============================================== Version 2.8.7: 30 Nov 2011 ============================================== Bug fix: where CT is installed on a machine where the user application data directory is inaccessable, CT would crash on the second startup and henceforth. Now fixed. ============================================== Version 2.8.6: 23 october 2011 ============================================== BUG FIX: * When a file was recorded with wrong encoding (e.g. as utf-8 when it was iso-8859-1), then CorpusTool would fail in the opening of the project, making it impossible to fix the problem. Now fixed, and reported properly in the appropriate context within the tool (thanks Margot). ============================================== Version 2.8.5a: 29 September 2011 ============================================== BUG FIX: * In sentence segmentation, initials in names, such as "Donald L. Smith" were treated as sentence endings, now fixed. ============================================== Version 2.8.5: 29 September 2011 ============================================== STATISTICS PANE: changed the way in which the Barchart option displays. Now, when you switch to Barchart (clicking on the barchart icon): a) For General stats: a barchart is shown for each of the general aspect, with one bar for each of the sets indicated by the user. b) For feature Stats, a barchart is shown for each of the features of the selected unit, with one bar for each of the sets indicated by the user. Note that now Descriptive Stats, which involves a single set, will no longer offer the option for Barchart view. SCHEME: Fixed bug in save PDF from the Scheme window. (Thanks Kerrill). Note that Foxit PDF Viewer will know to print the single huge page as multiple single pages, which can then be taped together. (Not sure about equivalents on the Mac). PARSING: For the English parser, there was a bug which stopped documents being parsed in some contexts. This has now ben fixed. ============================================== Version 2.8.4: 8 September 2011 ============================================== Minor bug fixes over 2.8.3. STATS: Saving Descriptive Statistics as tab delimited text: the list of column titles (filenames, or set number) was misaligned. Now fixed. Also, save feature stats as text now provides the count as well as the percentage. MERGE PROJECTS * Previous implementation of this function (accessed from the Project menu in the menubar) did not merge all files. It now should. You will be asked to select another project to merge with the curent one. In all cases where the same files exist in both projects (e.g., schemes, corpus files and annotation files), the file in the current project will be preserved. Files will moved from the other project where they do not exist in the current project. To ensure cleaner merger, delete any layers in either project which have no annotation associated with them. Results will be saved in a new folder with the same name as the currrent project, plus "-merged" added. ============================================== Version 2.8.3: 31 July 2011 ============================================== MAINLY FIXES BUGS IN 2.8.2. However, some new features Various bugs detected during the CorpusTool Workshop at the Institute before ISFC 2011. Thanks to Ryan, Christina, Ann, Filipa, Adegboye, Gillian, Bart, Libo, Kate, Victor, Leyla for helping me find these bugs. MACOSX NOTE: version 2.8.2 did not display non-western scripts appropriately on Mac due to problems beyond my control. Version 2.8.3 fixes this, but does not allow you to to launch CorpusTool by double-clicking on the .ctpr file in your project folder. Open from within the application. ------------ NEW FEATURES ------------ STATISTICS Pane * there is now an option to "Describe several datasets", which allows you to see results for more than two subcorpora at a time (e.g., if you have 5 text-types defined by a document layer, you can see one column for each of them) (for Mario, long overdue). AUTOMATIC ENGLISH PARSE: Each finite clause is now assigned a feature for the tense-aspect combo, e.g., simple-present or past-perfect. This system ("TENSE-ASPECT") can be used in the Explore interface, "Feature Counts" and the "textstream" display to see how tense combos change through the text. If you want these features added to files that are already parsed, you can open each file for annotation, and select "Resegment Document" from the "Other Action" menu. ------------ BUG FIXES ------------ PROJECT WINDOW:EXTEND CORPUS: when adding a new file of non-European scripts, all segments made in the first session were indented one character when opened in subsequent sessions. Now fixed (Thaks Anya). CODING WINDOW: Glosses did not display properly after the first one. Now fixed. SCHEME: * Removed "Add Gate" option from the feature menu. (a "gate" is a system with one feature, but this is rarely used, and simple to do by adding a normal system and deleting one feature). AUTOCODE * In some cases, when an autocode rule refers to a feature created in the same session, the rule would refuse to work when "Show" is pressed. Now fixed. * Autocode rules are no longer stored in the scheme, but in a separate file, which makes the scheme diagrams neater. * Where an autocode rule included a "in_role" specification, this was not saved. Now yes. ============================================== Version 2.8.2: 15 July 2011 ============================================== BUG FIX RELEASE: previous version broken when creating new projects, but fine with existing projects. Fixed the bug. Thanks Mafalda. ============================================== Version 2.8.1: 14 July 2011 ============================================== EXPLORE: Lots of revision to the Explore pane, which offers various ways to visualise patterns in a text, a subcorpus, or your project as a whole compared to another project). Open the pane and click the Help button for more information. Eliminated the "KeyWords" pane as all of this functionality is now available in the Explore pane. STATS: Added an option "per 1000 words", which will provide number of occurences of features per 1000 words. Select this from the menu with "local" and "global". ============================================== Version 2.8 beta 0: 1 July 2011 ============================================== NEW FEATURES EXPLORE PANE: Major addition of the "Explore" pane, which provides various ways to visualise patterns over a text. STATISTICS PANE: Added a "Subjectivity" measure to the General Statistics section of the Stats pane. This basically shows how prevalent subjective words are in a text, and how positive or negative on balance (English only as it depends on a word list I only have for English). SEARCH PANE: Minor revision of the Search window: * presentation in "Summary" format now NOT limited to 15 most frequent, shows all (in order of frequency). Also, words are grouped ignoring case (so 'maybe' and 'Maybe' are counted the same). * "Save as HTML" improved. If you Save the search when "Table" view selected, the search is saved one row per hit. If the search involved some form of "containing", then the table will be shown with contained segents in their own columns (previously only the whole containing string was shown). * Where a search string contains a wordclass feature which is not defined, the system used to tell you this for each file. Now the message is given once. AUTOCODE PANE: Added an option to "create segments" in a layer corresponding to string patterns within the document as a whole. Also fixed a memory leak problem with autocode. Also fixed display of the check boxes which tended to cross into the next row. TOKENISATION: Minor revision of the way English texts are tokenised, so that "won't", "isn't", etc. are treated as a single word, while "I've", "cat's", "I'm", etc. are treated as two (with the ' included in the second token). MENUS * Added an option "Show Project Folder" to the Project menu, which will open your project folder in the Finder/Explorer. SCHEME DISPLAY * Fixed a bug in display of glosses from recent versions, which were shown without splitting long lines. * You can now place URLS in glosses, e.g., To see more, press here. Be careful not to make syntax mistakes with the url structure. The text within the and will be shown in blue, and you can click on the blue text to go to the link. To open a file in your project folder, try: To see more, press here. ...where test.txt is a file in your project folder. Glosses are also hyperlinked in the Gloss box of the Coding window. (Thanks Jim Benson for pressing for this over years) ============================================== Version 2.7.4: 13 April 2011 ============================================== NEW FEATURES KEY FEATURES: In addition to presentation of Keywords, the Keywords pane now offers an option for "Key Features", which shows the annotated features in a