RunRev vs RealBasic

Jesse Sng jsng at wayoflife.org
Tue Jan 4 22:00:05 EST 2005


>Yes, I could use externals or you could argue that rev
>is UI/multimedia-centric and not intended for serious
>scientific computing, but does that have to mean that
>it could never be?

Why not just use Revolution for all the things that it is good for 
and then pass the rest of the nasty stuff over to another app?

On the Mac, you can AppleEvents to send the data (or even via a file) 
to another app that is faceless and this will work for both OS X and 
OS 9. If you want cross platform, then implement something using 
sockets to communicate XML data back and forth.

That way, you can have a natively compiled app that can multitask and 
do all the heavy hitting, without having to compromise on the time 
taken to develop a good user interface and operating environment.

The interesting thing will be if you have faceless BSD command-line 
type apps, you can fork multiple instances of that process and return 
the results, while you still have a front end that is free for the 
user to do other things.

I've done that kind of stuff before using Supercard and a C++ app 
running in the background and I didn't have to tie the front end to 
waiting for something to be completed before I allowed the user to do 
something else. I would think that this is a good way to go as 
Revolution doesn't support multiple threads of execution at this 
point in time.

The whole front end can be event driven in that once the faceless app 
finishes, it could send an AppleEvent back and this triggers the 
front end to do some appropriate handling and transformation of the 
results into something that it can use for display purposes.

That being the case, you could in theory have 1 front end app and 
have the these background apps run on multiple machines on the 
network and this would be very nice for scientific apps that may need 
quite a bit of computing.

I had a front end control a whole network of kiosk macs using 
AppleEvents before and everything was done via one single machine 
using a stack based interface. I even went so far as to Timbuktu 
remotely via a dialup connection into that administration machine to 
control the entire network remotely without having to travel 30 miles 
to get to that location.

I would suggest opening your mind up to the possibility of using Rev 
on the front end and using RealBasic, Objective C or whatever you 
wish on the backend and have a client-server type configuration. All 
this can be easily done using Revolution and RealBasic in concert.


Jesse


More information about the Use-livecode mailing list