Revision history of the TMA-Combiner during development.

Last development version was 0.99 beta (2-24-05).

• Modification: Some changes to the user interface were made, mainly that involving the Score Key, which was simplified to reflect the new capability of handling quantitative scoring systems.

• New Feature: A more general scoring system is now supported. The user must specify what numbers correspond to equivocal values and data-missing values, and these numbers cannot be ranges. Otherwise, semi-quantitative and quantitative scoring systems are now supported. Note: because we do not use such a scoring system, this feature has not been extensively beta tested.

• Modification: Further simplified the code and began implementing a more general score handling ability. Note - this version was never released.

• Modification: Cleaned up the code and added some comments to clarify the score combination rule section. Certain sections still referred to the process as "compression" instead of "combination", so those have been fixed.

• Modification: Significant changes have been made to the user interface (UI). The UI has now been cleaned up and compacted, in preparation for publication. Furthermore, the score combination rules have been renamed and shoehorned into 3 rules. Old Rule 4 = new Rule 1, Rule 2 did not change, and Old Rule 5 = New Rule 3. All references to "Compressor" have now been changed to "Combiner".

• New Feature: The Compressor can now detect a couple new file formats and will do its best to handle them. Please refer to the file format requirements for details.
• New Feature: If a number or symbol is used to delineate missing data values, the user can now specify that in the Score Key. The default is a blank cell.
• New Feature: If the column name contains an underscore ("_") character, everything after the underscore will be removed, allowing columns with common name stems to be combined together. Please refer to the file format requirements for details.
• Modification: Significant changes to the user interface, particularly a clarification on the meaning of unconverted and converted scores (it used to be "pre-deconvoluted and post-deconvoluted" scores. Also tidied up the user interface.
• Bug Fix: Compression of datasets consisting of 3 or more files results in the dropping of the rightmost columns of the combined dataset. This has now been corrected.
• Bug Fix: EWEIGHT row in output file would completely fill out all the way through Column IV with a value of 1, if there was only 1 data column in the combined dataset. This has now been fixed.

• New Feature: Rule 5 (Use lowest score among good data) was added.

• New Feature: Compression implemented for columns. Before this, the Compressor could only do row-by-row compression.

• New Feature: Partial implementation for compression by columns.

• Bug Fix: Various bug fixes (not documented).

• New Feature: Users can now specify if the scores are in a pre- or post-deconvoluted state.
• New Feature: Additional score selection rules have been implemented.

Modifications:
• Removed (beta code) message at beginning
• Changed the # to (2), and removed the * for (3) and numbers > 3
• Made the actual compression process less epileptic seizure-inducing
• Did some error-handing in cases where the user neglects to specify the output filename and where the user neglects to specify the input filenames.
• Made the extra NAME and SYMBOL columns go away.

Earlier versions date back to 0.65 beta (8/8/2001), but I doubt most users would be interested in the development details. Email me if you think otherwise.