subcopora which are special to it (e.g., which errors do low proficiency students make in comparison to higher proficiency students) SEARCH: As before, the magnifying glass icon can be pressed to view the segment in context. Now, the previously viewed segment is indicated by putting a border on its magnifying glass icon. ========= BUG FIXES ========= START NEW PROJECT * May have been a problem with the 2.7.3 release where the main window was drawn blank on startup. Now fixed (hopefully) SEARCH * Fixed bug in search for segment containing segment, would return segments not matching the feature spec (e.g., "np containing determiner-error" would return elements which were not np as well) * Fixed a display problem in Search where the search included a "containing" specification and the searched segment contained a line break. * Fixed the handling of "containing immediately" across layers. "np containing immediately 'the'" previously would not match any np, since 'the' is immediately contained within a subelement of the np. Now, the smallest np containing 'the' will be returned. * Cancel of a search previously would not actually stop the search due to a bug (although the interface did show the search as cancelled, the program kept on running). Now fixed. * In the Search window, the menu at the right allows display in terms of "by file", "table", "summary" or "lematised summary". Previously, "lematised summary" only worked for a query with a 'containing_string' field specified. Also the contained items, when specified with @ (e.g., @noun) would be lematised using the dictionary (English only). This has now been extended for nouns, verbs and adjectives in an English Grammar layer (the one produced by the automatic parser) * Fixed bug in search for item containing a string where the search string was a single word, and the segment's string was that word plus one char (e.g., 'will' would match 'wills'). AUTOMATIC GRAMMAR ANALYSIS * Cases of verb and negation conflated (isn't etc.) are now treated as a single item, rather than two. ============================================== Version 2.7.3: 4 April 2011 ============================================== BUG FIXES * Bug in the File list display area of the Project Window did not allow scrolling beyong 1200 pixels wide, due to a bug. Now fixed (Thanks Gergana) * Fixed bug in "Import Metadata" where the spreadsheet had spaces in the title of a column. ============================================== Version 2.7.2: 4 March 2011 ============================================== BUG FIXES * Incorporating files: for the last few versions, the program always presented incorporated text files as UTF-8, even if the file could not be displayed in that encoding. Now, the system should revert to earlier behaviour, making a good guess at the encoding. * Autocode window: trying to add a rule without any layers in the project gave no response. Now, a message advises that at least one layer should be in the project. ============================================== Version 2.7.1: 27 February 2011 ============================================== NEW FEATURES * You can now put hyperlinks in glosses of features. See the manual (v2.7) contained in the release for details. (Requested by Jim) BUG FIX * The opening window cannot be maximised. (not really a bug, but...) * Keyword search will now work when there are only 1or 2 files in a project.(previously, each keyword had to be in more than 1 file to qualify, to stop words ideosyncratic to one text from appearing in the results). ============================================== Version 2.7 beta 4: 28 January 2011 ============================================== NOTE: IF THIS VERSION INSTALLS BUT DOESN'T RUN ON A WINDOWS MACHINE. YOU WILL NEED TO INSTALL SOFTWARE FROM MICROSOFT AVAILABLE FROM http://www.wagsoft.com/CorpusTool/vcredist_x86.exe ADDED * On the Keywords pane, you can now find key words by comparing a subset of this corpus with either: i) the rest of this corpus (as before); ii) another named subset of this corpus, or iii) the corpus held in a different CorpusTool project. * Added option under the Project menu to "Save Standalone Project". This saves all (and only) the essential files of your project in a single file, which can be emailed to the maintainer if you need the project debugged. (this is actually a zip file, but the extension is changed to .ctpa). BUG FIXES * Fixed bug in the Keywords pane where displaying keywords and you don't have a layer called "document". * Keyword lists for languages other than English should work a little better now with abbreviations (if abbreviation data is provided in the lexicon of that language). * For an "Error" layer, the option to "Save Student Feedback" was broken for a few versions. Now fixed. * Fixed a bug in the production of the automatic English Grammar layer where segments assigned "progressive-aspect" were also assigned "not-progressive-aspect" (and similar with other features) ============================================== Version 2.7 beta 3: 18 January 2011 ============================================== * Fixed further problems with Autocode pane when editing a rule. * A layer created to segment into paragraphs would fail to recognise paragraph boundaries on some platforms for some documents. Fixed. Version 2.7 beta 2: 8 January 2011 ============================================== CHANGES * Added RST layers, which allow the editing of the rhetorical structure of text (basically the facilities of RSTTool, but within CorpusTool, one can also assign features to the text segments, and thus study the kinds of units which fill each relation. * Added importing of RS3 files from RSTTool. * Import of Systemic Coder files put back after a vacation of several versions. * Text Streams (from within the Actions menu in each Annotation window, have been beautified. * Concordance Summary: Added display option on the search window: "Summary", which shows the number of hits for each search combination (e.g., searching in a small corpus for "sentence containg_string "@verbal-verb @personal-pronoun that" gave the following result: told me that 3 told them that 2 warned him that 1 inform you that 1 Also an option "lematised summary", which lematises any element defined using @ or %, so that e.g., results for 'said' and 'say' will be merged. BUG FIXES * Fixed once again Autocode which did not work in some situations * Export Layer was not working in recent versions. Now fixed. (thanks Lee Chang Soo) * Fixed bug in display of search results where segment contains string or segment, and total segment length > 140 chars. ============================================== Version 2.69 Release: 4 December 2010 ============================================== Minor Improvements * When Incorporating files, you no longer have to press the Start button to start incorporation (which was a bit redundant). * Improved HTML saving of Statistical comparison of feature data. * Improved reporting of what is happening during the parsing of texts by the stanford parser. * Sped up display of search where search is of whole documents: only first 140 chars of each document displayed in Search and Autocode windows. Bug fixes * Clicking on the name of a file in the main window would not display the file due to a bug. Now fixed. * Lexical density reports now works for Portuguese. ============================================== Version 2.68 Release 4 November 2010 ============================================== Bug Fix release: * Fixed a bug which stopped annotation windows opening for some files (Thanks Susana) ============================================== Version 2.67 Release 1 November 2010 ============================================== BUG FIXES * Coding Window: When opening the annotation window for a "code-document" layer, the text wass shown scrolled to the end of the text by mistake. Now shows the start of the text. * The Windows installer was made using Python 2.7, which depends on some Microsoft files being on your machine. Microsoft does not distribute these files in all releases, so CorpusTool might fail to launch after installing. Now recompiled with an older version of Python. * Saving Search results was broken in the last few versions. Now fixed (Thanks Kesuma) ============================================== Version 2.66 Release 22 September 2010 ============================================== Minor revision: Layers defined as "code-document" are automatically provided with two fields: "subcorp" and "filename". Each document at that layer is provided with these fields, containing the name of the document's folder in the "suborp" field, and of the filename in the "filename" field. These are useful, as they can be used in searches (Search, Autocode and Statistics windows). For instance, one can find all documents in the "Narrative" folder by specifying a search: document with_field subcorp=Narrative Note also that search on