left horizon span went beyond the number of columns estimated by the window span, no hits would
be reported. In some ways, this was a good thing as it showed that something is wrong with the span
settings. However, the results no longer matched those in previous versions of AntConc that simple
ignored the mismatch and produced results for the largest horizon span possible. I have now edited the
storage routine so that 3.4.1 also ignores the mismatch and creates results for the largest column size
possible. I am still unsure if this is best approach to take. I am considering if a proper error or warning
should be reported when a mismatch between the context horizon span and the window span arises.
Although not strictly a bug, due to changes in the way the Clusters/N-Grams tool treated line breaks
introduced in 3.4.0, results stopped matching those from earlier versions of AntConc (e.g. 3.3.5). In this
version, I have introduced an option in the Clusters/N-Grams preferences to either replace line breaks
with a space or other character (as happens with the Concordance Tool - the default), or leave line
breaks as is. If line breaks are left as is, the results will match those from previous versions of the
Although not strictly a bug, on startup with a user settings file, AntConc would not remember the size
or position of the main window and instead show it at the default size and position. This has been fixed.
When AntConc is launched on a Windows system with a magnified view (e.g. 125%), some of the
settings options were hidden from view. To compensate for this feature of Windows, I have redesigned
the settings windows so that all options are viewable when first launched in magnified view at 125%.
Viewing at 150% is likely to still hide certain options. I have also allowed the settings windows to be
resized so that even at 150%, it should still be possible to resize the settings window to see the options.
The way the program reads in files has been changed to avoid a strange bug that occurred when
reading in some Unicode files that resulted in the last character on a line being misread on very rare
occasions. The way file names are displayed in the program has also been changed to avoid a
potentially similar problem (although I have heard of no reports on this to date).
The "Range" heading of cloned Clusters/N-Grams tool results did not show correctly. This has been
The "Stat" heading of the cloned Collocates tool results was shown as "Prob". This now appears
This is a semi-major upgrade fixing a few small bugs and adding several important features to improve the
output display of the Concordancer and Plot tools.
The Concordancer tool output now correctly centers the search terms regardless of language. This is
especially useful for Asian languages such as Japanese, Chinese, and Korean that have two-byte
character widths. Note that this is achieved by using two columns in the 'KWIC' pane. If you click CTRL-
A in the pane, only one of these hidden columns will be selected. Use ALT-A to select all columns.
Line breaks in the Concordance Tool output can now be replaced by a user defined character (default
being a single space). Again, this is especially useful for Asian languages because traditional
replacement by a single space might inappropriately split words that cross line-break boundaries.
The Collocates Tool no longer includes the center word in the collocation table. This results in more
meaningful collocate values.
Not exactly a feature, but Concordance Tool searches are now only centered on completion rather
than after finding the first line. This is generally a more pleasant experience.
Whitespace options in the wildcards menu have now been extended.
A database now runs in the background and is used to store KWIC and Plot concordance results. This
alleviates out of memory problems with earlier versions and also allows much faster sorting of the
KWIC. The new database model also improves the performance of general KWIC and Plot searches.
In the Advanced Search, if you add multiple context words, AntConc will look for *any* of these in
surrounding context of the search term rather than *all* of them (as in earlier versions). This change
has been implemented as a result of many requests for this function. I hope it does not inconvenience
too many people.
In the Collocates tool, if you click on a word, if will search for that word in the Concordance Tool but
restrict results to only those where the original search term is in the neighborhood of the collocate as
set by the Collocates tool. Again, this change has been implemented as a result of many requests for
Colors for the highlight and ordered words in the KWIC display have been changed to be a little easier
The font size for the main results is now a little larger.
Mismatches in results across tools (e.g. different frequency counts for the Concordance, Plot, File View
and Clusters/N-grams tools) due to the different ways in which tools handle line breaks and spacing
have been fixed. In particular, by changing the default processing to replace line breaks with spaces,
the frequencies reported in the Clusters/N-grams tool will now match the Concordance tool hits.
