the keys of myArray

Richard Gaskin ambassador at
Fri Sep 8 11:31:02 EDT 2006

jbv wrote:

>> There may be other ways to handle this, although if using the built-in
>> sort command will do what you need it's reasonably efficient.  It's only
>> one line of code -- was it causing a problem in your script?
> it's not causing any problem in my script.
> again, I was just wondering about any logic behind the keys order...
> actually, there's only 1 situation I can imagine in which it could be
> useful to know in advance in which order keys would be output :
> in heavy math treatments with loops in which arrays values & keys
> are needed for further processing, many lines of code (sort commands)
> would be saved and scripts sped up significantly...

If you're up for it feel free to post the code and the folks here can 
see what can be done to optimize it.  I've been very impressed with the 
results of that sort of group effort here and on the MC list, and I've 
learned a lot from those exercises.

A recent example:

The script colorization algorithm in the MC IDE was quite slow, taking 
more than 30 seconds to colorize long scripts like libURL on my 1GHz 
Mac.  Those of us who volunteer to work on the MC IDE were looking for 
ways to speed that up, but we considered caching the htmlText a last 
resort since that workaround adds its own difficulties (increases object 
size, adds stuff to the user's custom properties, and introduces 
additional complexity into the code base, requiring overhead to maintain 
and keep it in synch).

More importantly, this type of algorithm is likely useful in other 
people's apps, so coming up with a way to do this efficiently without 
caching would open doors for  Rev's customers, turning what was a 
annoyance and a support expense into an inspiring liberation.

I don't know much, but I know there are a lot of people in this 
community much smarter than I am, so I posted a note to the MC IDE list 
to ask for help in optimizing this.

We had four people come up with a variety of alternatives, all of which 
were several orders of magnitude faster and none of which required caching.

The end result is that there's now a script colorization algorithm 
available which can do libURL in about a second -- and that's on a slow 

This exercise not only gave us a very valuable algorithm, but more 
importantly reinforced the value of leveraging the collective wisdom of 
the community when facing a difficult challenge.

After that experience I'm sold:  as soon as I come across a task that I 
think is impossible, I invite others to look it over and I always learn 
a lot in the process.  And usually the code gets both smaller and faster 
as well. :)

> BTW here's a suggestion for a slight improvement of the "keys" function :
>     get the numKeys of myArray  -->  outputs the keys in numeric order
>     get the alphaKeys of myArray  -->  outputs the keys in alphabetic order

Those might be handy indeed -- have you submitted those requests to BZ?

  Richard Gaskin
  Managing Editor, revJournal
  Rev tips, tutorials and more:

More information about the Use-livecode mailing list