fields can use the wild-card "*". For instance, one can search for: document with_field filename=Ca* ...will find all documents whose filename starts with "Ca". ============================================== Version 2.65 Release 22 September 2010 ============================================== FUNCTIONALITY IMPROVEMENTS AUTOCODE * Added option for Creating Segments in one layer corresponding to search hits on another layer. For instance, if you have a layer with all NPs segmented, you could create segments on a separate later for each NP containing a plural noun. There will be no NPs created on the new layer for NPS without a plural noun. SEARCH UNITS IN ROLE In the Unit specification (in the Search, Autocode or Statistics windows), on cane specify to search for units filling a particular Role. This only works for layers that include role structure (e.g., STNFORD layer and the Automatic Grammar analysis). E.g., search for "np filling_role subject" ============================================== Version 2.64 Release 16 September 2010 ============================================== FUNCTIONALITY CHANGES CREATE NEW PROJECT: a) temporally removed the option to start Image projects, a separate tool will be used to crete and annotate image corpora. b) removed the option to specify files for the project during the "create new project" phase. This should be done on the main Project window (it may have been confusing to have two ways to do the same thing). IMPORT: temporarily not working, pending a total revision of all import. Use prior version if you need to import Systemic Coding files or XML. BUG FIXES EXTEND CORPUS: minor improvements here in the case where you corpus has no existing sub corpus folders. INCORPORATE FILES: a) Previously, the Start button could be pressed multiple times, which confused the program, incorporating the same files multiple times. Now, this button is disabled when it i pressed. b) The "Cancel" button did not work, now fixed. SEARCH: search for segments in large corpora would sometime hang. This has ben fixed, but the interface may not be as responsive while a search is in progress. AUTOCODE: Autocode stopped working in 2.6.1 or so, now fixed (it seemed to work, but the Code Selected button did not change the features of the selected segments). ============================================== Version 2.63 Release 22 July 2010 ============================================== BUG FIX * Previously, when "Incorporating All", UAMCT would use the encoding of the first file on all subsequent documents. Now, the software checks if the encoding is appropriate for that document, and if not, prompts the user to specify the encoding. ============================================== Version 2.62 Release 15 July 2010 ============================================== Fixed bugs in TextStreaming (called from the Other Actions menu in an annotation window. Seems to have stopped working, but now fixed. Also, where a segment doesn't select for the system of interest, the segment is not included in the textstream. This provides cleaner drawings. ============================================== Version 2.61 Beta Release 8 July 2010 ============================================== Aded Keyword extraction (words most associated to a subcorpus) and Phrases (word sequences most associated to a subcorpus, following Biber's Lexical Bundles) Se the Keyword tab on the main screen. ============================================== Version 2.6 Beta Release 27 June 2010 ============================================== * Fixed again the way the Stanford parser is downloaded and stored locally. Will maybe fix some problems. * Fixed way UAMCT finds a folder to store local data (preferences etc.). This gives problems on VERY protected systems. * The Automatic Clause analysis layer was revised. Should be a little more accurate. * Sentences of up to 55 words are now parsed (previous limit of 40). * Sentence segmentation with abbreviations such as "etc.", which usually terminate a sentence, was broken. Now fixed. ============================================== Version 2.5 Beta Release 23 February 2010 ============================================== CODING * If you set the option autocodeRepeatedText in the Options pane, then if, during the coding of a text, you select identical text to that already coded in this text, the same features will be assigned to this segment. If the guess is wrong, it is simple to recode the segment. very useful when error coding and the same errors re-occur. * When coding a segment of Error type, a new button "VC" will appear on the coding interface. Pressing this will display the text around the currently selected segment, with all errors replaced by their corrected text. Fixed bug with case matching in searches: if all of your search string is lower case, then searching is not case sensitive: 'eu' will match 'eu', 'Eu' and 'EU'. If your search string uses upper case at all, then matching IS case sensitive: 'Eu' will match ONLY 'Eu'. EXTENDED LANGUAGE SUPPORT * Now provides lexical density information for 17 languages (European, some Slavic and Turkish). Pronoun usage information for German, French and Spanish added (affects "General text statistics" in the Statistics pane, and "View Basic Text Statistics" under the Action menu in the Project Window). SEGMENTATION * Added option when creating a layer to segment into paragraphs or into words. Will probably only work for western languages, and possibly other languages depending on the writing system. IMPORT XML FILES CODER: It is now possible to leave on system uncoded but continue with other enterable systems (click on the "Gloss" and switch to "Other Systems") FIELDS: * You can now add string or number information to each segment. Ask me (michael.odonnell@uam.es) if you want instructions. PROJECT WINDOW: * the information tab for each layer now displays the type of segmentation selected for the layer. BUG FIXES * Fixed a few problems with saving of preferences under Windows 7 and Vista. * Fixed a few problems with the saving of the Stanford Parser under windows. * The English Grammar Analysis layer did not work due to a missing file in the previous releases. Should now work. ============================================== Version 2.4.2 Non-beta Release 3 January 2010 ============================================== NEW FEATURES * CREATE NEW PROJECT: on MacOSX, the save location now defaults to the Desktop (already does this on Windows). * SEARCH: Previously, when searching for segment CONTAINING segment, and a segment contained several of the specified segments, each hit would be included in the count. Now a segment is only counted once no matter how many of the specified segments it contains. (Thanks Cameron). Only the first hit is displayed with a magnifying glass (click on the icon to open the segment in context). BUG FIXES * CREATE NEW LAYER: layer names are now restricted to alphanumerical charachers plus ., - and _. If you use other chars, they will be replaced by _ (Thanks Eun) Also you cannot now use the name "Metadata" for a layer name (it is reserved for system use) (thanks Kilim) * ANNOTATION: fixed problem in display of codings as structure, where root feature of the scheme was shown instead of most delicate feature. * STANFORD PARSER: on some systems, the automatic download and install of the Stanford parser would break. Fixed. * SEARCH: All segments were wrongly displayed as having comments. This has been fixed (Thanks Cameron). * AUTOCODE: a bug stopped the searches from displaying matches to the autocode rule. Fixed. Additionally, If a scheme was saved with an invalid autocode rule, this could stop the loading of the project as a whole. Now, the invalid rule is just dropped. ==================================== Version 2.4.1 Non-beta Release 13 December 2009 ==================================== BUG FIX RELEASE * When incorporating files, the tool would always ask you to locate the Stanford parser, even if your project does not require files to be parsed. Now fixed. * In the 2.4 release, on incorporating a file, you were no longer asked to specify its metadata first. Fixed. * Turned off the Automatic Grammar analysis of German texts (it only works for English). * SEARCH: If searching for a segment CONTAINING other segment, the view mode selector showed the current view as "by file" while it was actually in the "table" format. * When adding a new layer, the layer name 'Metadata' is not allowed, as this layer name is used internally by the system (thanks Kilim Nam). ==================================== Version 2.4 Non-beta Release 10 December 2009 ==================================== NEW FEATURES * Major overhall of the automatic recognition of grammatical stucture (Subj/Pred/Obj, etc.) Only for English. To see this functionality, add a new layer and select "Automatic Systemic Analysis". Then open an annotation window for this layer to see your