From: "Pedro CĂ´rte-Real" Date: 2006-08-22T19:22:39+09:00 Subject: Re: Why Does Hash Apparently Reorder Its Internal Representation And Other Associated Ponderings On 8/22/06, Austin Ziegler wrote: > It would be ordered and have arbitrary key values, but it would not, > in fact, be a Hash. That's the point that some folks have tried to > make about this. It is a Hash, since it provides it's interface and general performance characteristics. It just supplies the new guarantee of insertion ordering. I've wanted one a few times, instead of nested arrays, mostly because writing something like ['a' => 'b', 'c' => 'd'] would look better than [['a','b'],['c','d']]. I'd still iterate both with somename.each{|key, value| ...}. I've written code where I needed both the order and the random access. An example was a list of fields, each with an associated hash of options. I ended up declaring an array for the fields and then a hash of hashes to store the options. An ordered hash would have solved this neatly. A lot of people seem to want this so why not create a standard OHash? A nice syntax like the [1=>2] one would be great as well. Pedro.