SUMMARY: Sizing illustrations

[Mark Levinson <mark -at- SD -dot- CO -dot- IL>]

Last Wednesday I put forth the opinion that when a book reproduces
the computer screen several times in whole or part, the ratio of
inches on screen to inches on paper should stay constant.  And I
asked for reactions.

Melissa Hunter-Kilmer put into words the attitude that, to me,
seemed to emanate from inconsistently sized reproductions:
"Let's see what I can get away with, and I'll do pretty much
what I want" is what work like that says to Melissa, who sneers
right back at it.

But my first reply came from Maria Townsley, who wrote that after enforcing
consistency for two years she started expanding most captures to the
margins and found the users happier with the improved readability.

On the other hand, Auli Ingman says:  "If the contents of the screen are
legible in the smallest version, then there is no reason to have larger
versions at all.  To enlarge them just because you have the space available
must be very unsettling to your reader."  Beverly Parks had the same
opinion.

I'm sure a great deal depends on what your screen is showing.  But I guess
Auli, Beverly, and I would ideally establish a legible standard and consider
that the benefit of sticking to it outweighed the benefit of giving maximum
magnification to each illustration.

David Dubin has a simple method of setting the standard:  he sets the
ratio simply by mapping the width of the screen to the width of the page.

Jan Bates, on the other hand, is also against random magnification but
thinks that exceptions may be necessary when the graphic doesn't fit within
the margins at your preferred magnification.  David's system would work
fine for Jan when the whole screen fits legibly into the page margins;
but maybe not everyone's that lucky.

While Jan would make exceptions for big illustrations, David himself
admits exceptions for small ones, which he would enlarge when he needs
"to point out something unusual or specific that may otherwise be missed."

Marilynne Smith says she's "always tried to keep similar elements similar in
size.  That goes for tables as well as graphics.  I would not change the size
differently unless I had a very good reason.  Also, be careful that reduced
graphics and tables are still readable when they are reduced."  Marilynne's
concept of "similar elements" seems to imply that different sorts of data
might be reproduced at different standard magnifications.

Ken Duffy too allows for variations on a standard magnification:  "I
think it's OK to have small, partial menus be enlarged a little from
standard.  And, especially, it's also OK to have entire windows
reduced.  If you have to reduce it to make it fit, so be it.  It's a
practical limitation, which (I think) can lead to an exception to the
rule/pattern.  The full-window case need not set a common-denominator
scaling factor that all other images (dialog boxes, etc.) must use.
That setting would make text less readable in those other cases,
for the sake of (arbitrary?) consistency."

Dee Philipp is for consistency "ALWAYS! No matter what: if it's a button of a
GUI, 50 percent; if it's an entire screen, 50 percent."  Dee says
"That's the way Microsoft, Novell, WordPerfect and many others do it."
Me, I'm not so sure about Microsoft.  Maybe it depends on which year's
manuals you're looking at.

Nora Merhar(?), Lori Moore, John Brinegar, and Carma C. Allen also wrote in
favor of using the same magnification consistently.

Chris Tarski believes in consistency too, but to Chris consistency
means stretching or shrinking each illustration to fit a standard
number of inches of page width.  (Most of the time, Chris is illustrating
the whole screen anyway.)

Dan Strychalski doesn't mind a little variation but "would never, ever blow
everything up to full page width regardless of its original size."

Sue Gallagher added a side issue, quoting research in which full screen
snapshots worked worse than button/icon graphics and worse than no
graphics at all.  She says, "Cut down on the number of screen shots and
don't be afraid to zoom in on portions of screens, buttons, and icons!"
Richard Mateosian says "I don't know why this surprises anyone."
Richard believes that ideally after "a few carefully annotated screen
shots ... screens should require no further description. The remainder of the
manual should focus on the tasks the user needs to accomplish and the support
the software provides for those tasks."

What does it all add up to?  By my count, we had two in favor of making
every illustration as big as individually possible, we had eight in favor
of sticking to a single factor of magnification, and we had five who
consider that magnification may be fooled around with a little but not much.

Back on Wednesday, I thought that maybe this was the issue on which I
would walk from my job.  But for the moment all parties at my office seem
satisfied with a two-tier standard-- one magnification for a particular
type of data that tends to take the full screen width, and another
magnification for all other types of data-- while admitting one off-the-wall
exception among all the particular illustrations at hand.  My boss has become
less insistent regarding how the book should look anyway, because we now know
it will include a diskette for people who want to examine the data closely.

But heck, jobs come and jobs go.  What's important to keep is your faith
in your own clear judgement, and you folks, mosly from thousands of miles
away, helped me to keep that faith at a tough moment.  Love you madly.
__________________________________________________________________________
||- Mark L. Levinson, mark -at- sd -dot- co -dot- il -- Box 5780, 46157 Herzlia, Israel -||
||                      - Death to fanatics! -                          ||