On Mon, Apr 12, 2004 at 10:39:05AM -0500, [EMAIL PROTECTED] wrote: > Do you think Vis could possibly be shoehorned into http://wxwidgets.org, > or perhaps the developers of wxwidgets could be convinced to adopt the > Vis architecture?
It's still a widget-based GUI library unlike Vis, the core idea of which is to not be concerned with the presentation at the code level. Of course there could be an 'ui-wx' module for Vis, but writing that is just as much work as a module for any other widget library. > Gnome, KDE, Windows, and OS X are in my opinion converging on some type > of mutated desktop metaphor placed in stone by the first macintosh. > There are still innovations in user interface design taking place within > that metaphor. But these innovations are in part kludges over bad > original design, Apple's expose for example. Yet even with this > convergence few applications are written to be crossplatform, by > utilizing something like wxwidegets. > Bearing all this in mind do you think Vis would ever have a chance of > getting a foothold in the developer community? I'm not overtly optimistic about it (see the Conclusions section of the paper). Good stuff seldom is popular (and this applies to almost everything else besides software). > The notion of user configured interfaces I do not find compelling > because the configuration burden placed on the individual is too great. Users should have the chance to configure their interfaces, not be required to do so. With something like Vis, someone can provide the fine-tuning stylesheet (if needed) for a particular UI style and then everyone else who likes that type of UI can benefit from it. > Also I think collaboration and user testing on interfaces is very > helpful. For instance I would really like to see some standard key > bindings for Ion developed by the community, e.g. Vi, Emacs, Joe [for > Tumo], Windows, Mac OS X. Also the uniformity of some type of standard > allows for easier adoption of new applications by users. I am always > reticent to proselytize for Ion because I can't easily tell someone > how to begin using ion when I have my own personal bastardized vi like > bindings. I've thought of making the Ion scripts page <http://iki.fi/tuomov/ion/scripts/> a darcs repository <http://abridgegame.org/darcs/> as a poor man's substitute for a Wiki that supports file uploads. apt-get install darcs darcs get http://iki.fi/tuomov/ion-scripts cd ion-scripts ... do some modifications and 'darcs record' them .. darcs send (to send a patch bundle to me by email) This way there would be one central place where anyone could submit their bindings and other scripts without going through lots of trouble. The repository could, of course, also include a Makefile or such to install the scripts making it even easier to start using one. > With this in mind perhaps development would be better focused on a > gui library on the same level as gtk or qt designed for expert > interfaces. This library could be used with existing libraries such > mozilla's gecko engine to build new expert applications. Software is seldom designed these days with anything but the least common denominator in mind. Who's going to write that "expert word processor" compatible with the ones for "normal people"? Of course, when I can, I just use LaTeX, but there are times when documents are wanted in some format the only way to produce which is with some application with an awful UI. XML formats are as bad as binary formats in this respect; they're not meant to be edited manually, and the programs available for that job them suck compared to almost any decent text editor and a non-verbose format. But if using the special purpose editor was almost like the using text editor I like... -- Tuomo
