Alphabet Priority

The SetAlphabetPriorityHRC function provides a hint to the recognizer about which alphabet the user intends to use while inputting handwritten characters. Some written symbols have more than one interpretation, depending on alphabet (this is especially true in Japanese, which uses multiple alphabets). A crosslike character like the English plus sign (+) could also be interpreted as a lower case "t", katakana "na" or "me", or Kanji "juu" (the number 10), for example. Moreover, there are both SBCS and DBCS versions of English and katakana characters, corresponding to half-and full-pitch widths, so it's useful for a user or application to be able to pre-specify widths and character set preferences to minimize recognition errors.

Applications may use the SetAlphabetPriorityHRC API to set width and alphabet priority for HRC scope, or the SetPenAppFlags API to set width (DBCS or SBCS) preference for application scope. The Tool Palette application may be used to set system-wide priorities. Any priority set into an HRC supersedes application priority, which supersedes system priority as set (and if set) by the Tool Palette; in the event of not Tool Palette, the recognizer's default priority is used. The ALC_GLOBALPRIORITY bit will be set in all default HRCs, unless it is explicitly cleared with SetAlphabetPriorityHRC. To specify no priority, as opposed to default priority, an application should clear the ALC_GLOBALPRIORITY bit and set ALC_NOPRIORITY.

  • The following constant has been added to the "Gesture" section.

Kana-Kanji convert. (Japanese version only)

  • In the "Unboxed Recognition" section, the following paragraphs have been changed.

An HRCRESULT object does not contain a normal text (ASCII for English) string representation of a guess. This is not possible since a guess might be made up of a gesture, shape, or other entity that has no text equivalent. Instead, an HRCRESULT contains a string of symbol values, which are 32-bit numbers type-defined as SYV.

Symbol values can represent geometric shapes, gestures, letters of the alphabet, Japanese kanji characters, musical notes, electronic symbols, or any other symbols defined by the recognizer. The Pen API provides the function SymbolToCharacter to convert the null-terminated symbol string in HRCRESULT to a normal text string.

  • In the section "Getting Results from the RCRESULT Structure," note the version change in the first sentence of this paragraph.

The RCRESULT structure applies only when an application calls either of the version 1.x recognition functions, Recognize or RecognizeData. In this case, the system sends a WM_RCRESULT message to the application. The wParam of this message contains a REC_ submessage that indicates why recognition ended. The lParam of WM_RCRESULT points to an RCRESULT structure, which contains all the results of the recognition.

  • Note the change to "text string" (formerly "ASCII string") in this paragraph.

The RCRESULT structure identifies the recognizer's "best guess," which is the guess in which the recognizer places the most confidence. With this information, an application can conveniently retrieve a text string of the best guess by calling SymbolToCharacter:

Software for developers
Delphi Components
.Net Components
Software for Android Developers
More information resources
Unix Manual Pages
Delphi Examples
Databases for Amazon shops developers
Amazon Categories Database
Browse Nodes Database