ROOT

From Wikipedia, the free encyclopedia

For other uses of "root", see root (disambiguation).
ROOT
ROOT logo
Example ROOT plot showing regions of interest in a 2D distribution
Example ROOT plot showing regions of interest in a 2D distribution
Developed by CERN
Latest release 5.18/00 / January 17th, 2008
OS Cross-platform
Genre Data Analysis
License LGPL/GPL
Website ROOT home page

ROOT is an object-oriented program and library developed by CERN. It was originally designed for particle physics data analysis and contains several features specific to this field, but it is also commonly[dubious ] used in other applications such as astronomy and data mining.

Contents

[edit] Description

Development was initiated by René Brun and Fons Rademakers in 1994. ROOT is written in C++. Some parts are published under the LGPL, while others are based on GPL software and thus are also published under the GPL. It provides platform independent access to a computer's graphics subsystem and operating system using abstract layers. Parts of the abstract platform are: a graphical user interface and a gui builder, container classes, reflection, a C++ script and command line interpreter (CINT), object serialization and persistence.

The packages provided by ROOT include those for

A key feature of ROOT is a data container called tree, with its substructures branches and leaves. A tree can be seen as a sliding window to the raw data, as stored in a file. Data from the next entry in the file can be retrieved by advancing the index in the tree. This avoids memory allocation problems associated with object creation, and allows the tree to act as a lightweight container while handling buffering invisibly.

ROOT's focus on performance is caused by the amount of data that the Large Hadron Collider's experiments will collect, estimated to several petabytes per year. Physicists are expected to analyze this data using ROOT. ROOT is currently mainly used in data analysis and data acquisition in high energy physics experiments — most current experimental plots and results are obtained using ROOT.

The inclusion of the CINT C++ interpreter makes this package very versatile as it can be used in interactive, scripted and compiled modes in a manner similar to commercial products like Matlab.

[edit] Criticisms

ROOT has been criticised by users, claiming it has a monolithic design, a tendency to re-invent existing technologies, a too-strong reliance on the CINT interpreter, and poor use of object-oriented programming methods, including OO antipatterns.[1] In attempting to provide all the functionality required by any data analysis task through a single interface, ROOT has attracted "no silver bullet" criticisms. Many of ROOT's fundamental design elements, such as global state, STL-incompatibility and aggressive single-owner based memory management, can make integration of ROOT functionality with stand-alone code difficult. The C++ wrapping is superficial as most common actions are performed by passing cryptic text commands to ROOT object member functions, so users must navigate a C++ interface plus multiple ad-hoc scripting languages.

Additionally, some users fault ROOT for generally poor documentation. ROOT classes do not always behave as advertised or as expected.

ROOT is very difficult for users to begin learning, largely because the ROOT documentation and webpage hardly elucidate in any way what the purpose of ROOT is. The documentation repeatedly stresses the internal design characteristics of ROOT over what features it actually provides to users. It says that ROOT's main features are "The Runtime Type Information System, The Object I/O System, Automatic Documentation Generation".[2] In documentation marked "read this first", the user is mostly told things like "the backbone of the ROOT architecture is a layered class hierarchy".[3]

From time to time, issues of ROOT's design and implementation are discussed on the ROOT users mailing list.[4][5]

[edit] Application of ROOT

Several particle physics experiments have written software based on ROOT, often in favor of using more generic solutions (i.e. using ROOT containers instead of STL).

[edit] References

  1. ^ The problem with ROOT
  2. ^ ROOT webpage: ROOT's main features
  3. ^ ROOT webpage: ROOT Roadmap
  4. ^ ROOTTALK: 27 July 2006
  5. ^ ROOTTALK: 28 July 2008

[edit] See also

[edit] External links