From kde-core-devel Thu May 15 10:50:44 2003 From: Datschge Date: Thu, 15 May 2003 10:50:44 +0000 To: kde-core-devel Subject: Re: Moving QExtMDI to kdelibs X-MARC-Message: https://marc.info/?l=kde-core-devel&m=105299584405234 I have to say I fully agree with Neil on this topic. Neil Stevens wrote: > "MDI should not appear in KDE without explicit user intervention to > turn it on." * First step: Every KDE App should first start in SDI mode in every case. Users in general are easily overwhelmed of complexities they don't experienced and understand yet. SDI allows users to easily organize the amount of stuff since everything they started is represented by exactly one window, connected with exactly one taskbar and exactly one Alt+Tab switcher entry. Closing the window gets them away with it again, and there is no hidden stuff which could be missed or even lead to loss of data. * Second step: "Primitive MDI", organizing multiple windows at once with Kwin. As soon as a user is familiar with the interface and behavior of the programs he uses he won't be afraid of using more and more of them at once. This raises the issue of available screen space as well as the missing possibility to combine several windows together based on specific tasks (let's ignore the rather obscure possibility to use several virtual desktops). This ability should be offered globally by Kwin, while retaining independence of every (former) SDI regarding taskbar and Alt+Tab switcher entries. * Third step: "Advanced MDI", the MDI we all hate unless we actually want it. Using the consistent MDI capabilities of Kwin the user will be able to quickly familiarize with it and most likely ask for additional goodies. Want to save and close all SDI's in one MDI at once? So be it. Want to hide all SDI's and instead show the one MDI containing them as taskbar and Alt+Tab switcher entry instead? So be it. Want to have additional features through communication by combining eg. a file list and an editor in one MDI? So be it. Let the user dig in and stick together what he wants to be together and offer logically resulting features. The end result is the full force and heavily blown MDI users use to hate, hate because they didn't get to decide themselves what they get there and how it works. Instead they were confronted with a ready made confusingly complex block called MDI to which they need to get used first while constantly complaining about missing features available at the fingertip but in the wrong program. So what we should do now in my opinion is doing the second step before doing the third one. First let Lubos implement a Kwin based MDI, using QExtMDI if it helps, not using it if it doesn't. After that everyone can work on communication between SDI applications through the MDI framework offered by Kwin. After that the whole SDI vs. MDI discussion in only up to the user's personal experience and preference. Thank you for your attention. Cheers, Datschge =)