Hi Jakob,
On 09/01/06 14:54, Jakob Lechner wrote:
We would like to implement "Smart Tags"-like functionality in
Openoffice Writer.
According to our current plans we want to realize this feature as
described below:
3. In case of MS Office there are simple and complex smarttag
libraries. For simple libraries just a xml smarttag specification
has to be provided (see [1]). Complex libraries require a COM DLL
that implements a special interface [2]. The actions that can be
specified with xml smarttags facilitate basically the invokation of
an URL. It is though possible to include the recognized smarttag
word as a parameter in the URL. In smart tag DLLs of course
arbitrary actions can be programmed. In a first step we want to
provide support for xml smarttags.
But both the 'xml files smarttags' and the complex smarttag libraries
will implement the SmartTagLib interface, is this correct? From this
point of view there actually is no difference between these two. So we
would have to focus on the design of the SmartTagLib interface and its
usage in the Writer, and the xml smarttags thing is first concrete
implementation of the interface, right?
6. In order to be able to provide access to multiple smarttag
libraries there should also be a Smarttag manager. This class loads
available smarttag libraries implementing the specified interface
and constitutes the access point to smarttag libraries for the other
parts of the program. The smarttag manager should be able handle xml
smarttags but later on it could be extended to deal with UNO
components as well.
See comments to 3.
9. Implementation details:
The smarttag implementation consists of three parts:
1. Underlining recognized smarttag terms in the opened document.
2. Xml smarttag classes that process xml files, provide the
recognize function and execute the defined actions.
3. GUI integration of applicable actions by a popup menu.
The process of recognizing and underlining words is very similar to
the online spell checking function. Therefore we would implement
another list just like the wrong list that is used for maintaining
misspelled words.
Let me sum this up and add some comments to see if I understood your
design correctly:
1. If the user types a word, the SmartTagManager is called from the
Writer core to check if there are any SmartTags available in the current
word (or better paragraph and start/end indices to avoid string
operations? Could be done per paragraph during idle time. Should we pass
the current locale to the Recognize function?)
2. The SmartTagManager basically is a container for all currently
registered SmartTagLibraries. The SmartTagManager calls the 'Recognize'
methods of all registered SmartTagLibraries.
3. If the recognize function of any of the registered SmartTagLibraries
returns 'true', the current text region is stored in a list similar to
the current wrong list. The regions stored in this list will be marked
in a special way during painting.
4. If the user activates the context menu over a marked region, the
getSmartTagRefsByWord function is called to obtain a vector of
SmartTagReferences
5. For each SmartTagReference in this vector, a string/action pair is
obtained from the associated SmartTagLibrary, the strings will appear in
the context menu and the invokeAction function will be triggered if a
context menu entry is invoked. A UNO textrange object can be passed as a
parameter of the invokeAction function to enable the SmartTagLibrary to
change the text etc.
One more idea: What about separating the SmartTagLibrary interface into
a Recognizer and Action part, the way MS does? This way it would be
easier for external developers to implement new smart tag libraries for
both MS Office and OpenOffice.org.
Best Regards,
Frank
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