file analysed. * Added in an Options pane, which wil allow various options in the tool to be changed. Currently allows changing of: - autoadvance: when all features on a segment are coded, move automatically to the next segment. - Skip complete codings: if autoadvance is set, autoadvance to next coding which is not complete. - Generate Missing Files: when searching or performing statistical analyses, if an automatically generated file (e.g., using the parser) has not yet been generated, skip it. * STATS: General Text Stats now shows the min and max length of segments (in terms of tokens). BUG FIXES * IMPORT LAYER: Fixed bugs in import project from coder files * SEARCH: Fixed display problem in Search results, with small columns expanded too much * SEARCH&STATS: Search of layers with more than 128 features would fail due to the speedups introduced in version 2.1. Now fixed. * AUTOCODE: when editing a search query, and changing the feature to select, the displayed condition should change to the same layer as the select feature, but didn't. Now fixed (Thanks Joshua). * PARSING: Fixed some problems with parsing German. Some sentences were not displayed in the annotation window. May still be problems here related to the Stanford Parser's handling of umlauts etc. ==================================== Version 2.3 beta 6 31 July 2009 ==================================== * When changing a scheme, the main window would come up over the scheme window (annoying). Fixed. ==================================== Version 2.3 beta 5 29 July 2009 ==================================== * UAM CT would fail to launch on some Windows machines. Due to change to version 2.6 of the programming language used (Python). I have reverted to Python version 2.5, which should fix the problem. * The previous 2.3 releases did not work on some PPC macs. This version MIGHT fix the problem. Let me know if you have problems. * Fixed possible problem installing Stanford Parser on Windows machines, introduced by bug fix in 2.3 beta 4. * Added progress bar during parsing of files by Stanford parser to indicate that something is happening. ==================================== Version 2.3 beta 4 18 July 2009 ==================================== * Added feature: global/local counting in statistics (in local counting, within a system, percents add up to 100%, within global counting, percents are in respect to the total number of units in the set). BUG Fixes * Fixed bug in installing Stanford Parser for the first time on Windows 7.0 Machines * In searching for item containing immediately some segment, some matches were missed. Fixed. ==================================== Version 2.3 beta 3 13 July 2009 ==================================== Substantial revision to Statistics interface. Improved parsing of English texts. Added Lexie Don/Peter White Appraisal Scheme to the set of schemes included. ==================================== Version 2.1 beta 3 22 January 2009 ==================================== The Macintosh version now uses new font display code which should properly display Unicode text. Hopefully this means that Chinese, etc. should display properly on the Mac (let me know if you have problems). NOTE that Image annotation is not yet fully fixed in this version, in particular Search and Stats. Minor bug fixes on 2.1 beta 2: * Save PDF in the Scheme window would fail if the scheme had non-ascii characters. Should be fixed now (Thanks Michele). * Save SVG on the Scheme window had a bug, now fixed. * Export Codings in a Coding window failed. Now fixed (thanks Mario). * The "Extend Corpus" option for image projects was not functional, now fixed (Thanks Ivan). ==================================== Version 2.1 beta 2 22 January 2009 ==================================== Major revision of many parts, especially the search and autocode. NEW FEATURES * Automatic segmentation into clause and NP for English. When creating a layer, specify "Segment into clauses" or "Segment into NPs". * Error Analysis Layers: * Save Student Feedback Search: * Search can be cancelled with the "cancel' button. * Much faster search and autocode, especially where you have lots of text files. * You can now show the file names even with KWIK table presentation. * New symbols for pattern search: you can now use $ in a search pattern to indicate the pattern must match against the end of the segment. E.g., '@noun $' will find segments ending with a noun. BUG FIXES * When a text file starts with a BOM code (sometimes present to indicate the character encoding used in the file), the system might misalign characters. Now fixed (I hope). (Thanks Jannika) ==================================== Version 2.0 beta 7 25 November 2008 ==================================== * Fixed bug in drawing of schemes where a system has multiple entry conditions (Thanks Bill) ==================================== Version 2.0 beta 6 15 October 2008 ==================================== BUG FIXES * loading old projects would not allow some of the documents to open. Fixed in this version. (Thanks Susana) * Incorporating a text would break where one of the layers specified segmentation as "clause". Now fixed (Thanks Cedrick). * Importing of Coder .cd3 files did not work. Probably fixed now. * Fixed mouse scrolling within annotation windows in Mac version (Thanks Monica). ==================================== Version 2.0 beta 30 September 2008 ==================================== APPEARANCE Version 2 of UAM CorpysTool represents a radical re-design of the interface to make it more user-friendly. In this version, the Project Window, Statistics, Search, etc. have been placed in a single window as tabbed panes, so it is easier to find the window you want. The appearance has been redesigned to be more pleasing to the eye: this includes color scheme, button appearance and the appearance of labels. Online Help has been extended, with each window/pane having a Help button to give guidance on using that pane. IMAGE ANNOTATION CorpusTool 2.0 has been extended to allow you to annotate images (jpg, gif, etc.) as well as text. When creating a project, you are given the option to make a text or image project. Image projects allow you to add images to the project, and annotate each image (drag a rectangle on the image and assign features to this segment). For instance, you can select all eyes in your image corpus, and code each pair of eyes as +contact or -contact. Later, searching for all eyes with +contact will show you those instances. Perform statistical studies just as with text projects (e.g. is eye contact more or less common in kids books than in women's magazines). STATISTICS * The statistics interface has also been re-designed to make it easier to use. * For contrastive studies (comparing two subsets of your corpora), the software now provides chi-squared statistics as well as T-statistics. METADATA * When selecting the text encoding of a file, the text is displayed as read with that encoding. This allows you to ensure that the encoding you select is the correct one for that file. TEXT STREAMING * There is a new option under the Other Actions menu of the annotation windows, which allows you to see visually how the selection of features in this text changes as the text progresses. See Text Streaming in the manual. BUG FIXES --------- * Apple-q now quits the application on MacOSX. * Search and Autocode: Fixed a bug which would make search freeze when 2 segments were exactly equal. * Autocode: Fixed a bug such that segments autocoded were not always saved permanently (appeared as coded for this session, but when opening again, were not coded). * Fixed a bug such that re-sizing of the Project window was sometimes not possible. * ENCODINGS: Fixed bug where some accented chars in utf-8 would not display correctly (but no problems with annotation). (thanks Cedrick) * TEXT CODER: Fixed bug such that a line of text with no spaces in it would display as lots of 2 character lines. * Import Coder Files: would break where realisations held characters which give problem for XML, such as <, > and &. Fixed. ...and many others which I have lost track of. Version 1.3.2 10 May 2008 ==================================== Bug Fix: AUTOCODE * The Autocode interface can be used to tag all segments matching a given pattern with a nominated feature. A bug in the code meant that features assigned to a segment might not get saved to disk, such that when re-starting CorpusTool, those tags were lost. Now fixed. SEARCH and AUTOCODE * If you have the same segment underlined twice, starting at the same point and ending at the same point, then searching for segments containing strings or segments may lead to a freeze. Now fixed. ==================================== Version 1.3.1 27 April 2008 ==================================== Minor Bug fix on 1.30b: SCHEME * Previous versions gave problems if you included a "&" in a feature name. Now fixed. (Thanks