Suggestion: Arrays in Rev3 beta

[David Bovill]

2008/8/28 Bernard Devlin <bdrunrev at gmail.com>

>
> Personally the returned keys being unordered has never been an issue for
> me.
>  Can you illustrate some situations where this is a problem?


In parsing data in order to layout in Rev I often use arrays. However the
order of the original data is often important. Take a recent example - I am
working on updating some scripts that parse wiki formatted text. I want to
extract all the headers (like html headers they can be of different levels)
and store the paragraph of html that goes with each section. Then I want to
lay out the data in a Rev interface. I create an outline from the wiki
headers and when the user clicks on the header the corresponding text is
displayed.

There are a number of ways of doing this - and if anyone has any suggestions
about the best way please chip in - you can use list and Rev based indexes,
you can use XML, or you can use arrays. XML is natural for these
hierarchical tree style data structures, but franksly a bit of a pain and
overkill. So I tend to use arrays, and looking very much forward to the
nested arrays in beta 3 for this.

However in order to "reconstruct" the original order of the document I
cannot simply loop through the keys of the array as the headers will be all
out of order. So I have to go to the effort of storing both the array and an
ordered index of the array. Then if I need to store this data and have two
things and two places to store it.

In general about 1/3 of the time I use arrays I also have to store an
ordered index.