All searches are now performed to find the maximum number of hits for a particular search. This
means that a search for "aa" in a string "aaa" will now produce two hits instead of one (as in earlier
versions of AntConc). For most 'word' based searches, very few differences will be noticed. However,
results can be quite strikingly different for string-based searches using wildcards. The new way of doing
searches is intuitively more sensible than the previous search approach. However, many other
software programs adopt the previous approach so you should expect some differences when
Generic line break characters representing line breaks on different systems (e.g. \r\n for Windows and
\n for Linux/OS X) are now used for all file processing.
The Collocates Tool sometimes use to report values of "-1" (when the only collocates of the center
word was itself) and "-2" (when the collocate value was not able to be calculated due to an out-of-sync
word list). The first case ("-1") no longer exists (see feature 3 above). In the second case ("-2"), the
value has now been replaced with "0" to avoid confusion.
In the Word List Tool, Alt-A now correctly selects all the panes including the Lemma list.
When no corpus file names were selected in the left panel, sometimes the ͞Close file͟ menu option
would delete the first file name in the list instead of reporting that no file was selected. The has been
After changing the size of the Search Entry font option in the global settings, the new font size would
not be reflected in the sample display. This is now fixed.
If the "invert sort" option was activated in any tool with this feature, an "Array XXX" error was
produced in the results window. I apologize for this very inconvenient bug.
Several other minor bugs including one that sometimes led to tags not being hidden properly after
applying tag settings have been fixed.
This is a minor upgrade fixing a few small bugs. One very minor new features have been added.
The "Order by Stat" sort option is now set as the default option in the Collocates Tool.
Fixes a problem that caused the window span controller values to become out of sync if the "from"
controller changed the "to" controller or vice versa.
Fixes a problem that caused the window 'same' checkbox controller to not work properly.
Fixes a problem that caused the scrollbars to not show correctly in the cloned view of the
Concordancer Tool window.
Online Remove password from protected PDF file
Online Remove Password from Protected PDF file. Download Free Trial. Remove password from protected PDF file. Find your password-protected PDF and upload it. break a pdf password; password protected pdf
Fixes a problem that caused the background color of the global-settings->character encoding options
to appear white on a Mac OS X system. This made the options impossible to read because the OS X
native foreground color is also set to white, which is different from that on Windows and Linux
This is a minor upgrade fixing a single bug. No new features have been added.
Sort by word end in the Word List tool produced a list in reverse order. This is now fixed.
This is a minor upgrade fixing a couple of bugs. No new features have been added.
A tab delimiter character incorrectly appeared in the KWIC concordance display. This is now fixed.
If multi-character delimiters were inputted, the highlighting became misaligned. This is also fixed.
A bug causing the software to keeping reopening after clicking the close button (top right of screen on
Windows) was fixed. The problem only appeared after importing user settings files or reloading the
This is a semi-major upgrade that addresses an important bug in version 3.3.1. I've also added a few small but
Performance has been greatly improved (> 10 times) for all but the Concordance and Plot tools. I hope
to improve the performance of these tools from now.
It is now possible to append user-defined characters to the standard Unicode character classes.
Users can now choose to ignore or not ignore white space in the search box. The default is set to ignore
Users can now choose which delimiter character to use in the Concordance tool.
The "treat all data as lowercase" option is now set as default on all tools except the Concordance and
Plot tool. I hope this is a much more sensible default setting.
Files created on different platforms (Windows, OS X, Linux) can have different end-of-line characters.
These are now treated correctly regardless of which platform AntConc is launched on.
Some user token definitions caused the output to display strange results that appeared to ignore any
word boundaries. This problem is now fixed.
This is a minor upgrade that addresses a few small bugs version 3.3.0.
All windows now resize correctly to show widgets even when the ͚screen display͛ setting on Microsoft
Windows is set to 125%. In 3.3.0, an enlarged display caused some setting areas to be cut.
Windows are now centered correctly regardless of the screen display͛ setting on Microsoft Windows. Also,
there is no flicker of windows moving from the top left of the screen to the center as they are positioned.
The checkbox for Level 1 sorting now includes a greyed out check mark.
Underline marks that appeared under list items (e.g. corpus files in the main window) have now been
A few bold headers have now been corrected to normal typeface.
