ExtPhr32 - Updates since 2001

2006 October

The package has been recompiled under Borland's free (100-year license) Delphi 2006 for Win32 Explorer.

2005 July

A new phrase collapsing routine has been substituted for the old one, which was unacceptably slow for longer phrase lists.

2005 April

A list of special stopwords across which phrases will not be extracted can now be specified.

2004 December

A "Copy as keywords" function now allows copying into a format suitable for pasting into the contents of an HTML "keywords" meta tag.

2004 September

A minor change in the file comparison tool now allows it to be applied to analysis files that show full phrases.

2004 May

Extraction from text on the clipboard is now possible.

2002 July

Output in lower case can now be selected.

Users can now set properties of the font used to display extraction results.

A list of additional ASCII codes that will be treated as letters can now be specified.

2002 June

A utility function has been added to calculate the similarity in vocabulary between two files.

2002 May

A new option allows for elimination of phrases that form parts of longer phrases elsewhere in the list.

On startup, the program now looks automatically for a script file extpconf.txt.

2002 April

A bug introduced in the previous update that caused an empty display when "Show full phrases" was checked has been eliminated.

2002 March

A list of ASCII codes across which phrases will not be extracted can now be specified.

The option of not including numerals as characters in words has been replaced with an option to treat all words beginning with numerals as stopwords.

2002 January

When comparison to expected is selected, stopwords and words with negative values are now omitted, and sorting is enabled.

Scripts may now be run; commands available include 'includeextendedascii', 'includenumerals', 'maximumwordsinphrase', 'minimumoccurrences', 'expectationfile', 'extractfrom', 'saveas', 'showcomparisontoexpected', 'showfullphrases', 'showonewordonly', 'showmultiwordonly', 'showrelativefrequency', 'sortbyfrequency', and 'stoplist'.

The DOS programs FiltPhra and CompPhra are no longer distributed with ExtPhr32, since the file format is incompatible. FiltPhra was of little use in any case with files produced by recent versions of ExtPhr32. CompPhra has been replaced by a new "Compare" function in ExtPhr32.

New options allow the user to set the maximum number of words in a phrase to any value from 1 to 20 or to specify whether numerals or additional ISO-8859-1 ASCII letters are recognized as parts of words.

The application has been ported to Delphi Professional version 6. One adverse effect of this is that the executable file is now much larger.

2001 September

When "Show full phrases" is selected, phrases ending in stopwords are no longer displayed.

A problem with long items being split during sort by frequency has been eliminated.

Extended ASCII letters with diacritics are now included as alphanumeric characters and are correctly converted to upper case. (Sorting is still in ASCII order.)

2001 May

The value of Simpson's l is now correctly computed for longer files (provided the total number of word repetitions does not exceed 2 billion). The word count for the file is also reported.

New display options have been introduced to show only single words or only multiword phrases, relative frequencies instead of raw counts, and, for single words, values derived by comparing observed frequencies with expected frequencies specified in another file.

Each frequency is now preceded by a tab rather than a space in order to make saved files more usable with other software.

2001 April

A minor bug that caused "Sort by frequency" to remain grayed when an extraction was performed with "Show full phrases" checked has been fixed.
Last updated February 5, 2008, by Tim Craven