We use two text boxes for the convenience of users. In the first box, the program shows some possible correct spellings for words marked as incorrect. During this suggestion-generation process, the spell checker has to move back and forth through multiple loops containing many logic rules (sets of conditions and steps) and arguments (values passed through parameters), while comparing the results each time with billions of words in the database. This makes the overall process slower.

For this reason, we also need to limit the number of words that can be checked at one time. To address this, we have introduced a second text box as an alternative. In this box, spell checking is performed much faster because suggestion generation part is skipped here.

Users who want to check spelling quickly—especially for larger amounts of text and simply to spot misspellings—should use the second text box.

If common words such as আমি (in Assamese) are marked in red, it may be because the program has been opened in the browser for the first time and is still loading the necessary data. In such cases, you can refresh the browser or close the tab and reopen the site.

Another possible reason may be a slow internet connection. A slower connection can affect the program’s performance, since the spell checker has to check billions of words in just a fraction of a second.

For technical reason.
If common words containing the special Assamese–Bengali characters য়, ড়, ঢ়, such as দয়াময়, গৌড়ীয়, পঢ়ুৱৈয়ে (in Assamese), are marked in red, it may be due to a technical reason. The Unicode system allows two different ways of writing each of these characters. One method uses a single glyph in which the bottom dot is inseparable from the base character. In the other method the characters are written as combination of two glyphs, where the dot is added separately to the base letter. This difference creates difficulties when processing words efficiently. Our program uses the single (inseparable) glyph form. However, if a word is typed using the alternative form (base character + dot), the program may fail to recognize it as the same character, eventually marking the word as an error. Unfortunately, Unicode system has left many issues in Assamese-Bengali script including not recognizing the scripts appropriately.

We are quite aware of several such efforts—both in India and abroad—that aimed at developing a high-quality and real-time spell checker for Indian languages. However, none of these efforts have yet met with considerable success. The difficulties are both technological and linguistic, with the linguistic challenges being particularly significant. Indian languages are highly inflected. Even a single root word can produce thousands of inflected forms through the addition of various inflectional morphemes—for example:

• root + case endings,
• verbal root + verbal inflections,
• prefixes + root,
• prefixes + root + case endings or verbal inflections,
• root + suffixes,
• root + suffixes + case endings,
• root + midle suffixes + suffixes,
• root + midle suffixes + suffixes + case endings,
• suffixes + case endings

As expected, we use morphological and orthographical parsing algorithms to analyze and generate these forms. However, the methods of generating inflected words are so diverse and the number of possible forms so large that the algorithms alone often cannot help effectively.

For this possible reason, even big technology companies like Microsoft and Google have yet to come up with a complete full spell check functionality for Indian languages. Of course, we do not claim that ours is the first perfect or complete spell checker (for any Indian language). Instead, we believe it is probably the first real-time spell-checking program (for any Indian language) to achieve considerable success in both technological and linguistic terms.

You cannot directly check non-Unicode text. But you can use the Text Converter to convert them to Unicode text and then paste them in to check the spellings.

You can contact us at eduserservices@gmail.com or use the YouTube link given below and leave a comment in the comment section.