Re: [tlug] Are ordered hashes useful?

[Raymond Wan <rwan.kyoto@example.com>]

On 02/15/2011 07:22 PM, Simon Cozens wrote:
> On 14/02/2011 03:20, Stephen J. Turnbull wrote:
> > But most of the examples you give there (eg, histogram) are really
> > special cases of dictionary.
>
> No, I don't think so, and this is the point. Dictionaries define a
> transformation between list A and list B. If you want to be pedantic, then
> formally, yes, a histogram is a case of a dictionary. But in terms of how it's
> being used, there isn't a second data structure B - it's just telling you
> properties about list A. What we often think of as mapping between two lists
> is more often asking questions about the structure of one list.


Hmmmmm, I'm actually with Stephen on this one...

Hashes, or hashing, takes some set of inputs like a set of 
strings and applies a hash function h() to it so that each 
string is mapped to a numerical position.  What you then put 
at that position doesn't matter...  Whether it's French 
words or a number indicating how many times the word appears 
in a document -- it is still a "value".

While they are both interesting uses of hashing, I'm not 
quite convinced that the two situations are that different...

Even if you store a binary tree at each hash table position 
-- which isn't a value per se -- it would be implemented as 
the memory address of the root of the tree.  So, I guess 
this too is some abstract definition of a value?

Ray