Update, Char table: Values of attribute MultiStateType redefined. Old values: 0=OR, 1=AND, 2=TO, 3=EXCL. New values: 0=EXCL, 1=OR, 2=AND, 3=TO/OR, 4=TO/AND. This was necessary for non-continuous ranges; it allows to differentiate between '1-3 or 5-7' and '1-3 and 5-7'.

Update, Modifier table: Attribute SkipOperator redefined; type changed from Boolean to Byte and renamed to Operator. Values now indicate operator meaning (AND/OR/TO) to improve language independence. All table updates are automatically performed when a backup of a previous version is restored.

Bugfix, DELTA import: Slashes ('/') not followed by a blank or new line where already correctly treated in character names, but they were erroneously interpreted as the end of the name in item and character state names.

Bugfix, DELTA import: Item description notes after ranges were erroneously imported to first state of range, i.e. '1-3<note>' became '1<note>-3'.

Bugfix, DELTA import: The test whether incompatible local numerical formats are defined in the local Windows version failed in some cases and was missing from HTML-form re-import. This could lead to '12.5' being imported as 125 if the decimal separator was a comma, and the 1000-separator was set to a period ('.').

Update, DELTA import: The default of the import option 'Item name comments are authors' was changed from true to false. Comments/notes after item names are thus written to the Notes attribute, instead of being appended to the item name. To restore the previous behavior manually select the option before importing.

Bugfix, HTML-form re-import: Long postings created using Netscape contain a '!' at position 922. Depending on whether this '!' occurred in a note, text, or elsewhere, data may have been changed or errors may have occurred.

Bugfix, HTML-form re-import: The last character was missing from the supplementary DELTA code showed to find an error. This bug only occurred when already another error occurred during re-import.

Bugfix, Identify: You could not specify an item name condition immediately after setting a condition using categorical characters.

Bugfix, Character subsets: The test whether all controlling characters are included in a character subset reacted to the 'No'-button with 'Yes' and vice versa.

Bugfix, MultiItemEdit: Essential controls remained permanently invisible, making the form unusable.

Addition, DELTA import: The file names 'CNotes' and 'CImages' are now recognized during Confor-compatible DELTA import.

Addition, DELTA export: An option to start each character in the item description on a new line (instead of being written continuos, separated by blanks) was added.

Update, HTML-form creation and HTML-form re-import: The major numeric attributes (Min, Lower Range, Mean, Upper Range, Max) are coded as 1-5 now. This is intended to make the HTML form transfer code more independent from DeltaAccess conventions. No changes to existing forms are required, the update is fully backwards compatible.

Addition, HTML-form creation: The index options have been moved to a separate page and a new option to create an internal index on the second word of the item names in the index files has been added, see Index file tab page. Some additional changes to the code were made, which should make it more robust, but which should not affect the use of the dialog box.

Addition, HTML-form re-import: An option to ignore the original item ID (thus not adding a note "Revision of …") and an option to directly overwrite the original item was added.

Addition, Item subsets: An option to directly select item ID numbers using the new multiple selection dialog (instead of always using Identify with a condition) was added. The new dialog also replaces the Reorganize dialog for selection of character subsets.

Update, Switchboard dialog box: If no projects are present (e.g. when first using DeltaAccess), only the 3 applicable tasks are displayed.

Update, DELTA import: Several data sets analyzed use the DELTA Heading directive with a trailing colon (example: '*HEADING: Amorales'). This usage is not formally described in the DELTA documentation, but actually used in the example in that documentation. Similarly, SHOW is used occasionally with a colon. Starting with this version, such leading colons are ignored. Also, #DATE and #TIME are removed from Heading. In static documents they can be replaced with the date and time of creation, but this is not appropriate for the title of a dynamic data set in a databases, which can be used at any time.

Update, HTML-form creation: The coding of text characters in HTML forms was changed. Example: CID 91 of Type TE used to be coded "T91_TE" and is now coded as "T91"! This change should be fully compatible, i.e. results from old forms can be imported without problems and results from new forms can be imported by version 1.5 of DeltaAccess.

