Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!csn!stan!morgul!toml From: toml@morgul.Solbourne.COM (Tom LaStrange) Newsgroups: comp.windows.x.motif Subject: Re: Neuron Data's Open Interface Message-ID: Date: 20 Jun 91 15:45:34 GMT References: <9106191950.AA25259@25259> Sender: toml@solbourne.com (Tom LaStrange) Distribution: inet Organization: Solbourne Computer, Inc. Lines: 27 In-Reply-To: jerryl@is.Morgan.COM's message of 19 Jun 91 19: 46:47 GMT > I wonder if these experiences have something to say about real world > object-oriented software in general. Tom, can you shed any light on > your experience with the performance of the OI toolkit over there at > Solbourne? The original developers of OI made the same assumption that everyone else seems to make the first time writing applications using X, that windows were cheap so it's ok to use lots of them. This turned out to be less than optimal for the cases I stated in an earlier message, i.e., large menus, scrolling lists, etc. Currently in OI the only objects that are windowless are menu cells. Menu cells are also the object used to get what motif would call a push button. Menus and buttons paint very fast and you as a programmer don't care that they're windowless. They can each have there own fonts, colors, and callbacks. We have some very large applications written in OI and about the only time we really saw performance problems were when we were dealing with thousands of windowed objects. Most of the times, these things showed up in scrolling menus. Since converting the menus to have windowless cells, many of these problems have all but disappeared. In the next release of OI we will have much better support for windowless objects as we are moving to do InterViews style graphic objects (lines, circles, etc.) -- Tom LaStrange toml@Solbourne.COM