Word and case checkbuttons in the main window are now correctly disabled if the Regex option is marked
as active in a user settings file.
This is a major upgrade introducing many new features and addressing bugs that appeared in previous 3.2.x
Now AntConc runs completely natively on all platforms. This is especially important for Macintosh OS X
users as it means no X11 installation is required. Also, the interface is considerably improved over
previous 3.2.x versions most spectacularly on Macintosh OS X but also on Windows and Linux systems.
The ͞range͟ information for a word has now been added to the Clusters Tool. This allows ͚lexical
bundles͛ of a certain minimum range to be generated.
The full path of a corpus file is now shown by hovering the cursor over the file name in the main
Loading of lemma lists is now greatly simplified.
Loading of reference corpus lists are now greatly simplified.
When loading reference corpus lists, checks are now made to see if the format is correct.
The interface has been simplified a little by removing little-used buttons, in particular, the ͞Reset͟
button and ͞Exit͟ button on each tool tab. Also the ambiguous ͞Save Window͟ button has now been
renamed ͞Close Results͟ to more accurately reflect its action.
Hit and frequency counts now appear above the results window in all tools.
The Clusters tool has now been renamed ͞Clusters/N-Grams͛͛ to more accurately reflect its functions.
The actions associated with spinbox arrows have now been reversed. This is perhaps a more natural
way for them to work.
The term ͞Language Encoding͟ used in 3.2.x versions has now been corrected to ͞Character Encoding͟.
Various default settings have been changed to reflect modern corpus linguists͛ preferences. In
particular, the default character encoding has been changed from Latin 1 to UTF-8.
This user file has now been updated to be a little easier to read and understand.
After opening or saving files, AntConc will now remember the folder location.
Selection highlighting of corpus files now accurately reflects which files can be viewed in the File View
AntConc now correctly remembers the positions of window panes after global and tool preferences
have been accessed.
AntConc now correctly remembers which was the last preference window accessed by the user.
Numerous interface oddities associated with 3.2.x versions of AntConc should now be fixed with the
introduction of a native interface for all platforms (see New Features 1 above).
File names with a different character encoding the file contents will now appear correctly.
This is a minor upgrade fixing two problems (one only applicable to OS X). No new features have been added.
Modified the OS X version so that lemma list files will load correctly regardless of the OS version. The
bug was more general and possibly caused errors when using any single file open file dialog box, e.g.,
importing settings files, search list files, and word list files. I only heard one or two reports of problems
on OS X, so perhaps the problem was introduced with upgrades to recent versions of OS X. It is difficult
Fixed a bug that caused the total number of lemma entries to not be displayed properly.
This is a major upgrade introducing several new features and addressing bugs that appeared in version 188.8.131.52.
Massively simplified the engine used for sorting and displaying results. This will allow for improved
Introduced a Concordance preference setting allowing tab spaces to be added around the search hit in
the KWIC concordance view. Introducing this means that the toggle shortcut key ͚x͛ to show/hide the
KWIC search term has had to be removed. This function can now be accessed via the (fixed) menu
option. I also hope to reintroduce this toggle shortcut key later.
Introduced the option to use word list(s) of reference corpus files instead of directly processing the raw
files. This allows fast generation of keywords and also allows users to generate keyword lists even
when the reference corpus is not available (but a word list is). See the Keyword List section for further
Introduced feature to output the counts of types and tokens for each tool when saving results as a text
Changed some of the default settings in response to user feedback:
Concordance Tool sort level s are set as 1R, 2R, 3R.
The Collocates Tool Stats measure is now displayed and calculated.
The splash screen that appeared when a user settings file was added has been removed. Now, the
words ͞User Settings͟ appears next to the title of the software at the top of main window, whenever a
user settings file is used.
The layout of the main window has been improved so that the main work area is maximized for all sizes
of main window.
Unicode ͞Marks͟ can now be added/removed in the token definition settings
Occasionally, the last bar of the progress bar remained blank at the end of processing. This has now
Under tag settings, the ͞Hide Header Tags͟ option will now work for headers than span across multiple
The main window and all pop-up windows now appear centered in the display.
