expected results on the first iteration and then tweaks
the input parameters to reach the goal.>>

Well, the above approach provides much of that theory
indirectly, but it's not possible for anyone other than a
chess master to keep all those possible permutations in
their head. Your goal should instead be to give enough
theory so that users can quickly reach the stage where all
they're doing is tweaking a single parameter. The
recommended sequence of activities will vary for each task,
as will the parameters you need to define closely. This sort
of problem is beginning to intrude on UI design more than
documentation, and seems tailor made for the sort of
approach that some photo editing software uses: provide a
palette of thumbnail images and let the user pick the one
nearest to where he wants to go.

--Geoff Hart @8^{)} geoff-h -at- mtl -dot- feric -dot- ca
Disclaimer: Speaking for myself, not FERIC.




Previous by Author: Buying a new monitor
Next by Author: $75K per year in tech. writing?
Previous by Thread: Re: Buying a new monitor
Next by Thread: More complex magic procedures


What this post helpful? Share it with friends and colleagues:


Sponsored Ads