On Thu, March 15, 2007 3:40 pm, Andrew Lentvorski wrote:
> Lan Barnes wrote:
>
>> Agreed, but data driven programs can be way-cool when done by someone who
>> knows how.
>
> I consider all programs to be data dirven, but I'll let you clarify.

I won't dispute your statement, although it's a little like saying
"everything found on planet earth is a 'natural ingredient'." ;-)

I can best illustrate what I mean by data driven programs by a simple
example. Say I have a task to develop a data base program which is front
end screens and back end server.

1. I can draw or hard code a gazillion screens.

2. I can take the time to write a generalized front end that finds how to
draw the screens by hitting the data base for such things as size, color,
background text, input fields, etc etc.

Method 1 is quick and dirty and guarantees a maintenace ... challenge ...
as things change, which they always do. PHBs really like method 1.

Method 2 has a big front end programming load, but makes screen
development on this and subsequent programs like this lots faster and
makes maintenance a relative breeze. Wherever there is a change in a
screen, you just add, delete, or edit the screen data (now considered
metadata) in the back end.

If you're really serious, method two is better. PHBs usually don't get
method 2 and consider it yet another case of the bit twiddlers wanting to
play instead of doing real work.

Method 2 is what I would characterize as data driven programming.

NB: There is a method 3 which I will be presenting in April since I can't
seem to weasel out of it any longer. It is much beter than method 1 but
not as costly (in the front end) or as cool as method 2. But I can't tell
you more right now because I'm trying to build suspense :-)

<some BCCs>

-- 
Lan Barnes

SCM Analyst              Linux Guy
Tcl/Tk Enthusiast        Biodiesel Brewer






-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to