Bugfix, HTML-form re-import: The problem with spurious '!' already addressed in 1.5 could reappear if the '!' appeared 1 or 2 characters after a '%' character. I hope this is the final fix for this problem!

Update, Identify: Added descriptor project and application information ("About").

Bugfix, Restore of project backups: A serious and very regrettable error was introduced in the previous version (1.51), causing a failure to restore saved backups. I apologize for the problems caused by this bug!

Update & Addition, Restore of project backups: It is now possible to restore either a selected project or all projects stored in a backup mdb-file. A new dialog box was added for backup and restore. Further, 1.6 is the first version where all linked projects and subset projects can be correctly restored. This includes mixed projects (i.e. an independent item project with a link/subset to the character definition of another project). Please note that while dynamic item subsets created in version 1.6 will be restored unchanged, dynamic item subsets created in previous versions of DeltaAccess are converted to equivalent static item subsets, i.e. they become based on IID numbers.

Addition, Installation: The handling of the restore function during upgrades to a new version of DeltaAccess has been improved. The Switchboard (the dialog box which comes up at startup of DeltaAccess) now directly offers an upgrade option if a new DeltaAccess version is opened (i.e. if no projects have yet been defined). This should make the upgrade to new versions of DeltaAccess easier.

Bugfix, Importing DELTA files: The following problem is due to a change in the libraries distributed by Microsoft. It does not affect customers with a US or British version of Microsoft Access, nor does is affect customers with any international version of Office 97 as it was originally distributed. So far, it has been detected only under the following circumstances: (1) Access running under German NT 4.0 with service pack 4 installed (service pack 3 does not yet produce the problem). (2) Office running under Windows 95 on a computer where the German Visual Basic version 6 has been installed. The change in the VBA libraries is not documented to the best of my knowledge and not logical either. It also contradicts the original documentation published by Microsoft. It basically means, that the settings for the decimal separator (English uses a period, but some other countries a comma) and the 1000 separator during conversion of strings to a number are now ignored. The number type seems to be hardcoded to the most frequent use in Germany, which means that numbers using a period (like DELTA does) can no longer be imported. This required a deep change in DeltaAccess, and I hope that no further problems result from this.

Bugfix, Importing DELTA files: Errors occurred if an item name in DELTA coded text files had one of the following unusual forms: (a) only a number, e.g. '# 1./'; (b) only a comment, e.g. '# <x>/'; or (c) nothing at all, e.g. '#/'. The item name is still required in DeltaAccess, but a prompt to enter an appropriate item name is now automatically added in the latter two cases.

Bugfix, Importing DELTA files: If a UNIT notation was used in categorical characters (as is not allowed in DELTA, but common in data sets originating from TAXASOFT) DeltaAccess would already in previous version attempt to append the UNIT to the character states. However, the unit string was erroneously not removed from the unit field and would therefore be duplicated in the output.

Update, Importing DELTA files: If an imported file contains character heading definitions, but no named character group definitions (= Define Characters directive in IntKey), the character headings are also used as character groups.

Update, Exporting to DELTA files: A warning is now issued if the character ID number are non-continuous (e.g. 1, 2, 4, 5). While DeltaAccess has no problem with this situation, many other DELTA compatible programs including Confor will not import such a data set.

Bugfix: If a text character was created or modified in the Character Editor Form of version 1.51, the corresponding TE state was not correctly created. This lead to problems in the item editor later on. The error did not occur if a project was imported from a previous version or from a DELTA text file. It was due to a tiny apostrophe erroneously deleted. NOTE: You must repair such a data set! Please go to "Analyze data" (e.g. using the Switchboard), select Analysis type: "Repair and maintenance" and Open analysis: "Add CS: Internal state TE for all text char (TE)". Click on OK, Answer YES twice, answer YES if an error message appears. This will insert the missing TE states. (Example for error message: If you have a total of 6 text characters, DeltaAccess will try to insert 6 characters. If the error message reports that 4 records could not be inserted, this means that 2 TE states have been added, repairing the 2 damaged text characters.)

