jvkr

Since I'm in feature request mode, allow me to make one more :)

It would be great when autocompletion could be adjusted so that an order can be made concerning Text Completion suggestions between objects inside the max folder, those elsewhere (3rd party) and abstractions. The best example as to why is when I want a groove~ object: the first candidates are grooveduck1 and grooveduck2. Instead of autocompletion making the proces faster, it slows it down, because either I have to type the full name including the tilde, or I have to scroll down using the arrow keys, or I can click with the mouse. And I am never going to use grooveduck...

Like this there are many examples, and quite regularly I instantiate the wrong object because of this. If objects inside the max folder would be listed first, then you could always count on blindly making the correct selection. But maybe I'm the only one?

autocompletion-candidate-refinement

Anal-retentive response: how does having to type "groove~" become slower than having to type "groove~"?

Without auto-complete you have to type seven keystrokes; with auto-complete you type seven keystrokes. I don't see how typing seven keystrokes is slower than typing seven keystrokes (if you see what I mean).

Ignoring semantics, your suggestion has a certain cachet. I can imagine that some people, however, would like their custom abstractions searched first, followed by 3POs, with factory externals bringing up the rear. (And it doesn't take much imagination to come up with four further orderings.) So, does this get a user option somewhere?

Groovy~ idea, but needs a bit of ironing out, I think.

One simply gets used to auto completion, and it's supposed to be helpful, but kind of gets in the way more often than necessary(especially in messageboxes..this is really annoying).

I think there has been a thread about this already some time ago.

In general, I'd suspect built-in objects to appear first in the list, then externals, then abstractions. Why? Because I think there is a tendency that complexity is growing that way, and therefore frequency of usage is decreasing in that order. (wild guess, but sounds reasonable to me)

Anyway, I think there should be done *something* about this, because the behavior at the moment needs some ironing more than the proposed idea I think.

I want to second the JVKR,
One simply gets used to auto completion, and it's supposed to be helpful, but kind of gets in the way more often than necessary(especially in messageboxes..this is really annoying).
I think there has been a thread about this already some time ago.

Anyway, I think there should be done *something* about this, because the behavior at the moment needs some ironing more than the proposed idea I think.
cheers


Another thing about auto completion: I'd love to see the gen environment knowing about gen abstractions..

autocompletion candidate refinement (feature request)