Christian Holm Christensen writes: > > On Wed, 24 Oct 2001 15:00:44 -0400 > Brett Viren <bv@bnl.gov> wrote > concerning "Re: More on Graphics abstraction (was Re: [ROOT] Qt ROOT)": > > Christian Holm Christensen writes: > > > However, since most modern Graphics libraries uses > > > signals and slots, and ROOT has that too, I guess what one could do is > > > to connect the Qt/Gtk--/... signals into slots in the wrapper, which > > > would then transform that into a ROOT signal. > > > > To butcher Orwell, "not all signal/slots are created equal, some are > > more equal than others". > > > > It's true that Qt's sig/slots are close to ROOT's Rt which isn't > > surprisingly, since Rt is dirrectly based on the ideas in Qt's > > sig/slots. > > > > However, Gtk--'s signal/slots, implemented in the libsigc++ library, > > are different enough as to make mixing with Rt a little tricky. One > > could connect an Rt signal to a libsigc++ slot but not, in general, > > vice versa. > > Oh, that was never ever my intention (if that's the impression you > got - sorry for not being clear enough). To slaughter "Highlander": > "there can be only one" GUI factory in any application, hence no need > to send signal from Qt to libsigc++ or vice versa. Not Qt to libsigc++, but Rt to Qt or Rt to libsigc++. I understood you to be suggesting interface Rt to Qt's (or alternatively, Gtk--'s) signal/slot mechanism. I am saying it would be difficult to connect Rt on the ROOT side to libsigc++ on the Gtk-- side in any general way with out severely limiting what kind of sig/slots can be used on the Gtk-- side. However, one could entertain the notion of ripping out Rt everywhere in ROOT and replacing it with libsigc++.... > > > This is because libsigc++ is compile-time type-safe while > > Rt is neither ... > ^^^^^^^ > ||||||| > Did you mean to say libsigc++ is run- and compile- time > typesafe? I guess you did. No, sorry, I was a little unclear. I was just issuing my standard complaint about Rt: It doesn't allow for type safe connections and any little errors (such as typos in the strings you must pass, forgetting to add the slotted class to LinkDef.h, etc) must show up only at run time, if they show up at all. Since libsigc++ is template based the compiler fails/warns if there are type mismatches. And since the signal connection is not based on passing strings, but real live objects, a typo is caught immediately. ... > I guess it really boils down to the question: > > Should ROOT provide a full graphics abstraction, so that user ROOT > GUI code can run on any platfrom, or should ROOT simply implement > the GUIs that ROOT need for all supported platforms, leaving the > user to use the GUI library she prefers. > > Personally, I prefer the former, but I think the later is probably the > most reasonable choice (graphics libraries are rather good at what > they do - and ROOT sure could use a SQL query mechanism). Yes, this is the question (although I don't understand the reference to SQL queries, maybe something in Qt?). But, I think the answer to this "A or B" question is "yes", that is, "both". And, thankfully, this is looking like what we will get (via QtROOT's trail-blazing). > > Anyways, this is all probably more than you wanted to read... > > No, not at all. I really wanted to some more spicy arguments, and you > delivered - thanks. Heh, my pleasure. Although, I didn't mean to be inflammatory. -Brett.
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:51:04 MET