On 12-12-22 12:14 PM, ron.vandenbranden wrote:
> On 22/12/2012 11:46, Saeed Sarrafzadeh wrote:
>> On Sat, 22 Dec 2012 09:39:55 +0000, Lou Burnard <[log in to unmask]> wrote:
>>> ... -- it really is meant only for cases where the target and
>>> the source are interchangeable, as for example where one translates the
>>> other. I also think it is meant to be bidirectional -- that is, if
>>> corresp(a.b), then corresp(b,a) ...
>> Thanks for your explanation. What I interpreted from the guidelines was a more general case for corresp attribute which can be any type of correspondence between two elements
> I must confess that the exact semantics of @corresp have always remained
> rather vague to me. Consequently, I might be abusing it in all kinds of
> ways to express /some/ kind of correspondence between two identifiable
> elements. If, on the contrary, this is explicitly not the indended use,
> I think the formal description 'points to elements that correspond to
> the current element in some way'
> is to blame...
I've always assumed the vagueness was intentional, and the flexibility
of the attribute is useful. It's up to the end user to decide what
"corresponds" means. It's a good idea to document your usage in the
header, of course (although I must admit I never have).