[ANN] Inspect without clicking in 2.5? FREE!

Richard Gaskin ambassador at fourthworld.com
Sat Sep 4 16:38:49 EDT 2004

Alex Tweedly wrote:
> See Bugzilla 1884. I submitted a report saying that it shouldn't send 
> the mouse message to the stack, and Kevin disagreed and marked it as 
> "Not a bug". I can sympathize with his reasoning - though I still think 
> it would be an improvement for minimal, if any, risk.

Therein lies the beauty of SuperEdit.

The SuperCard package ships with a separate editing and layout 
environment called SuperEdit, which lets you dive into your project just 
like the runtime environment with one critical difference:  there is no 
script interpreter in SuperEdit.  This extra layout environment was 
rarely understood by other xTalkers and often dismissed as being "too 
different", even though using it was fully optional and the idea is 
common with VB, RB, and other more traditional environments.

As many in the biz have said before, the difficulty with editing at 
runtime is that it's like operating on a patient while the patient is 
squirming under the knife.

Personally I feel the pros of editing at runtime outweight the cons by 
far, and I'm not advocating RunRev make something like SuperEdit for 
Revolution.  But it does serve as a contrast which illustrates the 
challenges inherent in the  xTalk edit-at-runtime paradigm.

The vendor of a runtime editor (the Rev IDE) has to make a choice, and 
neither is always optimal:

  - Trap messages without passing them,
    which may prevent an object from exhibiting expected behaviors.

  - Trap messages and always pass them,
    which may also lead to unexpected behaviors.

The purist in me says that all messages must always be passed, but the 
pragmatist in me recognizes that there are times, like this one, where 
that may be less desirable.

The team that maintains the MetaCard IDE tries to employ a principle 
which we could call a "Pure Transcript Initiative", in which the engine 
is considered the primary arbiter of behavior.  Any deviations from 
normal engine behavior introduced by an IDE must come with options for 
turning them off, so the developer can see true engine behaviors just 
like their users will.

This "middle path" approach may be useful in this case with mouseUps: 
add a Preference to allow whichever behavior the developer wants.

  Richard Gaskin
  Fourth World Media Corporation
  Rev tools and more:  http://www.fourthworld.com/rev

More information about the use-livecode mailing list