Rudi). * On Mac, if you used the "Change Entry Condition" option, the OK button did not work in come cases. Fixed. METADATA WINDOW * On Mac, when using the MetaData window (to set language, encoding or font of a text file), the OK button did not work in come cases. Fixed. (Thanks Cédrick) STATISTICS and FILE STATS * Changed the way words are counted so that a word must contain an alphanumeric character (thus excluding punctuation and numbers). (Thanks Wenting) SEGMENTATION * A minor but meant that when segmenting into sentences, the last text segment would not be identified as a sentence. ==================================== Version 1.30b beta 24 January 2008 ==================================== Minor bug fix: * Import Coder Study: fixed a bug which stopped the import of Systemic Coder studies as a new layer. ==================================== Version 1.30a beta 24 January 2008 ==================================== Minor bug fix. SEARCH/Autocode: * search for segment IN segment failed due to a bug introduced in recent changes (thanks Thora). * Save of Search results failed. Fixed. ==================================== Version 1.30 beta 23 January 2008 ==================================== Major revision of Corpus Search, addition of Autocode window (assign a feature to a set of segments matching a search query). SEARCH * Major revision to the Corpus Search window. - Now can use ^ in a search pattern to indicate the pattern must match from the beginning of the segment. - Added a keyword "containing immediately" in search queries which means that where segments are embedded within others, only the lowest level segment containing the search pattern will match (e.g., "clauses containing '@modal-verb'" would only match the embedded clause in [he said [she might go]] ). - Now you can search for segments having one feature but not others. Use the 'and not' conjunction when extending the search query. - Results displayed in a KWIK table if you search for a segment CONTAINING other segments, a string, or concordance pattern. - The KWIK table can be sorted on any column by clicking on the column header. - The columns of the KWIK table can be dragged wider or narrower. AUTOCODING * The main window now has a button "Autocode", which opens an interface which allows you to define autocode rules: rules such as: select passive if contains 'be% @participle' One selects one of these rules, and presses "Show", and all instances in the corpus that match are presented. You can then unselect any which are false matches (not truly passives) and then click "Code". All selected instances will be assigned the passive feature. Write your own rules for the features you are interested in. HELP * This version has a revised online Help system. Each window has a Help button, and clicking on it will show the Help page for that topic. CODING * Reorganisation of the Coder Toolbar. There is now a "Other Actions" selector which offers actions: - "Export Annotations" alllows the annotation of this file at this level is saved in an XML format, with the annotations embedded within the original text (CorpusTool annotation files usually use standoff annotation, not containing the original text, just char range pointers into the original text). - "Copy Features": remembers the feature so far coded for the current segment. - "Paste Features": assigns the features remembered with Copu Features to the current segment. STATS * Bug fix: Comparisons of features between two sets, results saved to file were wrong (in tstats and levels of significance) although those shown on screen were correct (Thanks Thora). Now fixed. ==================================== Version 1.29a 30 December 2007 ==================================== MINOR BUG FIX * Using "Extend Corpus" would fail if there were subcorpora folders in your corpus with non-ascii characters (thanks Giacomo) ==================================== Version 1.29 14 December 2007 ==================================== NEW FEATURES ------------ * Import Layer: You can now import studies done with Systemic Coder software as CorpusTool layers. If you analysed the same text file with different schemes in Coder, each can be imported as a different layer. See the manual for more detail. BUG FIXES --------- TEXT STYLING * Yet more bugs which firstly meant the styled text could not be saved to file and secondly, some embedded segments were not styled properly. Fixed CODING * Where a filename contained an '&' character, the document could be saved the first time, but subsequently would fail to open. Now fixed. * There was a bug in the segmentation pane: when you mouse over a segment which starts (or ends) off the screen, the segment would disappear (only from the display). Fixed. PROJECT WINDOW * The label "Layers in this project:" should be above the display of layers, but was below. Now fixed. MANUAL * Updated the manual to match actual features of 1.2 (was for 1.1). CREATE NEW PROJECT * The wizard now stops you specifying a save location where the folder already exists. ================================ Version 1.28 29 November 2007 ================================ Mainly a bug fix release over 1.27 UNICODE BUG FIXES * Reloading a project where a layer was created with a name in a unicode script could cause a crash at startup. Now fixed. (Thanks Giacomo) * Where the root feature of a coding scheme was non-ascii, the scheme edit window would fail to open. Fixed (Thanks Giacomo) * Sometimes an annotation window would fail to save the annotations because the filename was not ascii, nor in the encoding of the OS. Fixed. (thanks Giacomo) ADD FILES WIZARD * The title of this window was "Add Layer Wizard" while it should have been "Add File Wizard". TEXT STYLER * Yet another revision to fix problems which sometimes occurred when text was styled at different annotation levels. It should now be reliable. SCHEME * Changes to the scheme will now cause an automatic update to any open annotation window affected by that change. * Changed the way SVG documents are produced, which should give no visible changes, except for the header. The SVG file should open in more browsers now (try changing the extension to .xml) (Thanks Richard). * The font used in the scheme diagram may not have been appropriate. And the font saved in an SVG file in some cases was "A" instead of "Arial". ANNOTATION INTERFACE * Small revision to Segmentation interface. Previously space was not provided for newly created underlines until the page was scrolled, meaning sometimes underlines interfered with the text below. Now, space is allocated to the underlines as required. * Small bug: in saving annotation files, the language has an '%' appended. And each save added another. This version strips % from language, and does not add any. ================================ Version 1.27 13 November 2007 ================================ * In fixing file double-click on the Mac version, version 1.26 introduced a bug for Windows that stopped a project being opened by double-clicking on the .cptr file. Fixed here. * In the Text Styler, the "del" button didnt work. Fixed. (Thanks Monica) ================================ Version 1.26 11 November 2007 ================================ MAJOR BUG IN 1.25: I reworked the code for the Coder Window, and inadvetently disabled the save button, so it was not possible to save annotations. Fixed in this version. On the Mac, where the filename has non-ascii characters, double-clicking on the file may fail to open the file. This is due to problems beyond my control. Currently CorpusTool will recognise the problem, and open the usual Opening screen, allowing you to select your project using the "Open Project" button. ================================ Version 1.25 10 November 2007 ================================ Bug fix: Seems that version since 1.22 would not allow segments to be defined by swiping on Windows if the system's country setting is one which uses "," as a decimal separator. Fixed (Thanks Richard and Monica) MAC VERSION * Selecting "View Manual" from the Help menu didn't work. Now fixed. * Changed Icon (the old ying-yang icon was for Macintosh universal binaries) * GRAPHER : The menus which popup when clicking on features and systems now have a Cancel button, as the MacOS native popups lack the corner "X" to kill the window. (thanks Richard) * The was a bug such that double-clicking on a .ctpr file failed to launch the project, due to code put in to handle unicode file names which failed. Now fixed (I hope). Maybe give problems on non-western European Macs. Let me know if so. ================================ Version 1.24 9 November 2007 ================================ COLOR CODING: revised the window somewhat to make it more usable: 1) The window is no longer of fixed width, which made problems on small screens. 