MathGroup Archive 2008

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: A Package Function Tutorial


Just a small comment that probably only adresses beginners like me:

While a function with all the bells and whistles can have a lot of good
things, I often
spend too much time trying to figure out how a function works when i
download a package.
If it had less of all that nice stuff, I could understand how it works and
avoid feeding it with the wrong input (instead of not understanding it and
having the function tell me that it's a wrong input)
Of course, I am not suggesting people shouldn't write professional looking
functions; I'm only saying that sometimes there's a trade-off, at least for
beginners like me that are trying to understand how the functions work.

cd

2008/9/23 <blackhole at thebushman.net>

> I have written a brief introduction to writing package functions that
> implement most of the bells and whistles one would associate with a
> professionally written function.  Some of the topics covered are
> implementation of Automatic in functions, overloading of built-in
> symbols, and others.
>
> Yes, most of the material can be found scattered in the shipping
> documentation, but my sense is that there is no good place to find
> essentially a checklist of the features a good function should have,
> like argument checking, attributes, syntax coloring, etc.  You will
> find that I am no expert, but perhaps someone at Wolfram (or
> elsewhere) will be offended enough to contribute some good
> suggestions.
>
> You can find it at my website, http://www.thebushman.net.
>
> Here you will also find a package I published last month, a NURBS
> package providing a bridge from parametric curves and surfaces to CAD-
> friendly IGES format.
>
>



  • Prev by Date: Re: Diagonalize sparse matrix and documentation in version 6.
  • Next by Date: Function changes in a program
  • Previous by thread: Re: A Package Function Tutorial
  • Next by thread: Re: Re: A Package Function Tutorial