When viewing the global settings preferences, it was possible to accidentally activate the main window
by clicking on it. This has been fixed.
The list of files in the left pane was often too narrow to view easily. This has been fixed by introducing a
two-pane approach, with the left pane showing the files, and the right pane showing the results and
controls. The width of both panes can be adjusted with the default size being the minimum size.
It was possible to adjust the main window to be smaller than the default size of 800px by 600px. This
caused the positioning of some widgets to become misaligned. Now, 800px by 600px has been set as
the minimum size of the main window.
Various problems experienced when cutting, copying, and pasting text to/from the Search boxes in
each tool have been addressed by removing the ͞Spinbox͟ widget type and replacing this with a
standard ͞Entry͟ widget. The history of previous searches can be now accessed using the ͞Up͟
(previous) and ͞Down͟ (next) arrow keys.
After moving the cursor over the corpus files list while using the File View Tool, it should change into a
͞Pointing Finger͟ cursor. However, this cursor shape was returned incorrectly back to a standard ͞Edit͟
cursor when one of the files was selected. This has now been fixed.
Changed foreground to white when AntConc reports that the KWIC analysis has finished or has no
results. This should be easier to read.
Fixed a bug that caused the Concordance preference option ͞Hide search term in KWIC display͟ to only
show the hit. (The opposite of the stated function!). This is now fixed.
Fixed some typos in this Read Me document.
Fixed a major bug that caused a warning to always appear when using the T-Score statistical measure
with the Collocates Tool. Note that the warning could be ignored as the results would always be correct.
This is a minor upgrade addressing bugs that appeared in version 3.2.2.
The legacy character encodings for Japanese, Chinese, Taiwanese, and Korean were inadvertently left
out of the compiled executable for version 3.2.2 meaning that files encoded in these languages would
not open correctly in AntConc. These have now been added.
In the Mac OS X version, the compiled version did not include several important components which are
essential for the X11 graphical toolkit to run correctly. These have been added.
This is a minor upgrade addressing several minor bugs that appeared in version 3.2.1. It is also the first version
of AntConc to be complied with Perl 5.10.
Fixed a bug that caused back references to not work correctly when the 'Regex' search option was
selected. This now works correctly. However, note that the first back reference should be 2 and not 1
(due to an implicit back reference 1 that holds the entire search result.
Fixed error message when the sort by Stat option was selected.
This is a minor upgrade addressing several bugs that appeared in version 3.2.0, as well as introducing a few
new features requested by users.
Better display of long lists of fonts and font sizes in the global options/font menu. Now the lists appear
as an easy to navigate list with attached scrollbar.
When results windows are saved, the cloned windows now display summary results information.
Word range lists can now be used as lemma range lists.
New feature allowing tagged data to be searched while remaining hidden. See the tag settings
preferences. Pressing CONTROL and the START button (or the ENTER button if the search entry box has
the focus) temporarily disables the new feature, allowing the user to switch easily between a 'non-
tagged' or 'tagged' display.
New options in the tag settings preferences that allow embedded and non-embedded tags to be shown,
ignored, or hidden. This enables data of the form of the BROWN and BNC corpora to be processed
Improved the updating of the progress bar display, which may also improve the speed of processing in
Improved the images used for icons within the program.
Fixed bug that caused user defined token definitions containing special regular expression characters
from not working properly
Fixed bug that caused "Treat all data as lowercase" option to ignore the wordlist range and lemma lists
Fixed bug that caused the lemma list "Load" button to not ignore the currently opened file if a new file
dialog was opened and then "Cancel" was pressed
Fixed bug that caused file searches to not work if the search entry box was blank.
This is a major upgrade with a completely redesigned interface, several new features, and several bug fixes.
The new interface follows the basic design used in previous versions, although users should find it 'cleaner' and
more intuitive. In particular, all global and tool menu settings have been combined into two groups, where all
the related settings can be accessed and adjusted within the same window. This should dramatically improve
the usability. All tools now have access to the search engine (including the Word List Tool and Keyword List
Tool) and there is also a new advanced search window that can be used to perform list (file) searches, and
searches within a particular context. Due to the nature of the changes, this version will not be compatible with
the settings files for previous versions. Another huge change is that this version will run on Macintosh OS X
Completely redesigned interface
Added search and advanced search features to all tools (including the Word List Tool and Keyword
Created new list (file) search available in all tools.