2) the style options are at the bottom, allowing the text to be displayed full screen width 3) Added a delete key 4) Text displayed in the font specified by the metadata. Also, there was a bug meaning that searched for segments containing text sometimes showed segments not containing the text. ================================ Version 1.23 8 November 2007 ================================ BUG FIXES SEGMENTATION * Fixed a bug related to automatic segmentation into sentences introduced in version 1.20. Seems that sentence boundaries were not recognised properly. CODING WINDOW * Version 1.22 introduced a bug whereby one could not scroll to see the last 2 lines of a file. Fixed. * After resizing the Coding window, the top line should stay the same (as closely as possible). Now true. * Resizing the window larger could cause a redraw bug if on the last page of text. * Some chars representing open and close quote were not displayed properly. Now all converted to plain quote. ================================ Version 1.22 7 November 2007 ================================ NEW FEATURES * Save graphs as SVG (and then can use InkScape to save to other formats). BUG FIXES --------- PROJECT WINDOW * Clicking on the filename of an unincorporated file would not show the text file. Fixed SEARCH * Fixed bug such that concordance searching failed. CODER * Fixed bug in the Coding window such that making new segments and scrolling was somewhat broken in v1.2.1 * If you try to create segments in an annotation layer of type "code-document', you are warned this is not possible. ================================ Version 1.21 October 14 2007 ================================ * While the prior release allowed encoded text in text files, this version fixed remaining problems of non-ascii text in comments to codings, names of features, names of files or file-path. You can now give feature names in whatever script, could be in Chinese, etc. CODER * Totally revised how text is displayed in the Text Coding/Segmentation window. The display is more compact, only using sufficient space between lines as required by the segment markers. * Previously, clicking on the text without drag created a segment of either 0 or 1 char in width, which often became an invisible segment. The tool now checks for this and eliminates these segments. GLOSSES * Previously, the Realisation field of a feature could be used to write whatever text (e.g., a gloss of what the feature means). This version introduces a "gloss" field to each feature for this purpose. The "realisation" field should only be used by those working with formal system networks which use the realisation field for text processing. - if you have been adding realisations to features, CorpusTool will ask you if you want to convert these to glosses, and convert them automatically. * The Scheme Grapher now displays realisations and glosses separately. * The Options menu includes options for "Show/Hide Glosses" as well as "Show/Hide Realisations". * In the Coding window, previously, clicking on an offered feature showed the realisation field associated with that feature. Nw the gloss field will be shown instead. ================================ Version 1.20 October 9 2007 ================================ * Major overhaul of encodings in the program, allowing use of non-western texts, and more reliability within those European languages with accents. Each document is now provided with 4 fields: 1) Languag: The language of the text (e.g., Swedish). This will be used when loading in dictionaries, stop-word lists, etc. (currently not used) 2) Encoding: the character encoding used by the text. UTF-8 is the default, but if your text is in Chinese, Japanese, Arabic, etc., then you need to know what encoding your source text is in. The program will make a guess, but you can change this if it is wrong. 3) Display Font: What font to use to display the text file (useful if your language appears best in a particular font) 4) Display Font Size: default size your text is displayed in. * In the Project Window, each file now has a button 'i', which allows you to view or edit the above settings. NOTE FOR MACOSX USERS: Due to problems in the programming language used for CorpusTool (Python), most non-Western scripts will not display properly. The problem is being worked on and hopefully will be resolved in early 2008. SEARCH * It is now possible to search for segments containing strings with inflected characters. ================================ Version 1.15 July 29 2007 ================================ COLOR CODING * Added Save option to color coded documents, saving the color/font coded text as html. This can be opened in MS Word for use in research papers. BUG FIXES * OPENING: Selecting Quit from the Opening window caused a crash. Fixed. * HELP WINDOWS: The Help windows in Search and Stats: closing the window quit the application. Fixed. STATS * Copied a feature (with permision) from Wu Canzong's SysFan whereby the comparison between two subsets of the corpus is presented as a system network, with the probability of selection of each feature shown below the feature. See the "network" button in the Comparative Feature Stats window. ================================ Version 1.14 July 21 2007 ================================ COLOR CODING THE TEXT * On the main project window, click on a file name of an incorporated file to see the text. At the left, you can assign color and/or font effects (bold, italic, underline) to all text tagged with a given feature. This allows the patterns of selection throughout the text to be visible. E.g., use bold/italic/underline for appraisal categories, and color coding for clause type to see how appraisal is distributed. Bug Fixes: SEARCH * Save of Search results in either HTML or Text was broken. Now fixed. STATS * A possible bug when perfrorming a comparative study where the unit is set as "X in Y". Fixed. * All T-Stats and significance levels were understated in versions 1.12 and 1.13. Now fixed (was not a problem prior to 1.11). ================================ Version 1.13 July 15 2007 ================================ * Revised Manual * You can now see the features of the English lexical network under the "Show WordClass Network" under the View menu (these features can be used in concordance searching) SCHEME * Use the "Show/Hide Realisations" option in the Options menu to show or hide realisations. * Use the "Show/Hide System Names" option in the Options menu to show or hide system names. SEARCH * If you have a "Containg String" field, you can hit the Enter key to start the search. BUG FIXES * Double-clicking on a search hit in the search window would sometimes not show the hit in its file. Fixed. ================================ Version 1.12 July 5 2007 ================================ STATS * Added vertical scrolling of the stats result pages * There was a minor misalignment of headings and columns in the stats display, fixed. * Added option to generate per file statistics. This generates a table with mean, N and stad deviation for each file, for each feature under the designated search pattern. * Allow Save of all Stats pages now. * Added option to show/save statistics per document: one column for each document, to allow quick overview of how documents differ. ================================ Version 1.11 July 1 2007 ================================ OPENER * "Open Other Project" now asks for a CorpusTool project file (.cptr) rather than for a project directory. Less chance of mistake, and selection is easier. PROJECT WINDOW * Fixed the color coding of document windows to better reflect the coding status of documents: White: fully coded. Light blue: somewhat uncoded Blue: highly uncoded. * For projects with many layers and small screens, the file display area now has a vertical scroll area (thanks Rachel). CODER * You can now select a single space if you want, to represent an ellipsed element. In prior versions, this was treated as a mistake, and the selection was expanded. * Fixed a problem whereby when reloading a project, a segment might be expanded to include a preceding quotemark. * The << and >> buttons should take you to the prior or next incompletely coded segment. If the first or last segment is incomplete, it would not be shown. SEARCH * The number of hits within each file is now displayed (for Maria). * Now allows save search results, either in HTML format or as plain text. * Scrolling with the mousewheel on this window was broken. Now fixed. STATS * Now can save results of (some of) the Stats windows, either as HTML or tabbed delimited form for importing into Excel or Statistical programs. SCHEME * The scheme window now displays a meaningful title with the name of the scheme (previously, "tk"). * "Show Examples" in the Feature menu was broken in 1.10. Now works again. ================================ Version 1.10 June 28 2007 ================================ * Major revision of Search and Statistics Windows Now a MacOSX release as well. PROJECT WINDOW * The last pressed