Re: [oiic-formation-discuss] Informative clauses

[robert_weir@us.ibm.com]

Radoslav Dejanoviæ <radoslav.dejanovic@opsus.hr>
wrote on 06/23/2008 08:05:06 AM:

> Dave Pawson wrote:
> 
> > I think that is presumptive. It would misinterpret the standard?
> > If the author had wanted a shall or must surely he|she would
> > have put one in there?
> 
> When I write (or read) documentation, I always presume that something
> should be done as it is written, unless there's a choice, but in that
> case the choice should be (and usually is) clearly described.
> 
> For that reason I would say that it might be Ok to consider every
place
> where there's no clear choice given - as mandatory, non-changeable
item.
> After all, standards should be read and understood by engineers and
> developers who would like to have a clear picture of what they're
> supposed to do...
> 

Let's take a fictional example that we can all understand.

I'm writing a standard for text format.  I want
to define a way to indicate bold test.  I define it like this:

"The <bold> element is used to indicate
bold test"

I think this is clear enough to implementors what
is intended.  But from a formal perspective, I have not said "shall"
or should or "must" or anything. 

So a conformance test would not be able report an
error for an application that failed to render bold text as bold.  Since
the ODF standard defines conformance, any test of conformance must be traceable
to what the ODF standard says.

However, in an interoperability test I think we can
go beyond the limitations of conformance and test the runtime behavior
of bold, asking questions such as:

1) Is bold text visible?
2) Is bold text distinguishable from non bold text?
3) Does bold text render with a heavier font weight
than non bold text?

These are three reasonable tests.  Although the
standard does not come out and say "Bold text shall not be rendered
identically to non bold text", it is a reasonable and useful thing
to test for, from an interoperability standpoint.

-Rob