Documentation about the XperCdmAdapter¶

For easier understanding of Xper² and the Adapter

• Table of contents
• Documentation about the XperCdmAdapter
  • Architecture
  • Xper²
    • Where to find what

Architecture¶

The current architecture tries to follow the following rules:

• Data needed for all computations such as taxa (xp:individuals) or features (xp:variables) are loaded to xper during start up.

• Data needed only on certain requests and not for high performance algorithms are requested only on command and not stored in the system at all. Such data are media (xp:resources) and comments (xp:base.description, xp:individual.varcomment). For this we need to rewrite Xper in a way that retrieves such data always through the IBase interface or the IControlerBase interface, never directly from the base object it is attached to in the original xper version. This is to save memory.

• For modes(quantitative, categorical data) still need to be decided if loaded on start up or on request. Depends on computation demands for discrimination evaluation (accessable e.g. via right mouse click in taxon list).

• Changes to all objects should immediately be stored to the database no matter if data are available in main memory or not. Explicit saving will never be necessary.

• All taxon editing will be disabled. Discussion still needed on how to handle taxon grouping and hierarchies. If a workingset includes a classification only available for the workingset editing of hierarchies may be acceptable. However, most users will have a need to have classification in line with there general classification used in the CDM. In this case case hierarchies should not be editable. Groups are currently not availabe in the CDM. We could allow volatile group which do not persist. Maybe also a technical annotation could be used for attaching a taxon(node) to a group in a persistent way. Currently groups are disabled.

Xper²¶

Where to find what¶