On Mon, Nov 22, 2010 at 12:03 AM, neville holmes > My tack was that
the tacit style is the purest imaginable
> (to me at any rate) mode of functional expression and my
> experience was that the students agreed and coped very
> well, though they always went back to what they saw as
> the more expressive (I disagreed) coding styles as in
> Basic (ouch!) and C (shrug!) for example.
>
> The core issue seems to be whether you are tackling
> a calculation or facing up to a massive system.  I held
> the opinion that tacit J was basically calculation and
> that connecting to files and databases was peripheral.
>
> However, there are problems in calculating with tacit J.
> I recently put an essay in my Computer column describing
> a device I called the formulator (see
> eprints.utas.edu.au/9474) that was based om my experience
> teaching tacit J (though that essay doesn't explicitly say
> so).

I agree with you that tacit expressions can be very nice to
work with.

I also agree with your statement that there are problems
with it.  Though, personally, I think that these problems
were explicitly designed into it and I think that that means
simply that it is not suitable for everything.

Tacit J is a great tool for concisely expressing certain
kinds of calculations.  However other kinds of calculations
(including, especially, calculations with too many data
sources) are better expressed as explicit calculations
[or, in some cases, are better avoided entirely].

A related issue, though, is expressing the purpose and
domain of the vocabulary.  You can get a lot of mileage
here with well chosen names, and with good examples.
But (especially in large programs), some explicit notes
about a block of code's purpose or usage can be
essential for some people.

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to