Warning, /education/labplot/src/doc/data_set.dox is written in an unsupported language. File is not indexed.

 /**\page data_set Data set
 
 \section goals Main goals
 
 We want support for easily importing data from a variety of sources, including files, pipes, SQL-databases, measuring devices (real-time measurements?)
 and function generators. Furthermore, GUI should provide functions for easy editing of imported data sets without opening them in a specific view
 (spreadsheet, worksheet or editor).
 
 To set the terminology used in our project, we define some objects. The basic object for storing the data is 'DataSet'.
 Furthermore, we discriminate between following data sets:
 
 - Vector data set: data is organized in columns (vectors). Data can be numeric, text or date/time values,
 where the data type always applies to a column as a whole (in contrast to spreadsheet applications à la KSpread).
 - Matrix data set: data is organized in matrices, represents values of functions which depends on two variables.
 - Image data set: represents imported or generated image data.
 
 The actual origin of the data is encapsulated in the object 'DataSource'. Each 'DataSet' object has a 'DataSource' object.
 
 \section features Planned features (user's sight)
 
 - There is a container for all imported data sets - 'DataSetContainer'.
 - There is a 'DataSetManager' - part of the GUI, which provides all the functions for managing available data sets. 
 It is possible to generate new sets out of already existing, plots out of vectors from different sets etc. KST is worth looking at.
 - It is possible to convert the content of a spreadsheet to vectors and/or matrices and to add them to the 'DataSetContainter'. 
 May be this should be done by default. Say, in the 'DataSetManager' all the columns created by hand in different spreadsheets are accessible.
 - 'ProjectManager' contains a folder 'DataSets' which represents the data set container introduced above.
 - 'DataSets'-folder provides functions for adding new data sets, updating all available sets, removing all sets etc. in the context menu.
 - Each child node in the 'DataSets'-folder corresponds to an imported/generated data set.
 The different origins of the data sets (file, database or generated from a function)
 are pointed out in the 'ProjectManager' by using different icons and different strings in the column 'Comments'.
 - Context menu of each 'DataSet'-node provides functions for updating, deleting, showing the data in a new spreadsheet or text editor,
 attaching the data to a an existing spreadsheet or worksheet etc. 'Edit'-function opens the Widget associated with the corresponding 'DataSource'
 (currently ImportWidget), where the origin of the data (e.g. the file name) can be changed etc.
 Update-functions ('Watch the file', 'reload the file', 'stop/continue reading from the pipe') trigger the update in the corresponding views, if available.
 Data can be displayed in a spreadsheet or in a text editor in read-only mode on in a mode which also allows changing of the original data.
 Saving of the changes modifies the original data in the corresponding data source, except for data set which are generated with the help of a function.
 - There are 'Add data plot' and 'Add function plot' fictions in the GUI which help to create/generate a plot from imported/generated data set in one step.
 
 \section arch Architecture (developer's sight)
 
 (The contents of this section are still subject to discussion. Everyone is encouraged to provide suggestions and ideas to improve the architecture.)
 
 We need an object for storing imported/generated data. The general interface for that is provided by 'AbstractDataSet'.
 
 AbstractAspect
 ¦
 `- AbstractDataSet
         ¦
         +- VectorDataSet
         ¦
         +- MatrixDataSet
         ¦
         `- ImageDataSet
 
 'DataSetContainer' is a one the central parts of the application and contains all the data sets available at run-time.
 
 AbstractPart
 ¦
 `- DataSetContainer
 
 The view of the 'DataSetContainer' is the 'DataSetManagerWidget'.
 
 Data can be imported from a directory, a file on the disk or from a database. May be it's better to first concentrate on these sources only.
 Importing of data from measuring devices reduces on Unix-systems to reading out a device file in /dev/ or a pipe created and filled by the measuring device.
 Furthermore, we need an abstract interface that allows users to add support for data sources using plug-ins or scripts.
 'AbstractDataSource' provides the common interface for all possible data sources.
 
 AbstractDataSource
         ¦
         +- DirDataSource
         ¦
         +- FileDataSource
         ¦
         +- SqlDataSource
         ¦
         `- FunctionDataSource
 
 Different formats of data stemming from file- and sql-sources (ascii or binary, separator, number of lines to skip, numeric or date/time or strings,
 data organized in columns or rows etc.) are handled by pre-defined or user-defined I/O-filters,
 which convert the content of the data source to the internal representation in terms of vectors or matrices.
 Custom settings are made in the ImportWidget. It is possible to save the made settings as a new filter.
 Data in the source can be read once, checked for updates automatically ('watch the file'-function in the GUI),
 in defined time intervals or the update is triggered by the user himself. In the last case, the GUI signals if the source was changed
 (e.g. 'changed'-icon in the project manager).
 
 \section status Current status in the back-end
 
 The back-end code from SciDAVis already contains some concept of a data source: 'AbstractColumn'.
 A better name would probably be 'AbstractVectorDataSource'
 ('vector' in the sense of a mathematical vector in R^N where N is the number of rows and R can be real numbers (or rather the subset that 'double' supports),
 strings or dates; the corresponding data abstraction for 3D surface plots would be a matrix), at least from the developer's perspective.
 Currently, however, there is only one real implementation of this interface: 'Column'.
 
 Labplot1.x contains already some pre-defined settings in the 'ImportWidget' and has filters for reading CDF, HDF5 and Origin-files.
 Those should be casted to the new I/O-filter-plugin-Framework.
 
 
 */