> >> On Sun, 20 Feb 2005 18:43:00 +0100 CET, Jonas Sundström > >> <jonas@xxxxxxxxxxx> wrote: > >> > I think it would be much better if the system could provide > >> > the necessary uniformity, to keep everyone from rolling their own. > >> > I would much rather have a good system-provided scanner dialog, > >> > (with possible extensions, should they be necessary). > > > > Actually I got that idea recently, and, humm... Sorry, can't tell more yet > > :p > > > Argh, don't talk about it then! :) :P btw, seen Philippe Houdoin ? Anyone who can assert the license for Sanity ? (bebits says MIT but I'm not so sure. > > > > Yes, BeHappy does that with NetPositive. > > You just need to create a BMessage with the correct field and pass it to a > > BShelf. > > > So what we need it to make the Replicant framework easier to use, and > provide for higher-level mechanisms(sp?) for interaction between the > container and the embedded app -- preferedly, the client (i.e., embedded) > should have the Replicant stuff built-in as much as possible (I think the > default API is a bit trickier than necessary), and that a way of exposing > the methods supported by the client in a high-level definition language. > We have most things in place for supporting a module-based (task-based, > even!) environment, with the Application Kit and whatnot. I've been thinking about extending the url-handling protocol... (apps tell they handle files of type application/x-vnd.Be-URL-foo) telling if they can be replicated. That way a browser could handle plugins in a simpler way, apps could embed VNC or RDP session by searching for a replicant that handles vnc:me@foo:0 ... > OpenDoc Haiku-style :) embedding apps != embedding docs... I'd say sending the replicant a pointer to a BDataIO it could use to load/save its document would be generic enough... the holding app would then archive the (MallocIO or whatever) in the way suitable for the container doc (base64 for mail, ...) François.