translated to the local format at runtime [Re: Rev 2.02/New pricing]

[Alex Rice]

On Friday, July 18, 2003, at 08:33  AM, Geoff Canyon wrote:
>
> Yes. Well. Erm... I never said the bytes were in the same _order_, 
> just that they were the same. ;-)
>
> In reality, for most apps this has no consequence. Functionality 
> should be identical. The translation happens transparently, and except 
> in the case of _large_ quantities of text is extremely fast. Anyone 
> have a benchmark?

My project builds a single-file executable approx 5-6 MB in size, which 
includes 17 substacks and about 120 cards. The automatic 
isotomac()/mactoiso() text conversion by RR is very fast.

AFAIK what is being the converted is not scripts or custom properties, 
only the text on buttons and fields on the cards of the app.

I haven't got any actual benchmarks, but my feeling is that when you 
factor in varying CPU speeds, and disk caching, it becomes a non-issue. 
By disk caching I mean the initial launches take ~ 2-3 seconds on my 
mapp. Then after the first launch of the app, subsequent launches take 
< 1 second. Whatever the time it takes to do the text translations, 
it's not subjectively noticeable. It's less time than the initial 
reading of the app from disk.

But technically speaking yes this is one difference between builds in 
Mac and builds on Windows. ;-)

BTW do Unix RR engines use the same text encoding as the Windows RR 
engine does?

Alex Rice, Software Developer
Architectural Research Consultants, Inc.
http://ARCplanning.com