Bugfix: Character Heading Editor and data model for headings: In previous versions of DeltaAccess, cascades deletes were active between headings and characters. Deleting a character heading that was used as a CharHeading or ItemSubheading in the character definition would thus cause the deletion of all related character records, which in turn would delete all related character states and item descriptions for these characters. Since this behavior is unintuitive and dangerous, it has been changed. In tables or queries, headings that are not used as CharHeading or ItemSubheading you can no longer longer be deleted directly. To delete such a heading you must use the Character Heading Editor or the Reorganize form, where the links to the Char table (CharHeading, ItemSubheading) are now automatically removed before the heading is deleted.

Bugfix, Create new item project linked to a subset of an existing character definition ("mixed projects"): The setting of referential integrity relations for the new project was faulty. Projects could be used, but changes in the original project (esp. renumbering/reordering) were not reflected in the linked subset project. Also, it was not possible to directly overwrite an existing character subset project. The project had to be explicitly deleted, and then recreated.

Bugfix, Character subset: Excluded characters remained in the item description if no item subset was defined at the same time. The error would show only when opening the project queries directly, or when the special "ALL" view of the item editor was used. Most analysis queries would have been unaffected.

Update, Item subsets: The method to define subsets has been reworked, to simplify the management of subsets and to enable to change the item sequence (and the IID numbers) in the future. With previous versions of DeltaAccess, changes in the IID number would not be reflected in the subset. Character subsets were already resistant to changes in the CID numbers; a similar method is now used for item subsets. Please note that this does not apply to dynamically defined item subsets (a condition defined through Identify). These subsets will be invalidated by changes in the CID (i.e. the character, not the item sequence!). See also Further information about subsets.

Bugfix, Identify: Under certain conditions a single item would be listed multiple times in the item list (after clicking evaluate). All items should now appear only a single time.

Bugfix, Summarizing items: An error would occur, if descriptions with multiple, non-identical modifiers were compiled into a summarized item (and the option to ignore modifiers was deselected). Example: categorical character leaf number, species 1: "6-10", species 2: "about 6-10" would be ok, but if a species 3: "usually 6-10" would be added an error would occur because the combined modifier does not exist. The new version now ignores all but the first modifier. Any better ideas how to deal with multiple modifiers during Summarize?

Bugfix, Summarizing items: If the temporary Item-summarize table was not manually removed in the Summarize items dialog box, it would not be removed together with the other project tables when the entire project was later deleted. As a related upgrade, an option to automatically delete the summarize table was added to the Summarize items dialog box.

Bugfix, Item Edit Form, Preset Items: The button was not working any more; the list of items to select from was empty.

Bugfix, Editing character headings: On computers with a VGA screen resolution and large fonts the upper part of the form could become hidden behind the toggle buttons used to switch between heading and link-group editing. Also, the Notes and HID attribute were misplaced when the width of the form was changed.

Bugfix, Appending data to a project: Errors could occur if the project defined character headings or character dependency. Also, updated to add the project name rather than a number after the character or heading names.

Update, Deleting descriptor projects: A warning message is now issued if attempts are made to delete a base project without deleting the dependent linked projects.

Bugfix, HTML-form re-import: If the information posted back from a DeltaAccess HTML form is corrupted and cut off at some point during the transmission (esp. by email), the import process could fail to finish. The bug caused an infinite loop and the process would "hang". This has been fixed and appropriate error warnings have been added. However, please note that for old HTML forms (created by DeltaAccess versions before 1.6) there is no method to determine whether indeed all characters have been returned. You will not be warned if the truncation removes trailing characters at the "correct" point, i.e. as they could have been send by the HTML form itself. Forms created by DeltaAccess 1.6 have a special end of data marker ("DHF_END=1") which is checked.

Bugfix, HTML-form re-import: If the information posted back from a DeltaAccess HTML form is corrupted and cut off at some point during the transmission (esp. by email), the import process could fail to finish. The bug caused an infinite loop and the process would "hang". This has been fixed and appropriate error warnings have been added. However, please note that for old forms (created by DeltaAccess versions before 1.6) there is no method to determine whether indeed all characters have been returned. You will not be warned if the truncation removes trailing characters at the "correct" point, i.e. as they could have been send by the HTML form itself. Forms created by DeltaAccess 1.6 have a special end of data marker ("DHF_END=1") which is checked.

