On Wed, 2006-09-20 at 14:52 +0400, Maxim Udushlivy wrote: > Murray, I was thinking once that you present gtkmm as something superior > to GTK+ on purpose, but recently I changed my opinion. Your desire to > expose gtkmm in the most advantageous fashion is understandable. But how > can you explain this phrase from gtkmm website: "gtkmm is a GUI toolkit > and nothing more, and it strives to be the best C++ GUI toolkit"? There > are also strange things in gtkmm FAQ, for example "why use gtkmm instead > of GTK+" - note the word "instead".
There is no conspiracy. Really. C++ is in the funny position that you can use GTK+ *directly* from it, instead of requiring a binding. Gtkmm happens to be a C++ binding for GTK+, in case you prefer C++ idioms instead of C idioms. A frequent question from C++ programmers is whether they should use a C library directly or whether they should wrap it with C++ syntax and idioms. The Gtkmm FAQ gives a good answer to this question. You may feel indignant at the current wording of the FAQ, but it was not written that way to annoy you. You are just sensitive to wording. It's like when Arnold Schoenberg got sensitive that some poor musical chords were less likely to be used than others, and so he destroyed harmony altogether. > 1. State openly that gtkmm website contains subtle (unintentional?) > lies. (This was the most important for me personally.) Definitely unintentional. No more of a lie than saying "the sky is blue" in Mexico City. Everyone knows intuitively that the sky is supposed to be blue, even though it looks kind of brownish when you look out the window. > 2. Propose a position of a Moderator. This could automatically increase > signal/noise ratio of mailing lists; debates won't be endless; people > would afraid to lie; less politics and more activity; Gnome will be more > comfortable for people from Eastern Europe and Asia (some if not all > eastern cultures are not compatible with meritocracy!) This is not the first mailing list that suffers from endless discussions and high mail volume: look at the Linux kernel mailing list for a pathological example where people still manage to get work done. People have built tools to deal with that situation, and anyone can of course use them for GNOME's lists, too. You can ignore whole threads and nobody will mind. Making GNOME's customs more palatable to non-Western cultures is a *very* interesting problem, and one that we should definitely solve. Someone should start polling Eastern free software hackers into why they may feel that GNOME doesn't feel welcoming to them. Would you like to start such an investigation? [There *is* prior work in trying to answer that question... Alan Cox and others will be able to inform you.] > 3. Show the benefits of a split of Gnome and GTK+: this eliminates > technology and ideology mixture, so both projects could improve faster > since they have different goals. At this point in time I would say that GTK+ and GNOME need just a bit more integration than they have now, not more separation. A few years ago we had pretty much a genuine separation, and it turned out that GTK+ was not serving GNOME as well as it could. By "serving" I don't mean being subservient; I just mean that a slightly different philosophy would have been more productive for both. GTK+ tries to provide a good GUI toolkit for building applications, and GNOME is about building those applications and the environment in which they run. Obviously, GTK+ needs to be well-informed about the needs of GNOME, just as GNOME needs to be informed about GTK+'s plans for the future. And as usual, the best way to ensure that happens is by getting involved in both whenever you need it. > 4. Propose Gnome certification (originally I was thinking about it as a > replacement to the HIG, but I was shown that the HIG is needed, so it is > a complement now). This is in progress, please see http://live.gnome.org/GnomeCertification GNOME Certification was discussed extensively in this year's meeting of the GNOME Foundation Advisory Board during GUADEC. We want to use certification as a way of helping software developers know what they need to do to produce good software for GNOME, and we want a way to actually promote these practices: if you have a higher certification level, then it means that your application is better aligned with the goals of GNOME. > 5. Formulate the Gideon Principle and propose it as a cornerstone for > Gnome HIG and certification. What is the Gideon Principle? > 6. Discuss the possibility of contribution of Gideon to Gnome since > Gnome hosts GTK+. (I admit that I sounded frivolous on the > "Contribution" thread, but that's because I was not sure that such > contribution could happen and is fully sensible.) I'm sure that Gideon has a ton of interesting ideas and code that should be reused. Right now we have way too many GUI designers: Glade-2, Glade-3, Gazpacho, Stetic, Gideon, and probably others. Someone needs to take the initiative to stop reinventing this particular wheel, and see what is common across all of those GUI designers. Is it that they allow integration with IDEs? Is it the XML descriptions? Is it the mechanisms for plugging in new widget types? Etc. The task of "integrating libglade into GTK+" got postponed from GTK+ 2.10, because no one finished the work. That's the library side of things, but we also need the actual GUI designer and the integration with the IDEs. That's the initiative that someone needs to take to avoid so much reinvention. > 7. Propose to narrow Gnome by ideology (this is a short formulation of > an implementation style - not a HIG, but something like a motto: "Simple > interface, great functionality, coherent behavior") (In fact that's my > motto that I use when I develop GUI applications. It is a > user-understandable reformulation of the Gideon Principle) That's a nice motto, and it applies very well to some of GNOME's goals. It would be nice to distill that into some document. *Everyone* should read "The Inmates are Running the Asylum" and related books, but it would be nice to have all that philosophy distilled into a nice and short document for GNOME. Want to write one? Federico _______________________________________________ desktop-devel-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/desktop-devel-list
