Re: message probes

[Ginger Booth]

Hi, Manor,
    Time to add my 2 cents' worth...  I like Sven's rule of thumb:

> Note, I'm not saying I can't live with changes; I'm sure we all have 
> wishes for improvements we'd like to see to our toolkit. All I'm saying 
> is, before making a change, contemplate the ratio
> 
>       N = (everyone else's recoding time) / (Manor's coding time)
> 
> If it exceeds e, or perhaps pi, reconsider ...

And perhaps add another:  

    The purpose of Swarm and all of our applications, is simulation.  Any
changes requiring code overhaul should be justified by significant opportunity
in the field of simulation, such as batch job capability a la the last release.
(You'll note no one complained much about that, because we all understood the
necessity -for simulation-.)  

    The fact that Swarm provides gui tools is an also-ran.  Until and unless
someone builds an app on Swarm that has significant gui needs and a wide user
base, requiring overhaul for anything gui is not-nice.  So hold off gui changes
until releases where important simulation changes require modifications,
anyway.

    In practice, you'll find at least half your user base won't upgrade to 
new releases unless you come up with a very nice balance of benefits, and the
key to that balance is being clear on what matters most.  If you -have- to
make gui changes for some arcane reason, find a very nice carrot to go with it.

Ciao,
    Ginger