Ralf Hemmecke wrote:
> 
> >> (A) we allow (*) ... and then mark such places such that if ever
> >>     NNI and Integer don't share the same representation, these places
> >>     can be found easily, or
> >> (B) forbid such (potentially) dangerous "pretend"s.
> >>
> >> I'm somewhat in favour of (A) (restricted to NNI <--> PI <--> INT).
> >>
> >> Ralf
> > 
> > I agree with (A). The InnerSparseUnivariatePowerSeries is the
> > representation Domain for both Taylor and Laurent series,
> > so it accepts Integer (instead of NNI) as exponent.
> > 
> > So it is safe to use 'prentend' to replace 'nniToI'.
> > You should make a patch for it.
> 
> The patch is attached.
> 
> Question is what is a good form to recognize such places?
> Mentioning SubDomain is certainly a good idea, because it depends on it.
> But grepping for "pretend" alone is not the best to find the places
> where a "pretend" refers to an "Inner" domain. In fact, I will add a
> section "conventions" in order to state that "Inner" domains are not
> exported to the end-user, but rather only to library writers (of course
> the distinction between these two roles is not very strict).
> 
> Waldek, your opinion?

This is good use of 'pretend'.  Concerning finding such places
when needed: typically search will be limited to few domains
so looking at all 'pretend'-s in involved domains is not too
bad.

If some 'pretend' is likely to get much use we can wrap it
like 'qcoerce' and 'qconvert': both are as efficient as
directly using 'pretend' and their use is good indictation
of what happens.

I am sceptical of various manualy maintainded cross references
and list of usages: IME probablity that they are wrong is
large enough that it is better to recreate information when
needed.  And much less work overall.

-- 
                              Waldek Hebisch

-- 
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to