WD: How Find Fast Determines Which Method to Use to Index Files (136386)

SUMMARY

The following process is how Find Fast determines which files to index on specified local and network writable media:

File Name Extension:
1. If you choose to index Microsoft Office files, including Microsoft Word, Microsoft Excel, PowerPoint, Microsoft Project and Binder Files, Find Fast enumerates (makes a list of) all files matching these file name extensions on the specified drive.
2. If you choose to index all files, Find Fast enumerates all of the files in and below the selected directory with the exception of files matching the "exclude" extension list.
Find Fast then calls the OLE 2.0 GetClassFile function on every file matching the criterion specified in step 1 above to determine the file's Class ID.
1. If the Class ID identifies the file as a Word 6.0-7.0, Microsoft Excel 2.1-7.0, PowerPoint 7.0, or Office 95 Binder file, Find Fast uses its internal Word, Microsoft Excel, PowerPoint, or Binder filter to parse the file.
2. If a file is identified as a Microsoft Project file, Find Fast uses its internal text filter to parse the file.
If Find Fast does not recognize the Class ID, Find Fast then queries the Windows 95 Registration Database to determine if any external converters that support running in NoDialogs mode are installed. Find Fast then polls each registered NoDialogs converter that expects (recognizes) the file extension of that particular file. If a converter recognizes the file, the converter will be used to parse the file.
If a file does not have a Class ID that is recognized by Find Fast, does not include a file extension, does not have a file extension expected by any of the installed external NoDialogs converters, or does not have a format that is recognized by the converter(s) that supports its file extension, Find Fast reverts to use of its internal text filter to parse the file.

If a file is filtered by Find Fast's text filter, but the file does not include any intelligible textual data, Find Fast will still index the file, but truncate any "words" over 80 characters.

NOTE: The following file name extensions comprise Find Fast's "exclude" extension list. Find Fast will never index files with these extensions because these extensions denote file types that will not include useful textual data for indexing purposes:

.386, .BMP, .CHK, .COM, .DLL, .DRV, .EXE, .FFA, .FFL, .FFX, .FLT, .FRX,
.FTS, .GRP, .ICM, .ICO, .IDX, .JPG, .LEX, .MAP, .MMF, .NSF, .OBJ, .OBX,
.PAR, .SCR, .SWP, .SYS, .TIF, .TTF, .VBX, .VXD, .WAV, .WMF, .WPC

MORE INFORMATION

Find Fast is installed by Office 95 Setup for access through the Windows 95 Control Panel.

Office Extensions include:

      Extension              File type
      ---------              ---------
      .DOC, .DOT             Microsoft Word 6.0 or 7.0 document
      .XL*                   Microsoft Excel Workbooks
      .PPT, .POT             Microsoft PowerPoint Presentations
      .MPP, MPW, .MPT, .MPX  Microsoft Project files
      .OBD, .OBT             Binder files

The method of file name extension enumeration that Find Fast uses is the same method all Office applications use to populate the list of files in the File Open dialog.
As stated above, Find Fast includes its own internal Word 6.0-7.0, Microsoft Excel 2.1-7.0, PowerPoint 7.0, Binder, and Text filters.
"NoDialogs" mode is a mode during which an external text converter will not echo any dialogs to the screen. For example, if a converter would normally prompt for a missing linked graphic, it will not do so when called in NoDialogs mode. Converters that support NoDialogs mode can include a sub-key in the Windows 95 registry titled "NoDialogs". When a converter is set up, the default of the following converters is to be set with NoDialogs mode:

Excel 3.x-7.0
Lotus 1-2-3 2.x-4.0
RFT-DCA
Word 2.x for Windows
Works 3.x for Windows
Word 4.x-5.x for the Macintosh
Works 4.0 for Windows
WordPerfect 5.x
WordPerfect 6.x
WordStar

Although the Excel 3.x-7.0 converter supports NoDialogs mode, it is not used by Find Fast, given Find Fast includes its own internal Excel 2.1-7.0 filter.

The Word 2.x for Windows and Word 4.x-5.x for the Macintosh converters are not used by Word 7.0 for Windows for import because Word translates these older file types internally. These converters are, however, capable of import and are registered as import converters for and used by Find Fast for this purpose.
Word identifies some third-party file formats by scanning the first few bytes in their file headers. Find Fast does not support this method of recognition for third-party formats. Therefore, a foreign document that does not include a file extension recognized by any of the installed NoDialogs external converters will be processed by Find Fast's internal text filter.

For example, a WordPerfect 5.x document with no file extension will be processed by Find Fast's text converter because the WordPerfect 6.x import converter expects either a .DOC or .WPD file extension.
No converters other than those converters installed with Word or Office 95 that support NoDialogs mode will be utilized by Find Fast. This means that no 16-bit Windows 3.x converters or converters included with Word 6.0 for Windows NT or its associated Supplemental Conversions Disk will be used, unless the NoDialogs sub key is added. However, if this sub key is added to a converter that does not support the no dialogs flag, the result may be dialogs that come up during indexing, even if Find Fast is running in the background without an application window. This can cause erroneous behavior in both Find Fast and the converter.
For additional information on which converters are installed with each Word and Office Setup type and will therefore be available to Find Fast, please see the following article in the Microsoft Knowledge Base:

135782 WD: Converters Installed with Each Setup Type in Word/Office 95