Here is some answers to the post of Shahrokh Nasseri
 on the Geant4 user forum (analysis 159) to the two questions :

 1) What is the relationship between OpenScientist/Lab and ROOT ?
 2) What is the relation between AIDA and ROOT ?

  Most of answers can be found on the OpenScientist web pages,
 but I can't refrain to do here a little summary.

 About OpenScientist/Lab and ROOT :

  The OpenScientist/Lab package presents itself as a data 
 analysis package and is then a concurrent to ROOT. There 
 is no attempt and no wish now to rely on or be tied to ROOT
 in any way in the OpenScientist/Lab package. See the OpenScientist
 introduction web page for grounding motivations.

  The only remaining relationship is around the .root file
 format which is, helas, now spread in HEP for storing histograms 
 and tuples. The Lab package can read and write at this format  
 but by not using ROOT classes and by using the Rio package
 which is part of the OpenScientist integration. With Rio,
 OpenScientist can offer then a consistent environment fully 
 decoupled of ROOT. Someone must note that the .root file
 format is not the only one supported by OpenScientist/Lab, and that
 the AIDA-XML file format can do a lot without having to 
 "sign with blood" to any TUniverse.

  At some moment the CINT interpreter had been declared
 to OnX which is the core interactive package of OpenScientist.
 OnX can handle Python, Tcl or java and it seemed a good idea
 to handle also CINT, which would had been then a natural common 
 scripting grounding between ROOT and OnX. But the unability 
 to have a CINT standalone packaging that ROOT would use and would then 
 be shareable at installation time with OnX made this way unteanable.
  Another important point was that CINT was not able to compile the Inventor 
 headers, which was needed for doing graphic scripting in OpenScientist.
 (Is CINT able to compile the Inventor and Geant4 headers now ?).
  Moreover it appeared that CINT was awfully crashy compared to Python, 
 Tcl, or Java (or anything else in fact) ; so that, tired of painfull 
 debugging sessions, I turned to solutions with defacto 
 higher quality, and the OnX/CINT driver had been deprecated in the last 
 release of OpenScientist (15.0). Then this relationship to ROOT does not 
 exist anymore and be sure that it is a relief.

 ROOT and AIDA :

  AIDA is a set of pure abstract interfaces that presents itself as a user 
 API for doing data analysis. The ROOT founders never joined AIDA, propably 
 because they have the strong feeling that they are the only ones on this 
 planet that have the right to non-discuss and fix some API for histograms 
 and in fact for everything. AIDA, by principle of abstract interfaces, lets 
 free hands and a lot of freedom to develop various implementations with 
 various languages and this kind of ideas is clearly foreign to the ROOT 
 founders. Must be pointed out, that despite they claims against abstrat 
 interfaces, a lot of "abstract base classes" can be found now in ROOT.

  A priori it seemed to be a good idea to have an implementation of AIDA 
 based on ROOT. This had been attempted within the LCG/PI project at CERN, 
 but this project had been now killed by the nasty ROOT sociology existing 
 in this lab. In fact, put all together, it was a pure suicide to attempt 
 an implementation of AIDA over ROOT at CERN : if someone wanted to kill 
 AIDA at CERN, it was probably the best way to go.

  For more about LCG/PI, see the "Letter to LHCb about the LCG AA review" 
 written at the moment that LCG had decided to stop doing developement 
 on LCG/PI. It is on the OpenScientist web pages in the "slides, papers" 
 section.

 AIDA at CERN :

  In fact there is an extra question : AIDA at CERN ? If you have not
 guessed the answer, I let you discover some in the OpenScientist pages.

  What is sure is that this fancy lab, originaly created to federate efforts
 of European countries to do physics, never had the idea to promote 
 more advanced AIDA implementations done elsewhere ; especially the ones 
 done in the "state member" labs or non-European labs strongly participating 
 now to the CERN experimental program. Strange.

 ROOT and AIDA, some link around file format :

  The OpenScientist/Lab being an AIDA implementation ; a relationship
 still exist around the support of the .root file format for storing
 histograms and tuples. Must be pointed out that the jas3 java 
 AIDA implementation, done at SLAC, can also read .root files.
 (Then an extra question ; ROOT and jas ?). 

  For storage, we clearly want to look now to something more dedicated 
 and more mature coming from the open source and that would evacuate
 this problem once for all (HDF5 ?).

 ROOT and OpenScientist ; the JurassicLab package :

  In OpenScientist, there is anyway the JurassicLab package that gather
 various trial materials in integratating ROOT in various ways ;
 through the GUI, the graphic, the windowing, the interpreter, 
 etc.... For example, in JurassicLab, there is a Xt widget for the
 TCanvas (the RootCanvas widget). There is also the SoRootPad Inventor 
 nodekit, to try to capture the ROOT graphic within Inventor. Some years 
 ago I had even written code to have an OnX driver for the ROOT/GUI classes.

  But all these developements, that took a lot of time, finished 
 always at some point to be bugged down by some nasty ROOT technical points,
 most of the time related to the lack of a clean design of the ROOT classes.
 I have probably not the right fractal/holographic brain to follow 
 the intrication of the ROOT class diagrams and the "g" logic. 

  Due to the fact that there is clearly no way now, at the sociological
 and technical level, to put things on tracks ; my global 
 conclusion is that for most of things, if seeking a minimum of 
 software quality, someone has better chance if looking elsewhere.

  For the LHC it is now too late to do something. Key decisions had
 not been taken. The executive of LCG is now in the hands of the painfull ROOT 
 sociology and next year we will have to run with "that". The hope now 
 is around (young ?) people that will be responsible of software for further 
 generation of machines and experiments ; I wish strongly that they 
 will be more clear minded than the "LHC era".

 References :

  About the relationship OpenScientist, ROOT, CERN and AIDA ; most of the 
 argumentations can be found on the OpenScientist web pages :

   http://OpenScientist.lal.in2p3.fr

 See for examples my four CHEP'04 posters and papers that are in the 
 "slides, papers" section.

  With my best regards (if you are not yet adicted to ROOT)

   6 March 2006

    G.Barrand