| Version 5.3 |
This version adds some more data structures and efficient storage technologies for improving the 'Indexing' performance, making index database file more compact and saving more disk space. In addition, we've made the 'Web-style Hyperlinks' availabe within RTF notes; See below for details.
1. The newly improved Index feature changes the storage structure for better performance and size optimization. It saves indexes in a separate file (.dbindex) within the same folder where the .nyf database resides. Old index database files created by previous versions get obsolete from this version on. Therefore, if you don't want to open your .nyf databases and use the 'Find as-you-type' feature within previous versions any more, we suggest that you purge the old indexes and release occupied disk space by selecting 'File -> Maintenance -> Purge Indexes ...' and 'Empty Recycle Bin ...' menu items. Note that the 'Empty Recycle Bin ...' utility also wipes off all the recycled items in your .nyf databases.
2. Unlike previous versions, this version doesn't insert .dbindex files into .nyf databases, they're saved as '.dbindex' files within the same folder where the .nyf files reside. Therefore, building indexes doesn't bloat .nyf databases, just small SSG-path-IDs map info is inserted. Consequently, you may want to carry the .dbindex file as well along with your .nyf file while you carry a .nyf database, for example, on trip.
3. The more effiecent Memory Cache implemented, so building indexes within this version is much faster than ever before, it won't take too much time, in most of cases, just one or few minutes are needed. As we tested on an average PC with PM1.6G/256MB, indexing 55,743 tokens of 4,833 items took 124.0 seconds (including GUI responses), 25.66 miliseconds per item and 2.22 miliseconds per token in average, the generated .dbindex file is 5.6MB large in size; Note that the test results actually depend on several factors, for example, the PC's overall performance and the size of memory Cache used.
4. The size of memory Cache is adjustable within this version. If you have enough memory available in your system, you may set a larger Cache Size for better indexing proformance. Generally, only 10MB memory Cache would significantly boost the indexing process, so in most of cases for most of .nyf databases, the indexing utility would be running at maximal speed. Adjusting the Cache Size is easy, simply enter the number of Cache entries or drag a slider-bar by using the mouse.
5. Now you have the convenience of defining a list of file types and/or words to be excluded during indexing. For example, almost all documents may contain these words 'a; an; the; of; on; in...', these can be excluded from within an index database, to save disk space and increase performance.
6. The 'Asia Language Support' option added into the new Index engine, as per user requests. With this option selected, all DBCS characters in text notes and specific documents will also be indexed, so the 'find as-you-type' feature now works with certain Asia Languages (e.g. Chinese). Nevertherless it may slightly slow down the indexing speed and require some more disk space, and note that the current tokenizer for Asia Languages is still in its earlier stage, it provides basic functionality for tokenizing and indexing DBCS characters. In addition, The boolean search engine always parses the find phrase as a boolean expression on the 'Find As-you-type' feature, this may causes some mismatches in find results in the case of DBCS, and the find results list may be somewhat longer than you expected. In this case, the Asia Language Support' option is better than nothing. If this option does not help, we suggest that you try the highly customizable 'Advanced Search' utility instead.
7. The new Index engine provides the interface for external tokenizers and file analyzers to get integrated.
8. The "Web-style Hyperlinks" feature is now available in the RTF text editor. It works much more gracefully than the old 'file://' based text links. You can now create web-style hyperlinks to either internal info items, or Internet Resources, and or local files. The web-style hyperlink in RTF looks like this: Link Text. You can make selection into a hyperlink with ease, and the hyperlinks can also be edited like that in a webpage editor.
9. Improvements: The web-style hyperlinks feature also improves the effects of several links-related utilities such as 'Capture -> Insert Local File Hyperlink', 'Insert Local Folder Hyperlink' 'Insert Info Item Hyperlink', the 'Insert Internet Hyperlink' and 'Import Directory Tree'. All these produce web-style hyperlinks now, instead of the old 'file://' based text links.
10. Consistency: Moving info items that have been hyperlinked elsewhere automatically updates the related hyperlinks, while deleting the info items disables all relevant hyperlinks.
11. Improvement: double-clicking on an info item now expands/collapses the tree branch if the item is already open within a child MDI window, otherwise, the item opens.
12. The 'Edit -> Make Selection Lowercase/Uppercase' utilities added. It makes selected text into lowercase or uppercase in the RTF text editor.
13. The 'Paste Unformatted' inserted into the context popup menu associated with the RTF text editor.
14. The 'Capture -> Import disk directory' utility includes a workaround for those particular URLs containing DBCS characters, and the 'Insert File/Folder Link' utility now supports certain Asia Languages in the resulting hyperlinks.
15. This version automatically labels custom icons with the corresponding .bmp filenames when the .bmp files are inserted.
16. This version has solved the problem that input focus may lose when typing a phrase for the 'Find As-you-type' feature on the toolbar.
17. The compond file parser comes with the improved 'Indexing' feature in this version, and the 'Advanced Search' utilizes this file parser as well. Therefore, those familiar documents attached in .nyf databases, such as .doc/.xls/.html documents, can be parsed and searched very well.
18. The Found List now supports 'sorting'. Clicking on a column header on the list view sorts the find results by the target columns.