Speed differences between MC and Rev

[J. Landman Gay]

Rob wrote:
> Hi all,
> 
> I have suddenly felt the urge to un-lurk myself  :).
> 
> As the licensing of Revolution is now in the engine allowing full use of MC
> with the Studio version of Rev I am ready to purchase and I am 
> interested in this speed difference issue.
> 
> So I decided to have a bit of a look.
> 
> First I created an empty stack (no controls or anything) in Rev then I 
> made 2 standalones,
> one with the Rev builder and one with MC, these I then compared with a 
> hex editor.
> 
> Result: The Rev version has about 45k more of scripts in it including 3 
> references to stack "Home" (MC version has none).
> 
> It would appear that Rev is including scripts that are not actually 
> needed,

They are needed for any stack that uses the rev-based commands that are 
listed in the Rev dictionary. There are so many users who do call those 
handlers that I think it makes sense to embed that library as a standard 
extension. Tech support would collapse under the volume of complaints if 
it weren't in there; users would assume that Rev's standalones were 
broken. Now that I think about it, quietly embedding a library that 
doesn't get in the way of anything is a pretty good decision for 
supporting ease of use, especially for novice users who may not 
understand what libraries do or why they need them.

> how much this would effect speed is debateable though
> especially when run on modern hardware. Having conducted some rudimentry 
> comparisons on my machine
> (AMD 4200+ x2, 1gb DDR400, Win XP 32bit) using a variety of stacks 
> (including "ImageData Toolkit")
> I notice no difference in the execution speed.
> 
> the size difference between the Rev and MC versions of the ImageData 
> Toolkit
> that I made is considerably larger than between the two empty test 
> standalones (128k as opposed to 45k),
> speed does not appear to be affected.

I use both IDEs routinely, and I've never noticed any speed difference 
either. Something else must be going on with certain stacks.

The differences in sizes may be due to changes in the standalone 
settings; did you use the same settings for both sets of test stacks and 
their apps? It may be that more options were included in one set than in 
the other.

The extra 45k of scripts you noticed were most likely the revCommon 
library (which I mistakenly refered to as revGeneral before.) But Kevin 
confirms it uses almost no system message handlers, so it couldn't 
interfere unless a stack needs to intercept a lot of mouseDoubleUps.

-- 
Jacqueline Landman Gay         |     jacque at hyperactivesw.com
HyperActive Software           |     http://www.hyperactivesw.com