Re: [Groff] Manuals in pdf format

[Miklos Somogyi]

On 28/02/2006, at 8:55 AM, Werner LEMBERG wrote:

> > In general it is good to follow rules, most of the time they are for
> > our benefit.
> > Sometimes circumstances change and rules need to be modified.
> > This "i, ii, ..., 1, 2" system was all right in the age of paper
> > 'cause any-one page had only one page number.
> > In the age of pdf one has two numbers to describe the same thing.
> > Clickable links are good, very good, two numbers for the same thing
> > is perhaps not.
>
> Hmmm.  Since PDF version 1.3, there is direct support for this: It is
> possible to define a PageLabels dictionary.  To cite section 8.3.1
> from the PDF reference, version 1.6:
>
>   Each page in a PDF document is identified by an integer `page index'
>   that expresses the page's relative position within the document.  In
>   addition, a document may optionally define `page labels (PDF 1.3)'
>   to identify each page visually on the screen or in print.  Page
>   labels and page indices need not coincide: the indices are fixed,
>   running consecutively through the document starting from 0 for the
>   first page, but the labels can be specified in any way that is
>   appropriate for the particular document.  For example, if the
>   document begins with 12 pages of front matter numbered in roman
>   numerals and the remainder of the document is numbered in arabic,
>   the first page would have a page index of 0 and a page label of `i',
<...>

This thing remind me of OpenGL's handling of normal vectors.
Their bible says:
     "In other words, normal vectors are transformed by the inverse 
transpose of the
     transformation that transforms points. Whew!"

Also' I've read a list of traps OpenGL novices fall into.
The most prominent one is that they forget to re-normalize normals when 
they apply
transformations to the body.

Now the beauty of the thing is, that when you rotate a body, you rotate 
the normal vectors as well.
When you translate or scale, you translate and scale the vertices, but 
not the normals.
Now these "Whew!" guys transform the normals, always.
By doing so they need to make some extra effort. Then they have to 
re-normalize the normals
to offset the problem they created with extra effort.
And they need to warn programmers not to fall into the trap and do the 
extra work that would not have
been required if they did not do something extra.
I am not sure that I am 100% right in this paragraph, when I met this 
thing first time, I decided
to stay with IrisGL as long as poss.

My problem with this numbering system is: why go to such a long way to 
placate something that
does not seem to give me, the reader, anything useful?
Usually I am reluctant to throw-out things, but I would not be sorry to 
say good bye to i, ii, iii, iv, 1, 2, ...
and welcome 1, 2, 3, 4, 5, 6, ...

Miklos