Hi Anton, I don't see what is there in my answer to disappoint you. As you could have seen from my reaction to the ROQT announcement by the GSI people we are very happy that they made this important development and are very willing to integrate their modifications into the ROOT system so people can use the ROOT and Qt toolkits together (but as you noted, before that happens still some problems need to be fixed). Besides making new developments, Rene and I have to support a large number of users and collaborations. Our priority is to have a very powerful data analysis system (advanced very efficient object store, data analysis algorithms, vizualization, parallel processing, distributed data access, etc.). Having a native ROOT GUI is important but not fundamental. It's main goal is to allow users to make simple but non-trivial GUIs to access the data analysis components, all within one framework. Except for two or three widgets (like QLCDNumber and QSpinBox) the ROOT GUI supports the same widget set as Qt and now with the recent introduction of the signal/slot mechanism provided by Valeriy Onuchin the ROOT GUI is fully scriptable. Qt is quite nice but not a practical solution for ROOT Windows users. In that respect an interface to gtk might be better since it supports also Win32. From this fact, that people might want gtk support, it is clear that we cannot and will not provide all these interfaces ourselves. Just too much work for non-core functionality. If people want those interfaces, they can make them and, if properly written and documented, we will be more than happy to adopt and maintain them. Since three years we don't have anymore a core Win32 programmer in the team. This is the main cause that we don't have our GUI on Win32 and that we stay with our obsolete and clumsy attribute panels. While we have many users on Windows we have very, very few Win32 contributors. Maybe we should do like the TrollTech and make the ROOT Win32 version of ROOT commercial? Like everything else in the Win world. So like I said before, yes it would be nice to have a fully functional Qt interface and we fully support that, but we don't have the time to make one ourselves. Cheers, Fons. On Wed, Nov 08, 2000 at 05:25:48PM +0000, KOSU_FOKIN@garbo.lucas.lu.se wrote: > Hi Fons and rooters, > > Fons ... your answer as well as the overall situation around > ROOT GUI are rather disappointing. I am inside the "root > gui" business for about two years and frankly speaking > I do not see any radical changes or even a sign of them. > Java GUI, native ROOT NT GUI projects seem to be not very > successful and far from being completed in the nearest > future. We have also learnt that people come and go > but not much of their developments go into root, especially > if we talk about GUI where things are quite system-oriented > and require deep knowledge of low-level aspects of an > underlying API. > > I respect Fons and his attempts to create an in-house GUI > framework but at the same time I understand that it is not physically > possible to compete with the teams of specialized programmers > developing professional GUI libs. This means root will always be > behaind the leaders and this makes me sad. Whatever brilliant classes for > data analysis I/we produce it is not competitive if I/we put them in a box > standing below standards. Of course it is not even needed in most cases > to have a full house of fancy widgets but for serious applications > this is the most critical point. I do not even talk about commercial > use of root but about my experience with root native IDE, > a modification of my r-quant for daq and on/off line hep data > analysis. A number of people use it now and I am tired to answer > their questions about missing features in the gui/help system/etc. > They all work with windows and they expect the same standard features from > any program. So either you kill the Bill or you follow the standards. > > Well, in short, don't you think it would be nice to put some > effort in development of Qt-root properly working interface? > > We saw from Denis&co work that it is possible... in principle. > On the other hand I have noticed that their interface is far > from being completed, i.e. TRint should be included in the > event loop, QRootCanvas doesn't functioning properly in some > cases, etc. Also, I think it might be possible to use native > root widgets running under CINT... or even use two event loops. > Why not? > > I do not think that Qt license limitations can prevent us from > using it. First of all it is free under linux and root team doesn't > need to include the qt source in the root distributions if Qt use is > optional. It must be included if one wants to redistribute his/her > qroot application. Or buy a license. In fact I do not understand > all this noise around Qt licenses. There is no free lunch and good > things cost money. I prefer to pay 1.500 greens from my pocket instead of > waiting another two-three-??? years until root gui gets native simulation of > QTable or QTextBrowser or become ported on NT. > > Finally, I think Fons and Rene have made a very nice and outstanding > framework for data manipulation/analysis/visualization. Let's spend > time on those features, making root the best in the world of data > analysis. There is a number of roads to go. I do not think our goal is to > beat Qt or MFC with root gui. Let's not spend much time for that! > > > Regards, > Anton > > -- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland E-Mail: Fons.Rademakers@cern.ch Phone: +41 22 7679248 WWW: http://root.cern.ch/~rdm/ Fax: +41 22 7677910
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:36 MET