At 11:37 AM +0300 8/9/01, Douglas Wagner wrote:
>At 2:43 PM +0300 8/7/01, Douglas Wagner wrote:
>>>1 Given an application requiring, say, 20 or 30 windows, each of which is a
>different size, does it make any difference which window is defined as the "main
>stack"?
>
>And Geoff Canyon" <[EMAIL PROTECTED]> responded with:
>
>>Probably not, but first: what kind of user interface requires "20 or 30 windows?"
>
>Thanks for your comments Geoff. You've managed to give me more than I asked for.
>
>I've not counted the total I need for the work I'm doing. In fact 30 may be an under
>estimate. My FaceSpan project already has 24 and I'm not finished. At the moment I'm
>looking at Revolution as a possible replacement for FaceSpapn.
>
>1 My question speaks to methodology. I want to know which is the most practical
>approach to Revolution. Is there any overhead attached to having many stacks with a
>single window rather than a few stacks with a number of cards each?
>
>Even in the case were two windows are the same size, it seems preferable to have two
>stacks, rather than a single stack with two cards (and two layouts). I don't
>anticipate calling both windows simultaneously, although this may not always be true.
Ah! <light bulb goes on> You're not talking about 20-30 windows _at the same time_.
That's quite different. You can certainly do this as separate stacks, but you can also
do it as separate cards in a single stack, which would probably be the more
traditional way to do it. The drawback of separate stacks is that some things require
a little more work/a slightly unusual syntax. The drawback of separate cards is that
you can't show both at the same time.
>
>2 Rather, having each window in its own stack seems to make maintenance and
>coding more straightforward. Is that a reasonable assumption?
I would think the opposite, but as I said, both will work. Neither solution prevents
you from sharing code between displays.
>
>You have also addressed my question about speed. I was wondering just how fast
>Transcript, Revolution and an XCMD would be compared with with Facespan and
>Applescript as a way of addressing a database. (At first sight Revolution looks like
>a revolution compared to the former approach).
Depends on what's taking up the time. If it's the database access, then Revolution
can't do anything to speed that up. If it's the code in the system, then Revolution is
_much_ faster. If you're doing a lightweight database yourself (few hundred to few
thousand records) then Revolution is much faster.
>
>3 If you have experience using, say Java, to access a database. In terms of
>speed, would Transcript be closer to Java or to Applescript?
Closer to java.
gc
