Re: New proposal for math.e
Juergen Luethje wrote:
>
> CChris wrote:
>
> > Juergen Luethje wrote:
>
> <snip>
>
> >> I do not even know the meaning of a gamma() and an erf() function, so I
> >> can't provide code for them. But you probably can do so. 
> >>
> >> Regards,
> >> Juergen
> >
> > The Cephes C/C++ numericall library at www.moshier.net has more than what we
> > need, and doesn't seem to have any licensing issue, apart from reproducing
> > the
> > copyright notice.
> >
> > CChris
>
> That's probably a good source for people who understand C. Another good
> source might be "Numerical Recipes in C", see
> <<a
> href="http://www.nrbook.com/a/bookcpdf.php">http://www.nrbook.com/a/bookcpdf.php</a>>,
> and I'm sure there is more free
> stuff on the internet, especially in C (or C++).
> There are also big math textbooks, so literally dozens of formulas are
> waiting to be implemented in Euphoria.
>
> But we should soon come to an end (for now!). It is not necessary to
> include all formulas that we consider more or less important in the
> _first_ edition of "math.e". We should be somewhat modest regarding
> our first "community product". Nobody expects that the first product
> of a craftsman is already his masterpiece.
>
I'd be a little more prudent here.
It is true that we don't need more than two, three, four functions, because
hardly anyone in the vocal community - I'll say more on this later - is
interested in maths beyond a very elementary level. However, the small steps
approach often results in stunted growth, because said growth will impose changes
later, while changes are widely rejected bu the vocal community. Euphoria has
suffered a lot from that already, so don't be too cautious or shy - the cost
later is high.
> I think for now we should not add more than say two or three additional
> functions. Then we should try to get agreement about as much points as
> possible, and decide what to do with the points that we couldn't agree
> upon. Then the documentation must be written. This also will be a
> considerable amount of work. Maybe there also will be some discussion
> about the documentation.
>
The documentation for the very simple stuff being addded will really require
little work. However, I expect an immense amount of discussion around it.
> All this should not take too long, otherwise too many people will lose
> interest.
It is not possible to access posts 5 day old from the web interface at the
current rate, which may not collapse sharply anytime soon. The "not too long"
really needs to be very short, otherwise nothing consistent will be ever built,
even before people lose interest, as posts won't be available. All right, they
will through the search facility, but it is less convenient, not everyone will
use it, and the reference # of the post aren't even apparent on the search result
pages.
> But I think it's important to actually release a
> "math.e created by the OpenEuphoria community". This will be give a good
> feeling of success, so we should concentrate on this goal. Later, we can
> anyway add more functions to the library like we want.
>
Now we may have a problem. The number of posters on this list is (hopefully!)
largely less than the number of Euphoria users worldwide. I have no idea about
whether the opinions being voiced here, as well as their relative frequency, are
representative of the user base at large. Perhaps Rob can tell us something about
it - he is probably the only one to receive news from users who are not amongst
the group of 15 persons at best trading code, opinions and preferences at any
given time on this list, which I called the "vocal community" earlier.
CChris
> Regards,
> Juergen
|
Not Categorized, Please Help
|
|
