PhraseContext - version info.


Build 1.02.70 Revised Sentence data: Added Stop-button to interrupt long tasks. Also revised some captions for better clarity and added multidelete and clear results. Speed improved due to better code optimisation. Simplified sorting of Listview data by simply clicking the header. The procedure toggles ascending and descending sort order. Fixed minute error in sentence counting procedure. Clusters: Sorted out some nonsense in the this routine.

Build 1.02.69 Changed Lemma search window so that it can be as wide as the screen to make concordancing of very long strings ( >10 words) easier. Added sorting of wordlist re: wordlength (ascending and descending). Changed click on Wordlist's word column so that it toggles between the 6 sorting types: Alphabetically (Asc + Dsc), Morphologically (Asc + Dcs) and Wordlength (Asc + Dsc); These options are still available from the Words menu item. Along with this the wordcount entry in the statusbar states the sort order. Added Word length chart: The length of all the corpus words are displayed together with the number of words of these lengths. Data presented in both numerical and graphical terms. (March 2006)

Build 1.02.68
Changed formation of collocations so that this procedure follows the general setting of word delimiters on the Preferences screen. The advantage of this change over the old one is that statistics are always calculated correctly since there can be no differences between the word forms that appear in the wordlist and the collocations scooped up by the collocation procedure. Fixed: Change case now also applies to words in the statistics grid.

Build 1.02.65-67
Fixed bug: If a saved wordlist was loaded the wordcount was sadly missing, thus a division by zero error occurred whenever this figure was needed for calculations. Changed the Compare Corpora procedure so that words not found in the default corpus are not just discarded (a lot of thanks to Wu Liangping for pointing this problem out to me and for testing the revised version). Clarified some captions concerning this function too. Added saving/loading of comparison data. Copying (to clipboard) and export of multiple selected citations implemented. (June/July 2005). Tiny bugs in word counting procedure and collocations engine discovered and fixed. Improved control of interpuctuation signs in collocations procedure and implemented user choice of span 1 - 5 left and right. Added count of types/tokens of keywords - relevant, for example, if a regular expression or another more complex search has been performed. Added delete of citations as pr. keyword. Also added a small option here and there. Try rightclicking to see if anything new has cropped up. Improved Collocations' Change Case so that it is both more reliable and faster than before. Still looking for a nice secretary to write the documentation!
Build 1.02.60-64
Collocation engine revised, speed improved considerably. Small fixes here and there.
Build 1.02.59
Annoying error when no default corpus text is specified fixed.
Build 1.02.58
Added comparison of corpora using Log-likelihood significance test. Implemented both support of wls and XML-wordlist files for corpora comparison. The former being by far the faster of the two. A number of tiny shortcuts and other utilities added. (Still at experimental level.)
Build 1.02.57
Revised registration details. Registration codes issued prior to this version will not work, but already registrered versions will work on the PC it was installed on. People in possession of legal codes can obtain new ones free of charge by contacting me.
Build 1.02.54/55/56
Special builds, no major changes.
Build 1.02.53
Small changes concerning interpunctuation in connection with clusters added to improve precision and clarity. Small error in registration procedures mended, - thanks to Stan Mikhalev for trying to bust me.
Build 1.02.52
Incorrect calculation of keyword index corrected. Revision of show/hide progressbar during search so that speed is actually improved if the progressbar is turned off.
Build 1.02.51
Tiny cosmetic changes to code and appearance.
Build 1.02.50
Added functionality for saving and loading a lemmalist (default extension:'lst'). Also relevant menu items for adding words from the wordlist added. Drag-and-Drop from wordlist to Stoplist and Lemmalist implemented. Major revision of the basic search procedure(s) to implement a standard wildcard search with ?, * and [..] expressions. Added algorithm to choose the fastest possible search procedure in any case.
Build 1.02.49
Additions: Lemmasearch: Scrollbars, Sorting routine, remove duplicates, warning with delete all, right-click edit menu. Citations: Copy selected to clipboard (missing space with export to Phrasebook added). A couple of warnings here and there. Small changes to main wildcard search routines. Typos corrected. Added morphological sorting of wordlist. Improved speed and functionality of wordlist's "Remove Html"-procedure. Simpler keyboard shortcut (Ctrl+l) for adding words to stoplist plus alert window (the former suggested by Andreas Aagaard). Speed of Batch Type/Token calculation massively improved. Wordlist generation window revised - and support for other mark-up characters than '<' and '>' added.
Build 1.02.48
Type/Token ratio calculation procedure added. First limited beta-release.
Build 1.02.45
February 2004: Completely new concordance engine based on the Boyer-Moore-Horspool search algorithm developed and installed. Reduced searchtime dramatically (Before: Finding and displaying 14.000 concordance lines in a 200,000 word corpus took 3 Sec., now just 440 mSec). It is now possible to work with very large corpora indeed. Faster loading of big single file corpus added. Clear citations, and some more citation sorting routines added. Another small bug in collocations engine discovered and fixed. Simple syntax highlighting of Regular Expressions added. The idea is to keep literals in black (unless enclosed in square brackets) and colourize everything else.
 
 

Back to PhraseContext frontpage
 
 
Hans J. Klarskov Mortensen (c) h@hjkm.dk