Bugfix, HTML-form creation: Radio buttons had no "neutral" position. The HTML form could start up with no button selected, but if anything was erroneously selected, it was impossible to revert to "I want to say nothing". An additional radio button with the default labeling "none selected" has now been added for this purpose. (Suggestions for an improved label text are welcome!)

Bugfix, HTML-form creation: The color of states was ignored (the other color settings worked).

Bugfix, HTML-form creation, Localization strings: The labels for numerical units (default 'Units:'), special text in a categorical character (default 'Other (enter as text):'), and for the new item in the index file (default 'Add a new item') remained unformatted. This could lead to problems if the localization string was changed and contained accented characters or special characters like '>', '<', or '&'. Additional strings that were fixed in previous versions were added to the user interface to allow localization. Update: Some localization string defaults were changed.

Bugfix, HTML-form creation, Output options: For radio buttons and pick lists (i.e. in the case of exclusive characters), modifier and notes field were always displayed, irrespective of the setting in the dialog box.

Update, HTML-form creation, Formatting options: The default for the number of character states up to which radio buttons were used (a pick list is used if more states are present) in exclusive categorical characters was changed from 7 to 4. The number now refers to the actual radio buttons displayed on the form. Previous versions of DeltaAccess counted special states present in the character definition, even when they were excluded from the HTML form.

Update, HTML-form creation: Two new special fields have been added to the form definition. This updates the Descriptor HTML interchange format from version 1.0 to 1.1. Both versions are recognized by the HTML-form re-import starting with DeltaAccess version 1.6. The first change introduces a hidden data element indicating the end of data. This element ('DHF_END=1') is evaluated to detect truncated transmission, which may occur especially if the data are returned by email. The second option adds a field to enter a required user name or email address of the person filling in the form. This is especially useful when using a CGI-gateway script that can reject a form where this information is missing. The user name field can be suppressed by leaving the localization string (tab page Localization strings) empty.

Update, HTML-form creation, Formatting options: Text characters which are set to several lines height (i.e. where a TEXTAREA with scrollbars is displayed) are now created with the attribute WRAP=SOFT. Thus lines wider than the display area wrap to a new line, without being broken in the data posted back to the database. Hard returns can be inserted as before and are transmitted.

Addition, HTML-form creation: If character headings are used, and if an internal index to character headings exists at the top of the document, a new option has been added to insert links back to the index of headings after each character heading (new checkbox on tab page Output options, phrase on tab page Localization strings). The default is the text "-- back to index", but it is possible to use any text or a small image instead. If the phrase entered here ends with '.gif', '.jpg', or '.png' it is assumed to be a relative or absolute URL to an image that you supply. Please contact me if you want to use other file extensions than the 3 mentioned.

Addition, HTML-form creation: Optional hyperlink to return to the index of item descriptions file added.

Addition, HTML-form creation: You can now choose to overwrite all files instead of being prompted each time a file is already present.

Update & Additions, Reorganize definition dialog box: The formerly independent reorganize character and reorganize character state dialog boxes have been joined and facilities to change the sequence of character headings and items have been added. Further, new functions have been added to insert a character, character heading, or item above an existing one, and to copy them to another project. Finally, a special function to reset the sequence of categorical states in the item description has been added.

Addition, Reorganize subsets: The Reorganize definition dialog box now recognizes if it is used on a subset and warns if the character headings, states, or description tab pages are selected. It enters a special mode for characters and items, allowing to add or remove them from the subset definition or a linked project (not possible for dynamic item subsets).

Bugfix, Reorganize character state: When, while joining two states within a single character, the new state code was accidentally left empty (i.e. the command was given to join a state with itself), the state and all linked item descriptions were deleted (because they were treated as remaining duplicate item descriptions!).

The formerly independent reorganize character and Reorganize character state: dialog boxes have been joined and facilities to change the sequence of character headings and items have been added. Further, new functions have been added to insert a character, character heading, or item above an existing one, and to copy them to another project. Finally, a special function to reset the sequence of categorical states in the item description has been added.

(Cosmetic) Bugfix, HTML-form creation: Some Output options: were displayed in such a way that they covered the label, making the label unreadable. This problem affected only the user interface, not the functionality of the program.