Evri release management

Eigenvector Research Inc. follows industry standards for software production. EVRI release management process includes the following:

Release Management

EVRI release management process includes the following:

• Continuous development - EVRI staff all work from the latest version of products so any changes are immediately tested in real world use. Typically a staff member will update from the repository at the beginning of work day. Note, occasionally staff will work from an existing release to maintain version compatibility with a customer.

• Version Control - EVRI maintains modern version control software and processes to manage software development.

• Issue Tracking - A software project management system for bug and enhancement tracking is used to document and manage issues.

• Testing - A mix of manual and automated testing is conducted to verify product quality. Each major release is regression tested to prior releases to assess numeric consistency.

• Release Notes - All significant changes and known issues are documented in the release notes.

Installation and Tracing Information

• All software is tested for common errors at the time of installation. A message appears in the installation window indicating the tests are being run. If an issue is encountered a warning dialog box will appear, otherwise the installation will proceed to open the main window.

• Solo and PLS_Toolbox have some tools to help trace model development but these are not a part of a certified regulatory environment. Tracing model development can be done using the following functionality:
  • Model detail.history field contains an entry log of datetime and changes made to the model.
  • DataSet .history field contains a running history of commands that have modified the DataSet contents.
  • The Modelcache stores models and data upon calculation in the analysis interface. The model cache interface keeps a complete history of calculated models.

Model Validation Information

It is recommended that customers develop a model within Solo/PLS_Toolbox as COTS (commercial off-the shelf) software. Then validate the model to their own standards paying special attention to how a model performs with previously unused data that spans the range of the samples you expect the model to work with.

• Steps to consider when validating a model.