[Insight-developers] Notation for _Insight Into Images_

[Mark Foskey]

I generally liked the proposal for theory book notation, but I would 
like to raise one issue.  I do not think we should use Roman type as a 
way to indicate discrete variables.

Using italics for stand-alone English letters (not only variables) is 
not just the LaTeX standard, it is the standard for typesetting 
mathematical formulas across fields.  There is a good reason for this 
in that it distinguishes mathematical symbols from running text, and 
also from standard functions like log and sin.  Of course you can 
generally tell from context what was meant, but it is distracting when 
the convention is not followed.

Also, since it is such a well-established convention, not following it 
tends to make materials look less professional.  Take a look at a CS 
book with mathematical content whose appearance you like and see what 
it does.

Terry pointed out at the beginning of his proposal that this book is 
for programmers, not mathematicians, and that our notation should be 
motivated by the notation that would be used in implementations.  In 
response it is worth bearing in mind that programmers also do not use 
changes in font to distinguish ints from floats, since they do not have 
that option.

Finally, there are well-established ways to indicate discrete values, 
and they are actually in the proposal as well.  I suggest that we can 
rely on them.

Again, look around at good CS books, in particular good image analysis 
and pattern recognition books, and see what they do.

Just wanted to get this in before we froze this....

-- 
Mark Foskey    (919) 843-5436  Computer-Aided Diagnosis and Display Lab
mark_foskey@unc.edu            Department of Radiology, CB 7515, UNC
http://www.cs.unc.edu/~foskey  Chapel Hill, NC  27599-7515