buton in the file space is highlighted. * Control-q to quit application (Command-Q on Macs) ADDING LAYER * When adding a new layer, if the layer name included a space, then this name with space was used as the root feature of the associated scheme. However, feature names with spaces give problems in other parts of the program. Now, spaces are automatically replaced by hyphens in the root feature. - if you have files where this is the case, you should rename the root feature to one without spaces. SEARCH: * Specification of search query now via menus, not by typing. * Search now allows search for concordance type patterns such as "be% # @participle" (any inflection of be followed by a participle verb with a possible intervening word). (English only currently) * the Search window now shows the number of segments that match the search query. * Previously, the column indicating degree of coding of each segment was always P (partially coded). Now displays P only for those segments which are not totally coded. * Search now displays hits in each file as they are encountered, instead of waiting for all files to be processed. STATS: As with Search, specification of subcorpus to describe or compare is now done via layers. - Added options for Descriptive Statistics as well as comparative stats. - Added new types of analysis: general statistics shows the word counts, sentence counts, lexical density, etc. for the specified subcorpus. Also Word Statistics shows the words which are most "special" to the subcorpus. SCHEME * Fixed a problem with schemes with a gated entry condition. MACOSX * Produced a version for the Mac. ================================ Version 1.09 May 25 2007 ================================ * For the Info pane, the count of sentences was wrong for texts saved with unix line feeds \r\n * In the Coding interface, some Microsoft quote chars did not display properly, now display as plain single quote. * In the New Project wizard on MacOSX, the Next button did not always enable properly. Fixed. ================================ Version 1.08 March 29 2007 ================================ * Text files containing linefeed rather than newlines were not displayed properly in the Coding window. Fixed. (mainly affected files taken from unix). * Files with parentheses in the filename gave problems. Fixed. ================================ Version 1.07 March 18 2007 ================================ * There was a fairly serious bug: if you open an annotation window, create a new segment without coding it, then open the scheme window for that layer and rename a feature, the system would crash. - partially fixed. (you will now be asked to save any open coder windows in this context). * When closing an annotation window, and asked to Save the codings, if you said no, the codings were not deleted, and shown again if you opened opened the document again in the same session. Now, saying know will not save the changes. ================================ Version 1.06 March 7 2007 ================================ Mainly a revision of the Search interface PROJECT WINDOW * When deleting a layer, the buttons for that layer associated with each file were not deleted. Now fixed. (Thanks Eveline) SEARCH * Where the same feature appears in two layers added ability to refer to a feature in a specific layer. use "layer:feature" rather than "feature". * Added a Help button to show the syntax of the search query. * Fixed a bug such that one could not search for feature names including '-' or '_'. (Thanks Cesare) * If one analysis file was corrupt, search would find nothing. Now fixed so that one corrupt file does not stop searching. * Fixed a bug such that when one clicked on a hit to open the file, the segment was not on the visible screen, but one line below. (Thanks Cesare) * The search form: feat containing feat wasn't working before. Now works. SCHEME Added option for "Copy to clipboard" which allows the scheme to be copied to the clipboard under windows, and pasted into an MS Word document (or other locations which allow paste of wmf files). ================================ Version 1.05 Feb 23 07 ================================ CODER * TABS in the original text caused problems when segmenting text. Now fixed. (Automatically converted to spaces for display purposes only) * Added toolbar in code-documents window as well (options for Save, Close, Show XML) BUG FIXES ========= SEARCH - fixed a bug which stopped this working in the last version OPEN: fixed a bug which caused a crash when opening by clicking on a .ctpr file if the folder path included accented characters. ================================ Version 1.04 Feb 13 07 ================================ MENUS - Added menus for: PROJECT MENU * Switch Project: close this project and open another HELP MENU * View Manual: Show the pdf version of the manual * View Release Notes: shows this document * View Licence Terms: shows the COrpusTool licence * CorpusTool Home Page: opens a browser window with the homepage on Wagsoft * About CorpusTool: gives version of your installed version. ================================ Version 1.03 Feb 09 07 ================================ FILE INCORPORATION: Substantially modified the way files are added to a project. Each project has a Corpus folder, which contains subfolders (subcorpora) holding text files. These files form the "Corpus" of the project. In previous versions, the user could "Add" files to the project from this corpus (with the "Add File" button). Files in the corpus but not in the project were not displayed. Now, all files in a project's corpus folder are displayed in the Project Window. Those which have been "incorporated" display buttons for annotating each layer. Those which have not been incorporated just have a button to allow the file to be incorporated. The reason for this 2-state approach to corpus files is that a user might want to distinguish files that they are currently working on (annotating) from those they may wish to add at some future time. The Files section of the Project Window thus has two sections, one for incorporated files, and one for pending files. ================================ Version 1.02 Feb 07 07 ================================ ADDITIONAL FUNCTIONALITY * ADDED mouse scrolling to project, stats and search windows BUG FIXES FIxed bug such that when a new layer was added, the buttons for the layer were not displayed immediately. IMPORT CD3: fixed some bugs which stopped this option working MOUSESCROLLING: the mouse scrolling, introduced in last release, scrolled in the wrong direction (up for down etc). Fixed. ================================ Version 1.01 Feb 05 07 ================================ PROJECT WINDOW * Previous version required the user to manually move files into the Corpus folder. Now, under "Add Texts", there is a button for "Extend Corpus" which asks the user to nominate a file or folder of files to add to the project's corpus. * Files are now listed in alphabetical order * TOOL TIPS: started putting in tooltips for some items. When you hold you mouse over a button for 2 seconds, a small message pops up to explain what it is for. CODER, GRAPHER * Added mousewheel scrolling. BUG FIXES --------- CODER * If selecting the start of a segment boundary, and the end of that segment was offscreen, the screen redrew to put the whole segment onscreen, which interferred with the moving of the segment boundary. * Fixed yet another problem involving UNICODE, this time with spanish characters in comments. * Fixed a rare problem in drawing underlines. PROJECT WINDOW * There was a problem when resizing the project window if there were many files in the project. The window would in some cases resize itself back to the original size. This is only partially fixed. ================================ Version 1.0 Jan 04 07 ================================ ALL * Buttons have been changed to a 3D Aqua design, not only to look better, but these buttons work equally on Macintosh as in Windows, which was not the case with the standard built-in buttons. CODER * If the segment curently selected is ignored, the highlight is shown in grey instead of the usual green. * Previously, when placing the cursor over a segment underline,only that part on the current line was highlighted. Now the underline of the whole segment is highlighted. * When cursor is over a segment endmark, it turns red (as usual), but now the underlining of the whole segment is highlighted, to show more clearly which segment this is the endmark of. * The spacing between underlinings of overlapping segments was increased slightly to allow selection easier, and the spacing between text lines was also increased. ================================ Version 0.99 Jan 04 06 ================================ NEW FEATURES ------------ CODER * Added buttons '<<' previousIncomplete and '>>' next incomplete which move to the next segment which has features to