Created new context search option in all tools except the Word List Tool and Keyword List Tool where
it has no meaning.
Busy cursors are used to indicate when very long sorting operations are being carried out (e.g. when
sorting large N-gram list results).
Case options affecting whether or not data is converted to lower case are now more intuitive. For
example, the 'Case' option in the main window now only affects the operation of the search itself and
has no impact on the data under observation. Data can be treated as lowercase (for example in Word
list tool) by choosing the 'Treat all data as lowercase' under the appropriate category in the 'Tool
The number of corpus files (and reference corpus files) being analyzed is now displayed.
Correct some mistakes in this readme file
My name has been removed from the top of the main window! However, please remember that my
name is Laurence with a U if you are ever citing me in your research papers!!
Now works with Macintosh OSX
The program no longer crashes when the 'All Values' option is chosen as the threshold value in
Negative keywords are now highlighted correctly when the 'Show Negative Keywords' option is
The KWIC lines are now aligned correctly even when the hit appears near the very start or end of a file.
Collocates frequency values are now correctly calculated even when the span extends further left than
the start of the file
The action of the 'one word only' wildcard is now more intuitive.
Some operations (e.g. creating a word list) now do not crash after restoring the default settings and
then performing an operation.
Bug fixes (since beta1 version):
The program now (correctly) only shows files that generate hits in the Concordance Plot tool.
The sort function in the Keywords Tool now works correctly. In previous versions, even when the
'Frequency' option was selected, the sort would be based on Keyness. Also, in some cases inverted
sorting did not work.
Bug fixes (since beta2 version):
The program now (correctly) hides the various Concordance Tool panes depending on the chosen
Display Options. In the earlier beta versions, the options were ignored.
The default file type to use when opening directories now works correctly. In previous beta versions,
after hitting the apply button, the default file type reverted back to the .txt type.
Fixed a bug that prevented the n-grams option in the Clusters Tool from working when the search
term entry box was empty.
Bug fixes (since beta3 version):
Fixed bug that caused the program to not be able to open files with non-English names correctly if the
full-pathname option was selected. There are potentially many problems with non-English filenames,
so I recommend that users use English filenames for their corpus files, and also save them under a
pathname which only contains English characters.
Fixed bug that caused the 'OR' wildcard to not work correctly if a character other that '|' was user
New Features (since beta4 version):
Made some small changes so that the program could be more easily ported to Macintosh and Linux
This is a very minor upgrade with just one change:
Corrected problem that caused the No. of Hits to not be indicated correctly in the Concordance Plot
Tool display when more than one corpus file was being used.
This is a very minor upgrade with the following changes:
Corrected problem which caused the program to not launch when the path of the default temporary
folder on the system contained non-English characters.
(Linux only): Corrected problem that caused the Open Dir menu option to not work correctly.
(Linux only): Corrected problem that caused font selections to not work correctly.
New Features: Improved speed and memory handling when calculating collocates. Over 10 times faster than in
previous versions (including version 3.1.3).
This is a minor upgrade containing an important bug fix that prevented files with non-ASCII filenames being
used. There are also some major performance improvements. For example, n-grams will now be processed
over 10 times faster on small corpora and many more times faster on larger corpora. A list of all important
changes is below:
The history feature for search term entries has been changed. I have heard two reports of the 3.1.2 version not
starting on computers. Hopefully, this change will allow the program to start correctly on all machines.
The performance of tools such as Collocates, Clusters and N-grams, has been significantly improved. (Over 10
times faster on small corpora and many more times faster on larger corpora.)
The Open Dir option now open files in all sub-directories too.
The program will automatically look for a user defined settings file named "antconc_settings.ant" in the
directory where the program is saved. If this file is found, this settings file will be used instead of the default
Documents you may be interested
Documents you may be interested