Item Description: Recipe-controlled data manipulation.
Review Synopsis: Very powerful and flexible.
Description
The Data::Flow module
allows you to specify a hash of "recipes"
that fetch or manipulate data.
Each recipe has a name (the hash key),
which you can use to set the input or fetch the output
value of the recipe
using the Data::Flow set and get methods.
A recipe can depend on
other prerequisite recipes,
which means that the input of one recipe
can be taken from the output of other recipes.
You can use this to create a
hierarchy (or other structure) of dependent recipes.
The cool thing is that
you then call for a specific type of output,
and the data flows from its input recipes
through all of the dependent recipes to your output,
without you having to code the calling structure directly.
This allows for a lot of flexibility,
because you can manipulate your calling structure
by changing your hash recipes and dependencies,
not by having to find all of the places
that subroutine X is called in your code.
Why should you use it?
The Data::Flow module would be especially appropriate
if your data can be processed in a lot of complicated
and inter-related ways,
and you want a lot of flexibility
in how you'll want it processed,
now or in the future.
Why should you NOT use it?
For straightforward processing of your data,
the recipe structure of Data::Flow can be overkill.
It's not as intuitively obvious
as just hacking the code.
Related Modules
The C::Scan module
uses the Data::Flow module.
(Both were written by Ilya.)
Example
The Data::Flow documentation
provides an okay example,
but looking at the C::Scan module will
teach you more about using Data::Flow powerfully.
|