select. PROJECT * The "Add Layer" button now brings up a Wizard to collect details of the new layer, as with "Create Project". SCHEME * The Options menu was cleaned up, and the "Save Diagram As PDF..." button prompts the user for a save location. ================================ Version 0.98 Jan 04 06 ================================ MAJOR CHANGES ============ PACKAGING The CorpusTool is now distributed in an Installer which places all program code in the user's Program Files folder It also creates an association on .cptr files such that double-clicking on these files (in your project folder) will open that project directly. There is no longer a Projects folder in the CorpusTool folder. Basically, users store their project folders where they like. The system thus only keeps track of the last project opened. SCHEME * Any change to the scheme is now propagated throughout all annotation files which use that scheme. For instance: i) Rename feature - renames feature in all files ii) Delete feature/System: all codings involving the deleted features or those below are deleted. iii) move feature: where this makes a coding inconsistent (e.g., seelction of two features in one system) then one of the inconsistent selections is eliminated. As a result of this, any change to a scheme needs to be saved immediately. As of this version, there is no save button on the scheme window, as the scheme is saved whenever changed. Also, as a result of changing the scheme, any and all annotation files of that layer may be updated and saved. Be warned, there is no undo at present. MINOR CHANGES ============= GENERAL * Fixed problems related to storing projects in paths with accented chars (e.g, /César). SCHEME * The "Graph From" button now shows at the top the features closest to the root of the network. CODER * Added a "Edit Scheme" button which opens the scheme for editing. TEXT * Segments which are ignored are shown with a grey underline, not green, to better indicate their status. BUG FIXES ========= OPENER * Was possible to press Open Project, go look at the desktop, and then select the CorpusTool window, and the Dir selection windows was hidden. Now it will always come to front with the CorpusTool window. SCHEME * Inclusion of an "or" complex entry condition led to a bug in loading annotation files for that scheme. Now fixed. CODER * Fixed a bug in relation to deleting segments after a comment was changed which caused the window to freeze (Thanks Cesar!) * Comment lines longer than window width now wrap to new line. ================================ Version 0.96 27 Nov 06 ================================ NEW FEATURES ============ SCHEME * SHOW Examples: there is now an option under the feature menu which, when selected, will show all examples in the Corpus of segments with that feature. BUG FIXES ========= GENERAL * Sometimes it was necessary to click several times to close a window (e.g., the Change Entry Condition Window). Hopefully fixed. * If you have an unsaved annotation window open, and quit the main window, prviously you would not be asked to save. Now, you will be asked to save each document before closing down. SCHEME * Previously, it was possible to open a Scheme window several times, each with different edits. Now, clicking on the Edit scheme button will bring forward the window if it is already open. SEARCH/STATS * Previously the Search function would match segments marked as ignored. Now only unignored segments are matched and displayed. Since the same search code is used in the Stats window, now ignored codings are not included in the counts. CODINGS * Fixed a problem such that most codings without any marking as ignored or not might be taken as ignored in some contexts. IMPORT CODINGS * Fixed "Import Codings" so that it stores the corpus within the new Project folder (as required in p.95) * Fixed a problem such that "Import Codings" dropped the ignore field of the Coder segments. ================================ Version 0.95 24 Nov 06 ================================ REORGANISING ============ * MOVED CORPUS FOLDER: Changed the system so that the Corpus folder is stored WITHIN a project folder, so that if one wants to move a project from one computer to another, the project folder contains all the information one needs. The toplevel Corpus folder is no longer used. FOR YOU TO DO: If you have an existing project, move the toplevel corpus folder into the particular project's folder (e.g., into the Instructions folder). * CREATED A PROJECT DEFINITION FILE In the past, the definition of Layers (name of layer, kind of segmentation, which scheme to use, etc.) was done via the Layers folder, with one file per Layer. However, the order of loading these layer files seemed to change daily, as Windows was in charge of this. Now, all Layer definitions are in a file at the toplevel of your project folder, with extension .ctpr. In future, this will allow: - Double click on the .ctpr file to open the project - Permanently reorder the layers in your project, so that the buttons apprea in the order you want. FOR YOU TO DO: Nothing: if CorpusTool does not find a ctpr file, it will make one from the contents of your Layers folder, and then delete this folder. NEW FEATURES ============ SCHEME * You can now move systems up or down. Click on the system name and select "Move System Up/Down". BUG FIXES * If you try to rename a feature, and the new name already exists, then you will be asked if you want to move that feature to the current location. (unless the original feature has sub-systems, in which case the rename is not allowed). CODER * The box showing systems still to code has been replaced by a box which will show the realisations of one of the features to select (click on the feature to show) BUG FIXES ========= SCHEME * Systems with a complex entry condition were deleted when loading the network again. Fixed. SEARCH * Long segments now appear totally on one line rather than coming on two, which was unsightly. CODER * The progression through systems as you code is more logical (depth first rather than breadth first) ================================ Version 0.94 19 Nov 06 ================================ NEW FEATURES ============ * The opening of Windows is now properly controlled, such that you cannot open a coding document twice. Rather, the one already open is presented to you. This is most noticeable in the Search window: clicking on sucessive hits in the search window will mvoe you through the Coding document's window. BUG FIXES ========= SEARCH * Before, clicking on a hit which appeared more than 100 lines into the file would produce a blank screen in the Coder window. Now fixed. CODER * If the coder makes changes to a segment's coding while loading, the document should be marked as "changed", and thus when closing the window, the user should be prompted to save, even if the human did not change anything. Now fixed. * When, during coding, a feature is added to a segment, and this feature is the entry condition to the system, then the feature in the system is now added also. * If you delete a selected feature, currently selected features which depend on the deleted feature are also deleted (recursively). * You can no longer delete the root feature of a segment (this feature applies to all segments). ============================== Version 0.93 17 Nov 06 ================================ NEW FEATURES ============ CODER * You can now press the delete key to delete the current segment. * A Save and Close button were added. BUG FIXES ========= CODER * Fixed problem when including non-ascii text in comments. - they should save properly now. * Errors in saving a document were not reported to the user, fixed. * Comments recently entered were wiped if a feature was selected or deleted straight after. Fixed. * Several fixes in regards how features of a segment are handled: - When loading a document into the coder, any features in the coding but not in the associated scheme are not deleted. - Any feature absent from the coding but implied by the other features is automatically added. - Features in the coding box are now sorted in terms of delicacy. GRAPHER * Fixed problem when pasting non-ascii text from a source outise the CorpusTool. * Errors in saving a scheme were not reported to the user, fixed. SEARCH * Search stopped on first found when searching "A in B" ================================ Version 0.92 6 Nov 06 ================================ Improvements & Bug Fixes ======================== GRAPHER - Fixed errors resulting from deleting a system or feature. Network became unloadable!!! - When all realisations of a feature were deleted, the program continued to show an empty box. Fixed - Long System names now drawn on 2 lines to give more compact graphs CODER - Selection of segments by clicking on segment underline should be easier. - Dragging of segment boundaries should be easier. - Creating new segments should be easier.