What's New in ExpressSpellChecker v2
Table Of Contents
New Hunspell Format Attributes
The following format attributes have been added to the Hunspell-based dictionaries:
- Compound words and options for compounding.
- Twofold suffix stripping.
- Extended affix classes.
- Homonyms.
- Prefix-suffix dependencies.
- Circumfixes.
These attributes extend the range of dictionaries that can be seamlessly used with the new version of ExpressSpellChecker.
To use a Hunspell dictionary in the application, in the Object Inspector add a new dictionary to the spell checker component. For the newly created dictionary, specify the Hunspell type and paths to the dictionary and affix files.
The OpenOffice dictionary uses the Hunspell format. This dictionary is retained for backward compatibility with v1. It’s recommended to use the Hunspell dictionary instead.
Complex Language Dictionary Performance
In v2, the following development was made, to improve performance of complex language dictionaries:
- Memory use for complex language dictionaries has been reduced significantly.
- A new load algorithm specific to complex language dictionaries was developed. As a result, load time has been reduced substantially.
The table below compares load times and memory use with v1 on our reference test machine (using Intel(R) Pentium(R) 4 CPU 3.20GHz).
|
Dictionary
|
Load Time, sec.
|
Memory Use, MB
|
| v2 | v1 | v2 | v1 |
|
Czech
|
0.75
|
32.92
|
31.46
|
181.50
|
|
English
|
0.16
|
0.75
|
7.39
|
5.78
|
|
French
|
0.31
|
0.39
|
9.46
|
2.45
|
|
German
|
0.45
|
5.69
|
15.12
|
32.40
|
|
Italian
|
0.23
|
198.63
|
10.69
|
450.73
|
|
Russian
|
0.34
|
6.39
|
14.51
|
52.99
|
|
Spanish
|
0.22
|
3.95
|
9.19
|
27.23
|
|