Re: [stack] function "adjoinment"?

Perhaps because of 3rd [going 4th] world connectivety, I couldn't get
the ltu - yet.
But since I've 'got you on the line' I'll slip in MY topic "cat-like
for productivety".

There's no problem with people playing cross-word puzzles, but
when I queried cat-like's contribution to productivity, W. Tanksley changed the
subject and pointed out that 'greatest productivity increases are from
team-building,
hence the increased use of java'. I don't dispute this, nor the claim that any
productivity increase is bad, because it creates unemployment.

However, altho' the joy-author doesn't seem to directly say so, as
Prof. Herzberg
does in his reference to 'conquering complexity', my reference to PRODUCTIVITY
is the same as theirs: the individual algorithm-designer/coder; and not some
higher 'what is the purpose of life' consideration.

Admittedly the topic of productivity increase is related to psychology and fuzzy
stuff which we technologists don't formally know about.
While re-reading my files related to 'this' I find repeated implications towards
that "structuring helps human cognition and hence productivity"; but
like Victorian
sex it's never spelled out. Eg. is point-free good because it avoids the need to
remember names ? Or is the replacement of names by "it" really better,
when you might
NEED to know/remember whether the data is in the soak, wash, spin or
dry function?

Consider how many billions of people CAN ride-a-bicycle, but few KNOW how to.
Analogously, IMO computing productivity [in my above defined context] must start
from human-cognition/psychology. Knowing the goal is more important
that discussing
cute steps which allow you to move - perhaps towards some undefined goal.

Here's a paste  which is relevant to individual
productivity:--------------------
I had great productivity increase results with mc > cedit > User-menu/F11
for scheme-code.
My previous notes, stated that I didn't like the 'structured editor' facility,
because I had to lose sight of the text while the process inserted
the new text. Yes the annoyance of 'loss of sight of the subject' for half a
second, is related to the annoyance against top-posting ?

Once I'd accepted this small annoyance, and I'd patched cedits 'lisp' facilty
towards scheme, I found that the syntax-colouring and "just enter a
full structure
at a time instead of single-characters" was a MASSIVE productivity booster.
It's especially good for scheme, which has only a few syntax elements.

The beauty of such a system, is that when you work with a new language, or
a "haven't used for years" one, you re/familiarise yourself with the syntax
while building the cedit-templates. And 'capture' the knowledge of the syntax
-- that's how 'expert systems' work.  Then you don't have to waste your creative
mental resources on syntax trivialities. After all, an IfThenElse construct has
only 4 elements: its name, the predicate, the consequent, the alternate.

Why enter 6 keys for 'banana' when it's just a SINGLE-concept !?
I want to extend cedit's user-menu to be able to enter IDs once only, and then
just pick-them-off-a-menu for later insertion.  Scoping won't be possible.

So: avoid complexity and adding too many features to mc/cedit, but I
found cedit's
syntax hi-lighting and template-inserting editor a big productivity booster.
The hi-lighting/colouring and predefined layout/pretty-printing makes eg. the 4
components of the IfThenElse construct evident/identifiable from amongst the
mess of brackets.

I can't understand why they haven't eliminated char-by-char code writing, and
replaced it by menu-driven -- a bit like spreadsheets. The only things
you need to
originate, besides the 'sequence', are ID names.
When you want to instruct an elevator to go to floor 9, you select ONE button,
instead of writing a syntactically valid message!?

== Chris Glur.



On 2/3/10, John Nowak <[email protected]> wrote:
>
> On Jan 31, 2010, at 6:22 PM, John Nowak wrote:
>
>> We're familiar with the notion of function composition:
>>
>>    (a -> b) -> (b -> c) -> (a -> c)
>>
>> However, is anyone familiar with a notion of function "adjoinment"?
>
> For what it's worth, I've asked a similar question on LtU to cast a
> wider net:
>
> http://lambda-the-ultimate.org/node/3803
>
> - jn
>
>