Abstract:

Poem writing is not an easy job. It requires knowledge of poem writing elements (PWE) like linguistic knowledge, creativity, selection of vocabulary, expression of thoughts, level of content complexity, the art of expressing the feelings, poem architecture and other linguistics features. Poets wisely choose and tune these PWE's keeping in mind the target audience while designing their poems.

Learning poem writing styles like the sonnet, sestina haiku, villanelle, ode and others took years of practice. Few online English poem authoring tools are available, which helps native English poets to design their poems. However, no such poem authoring systems are present for native Indian languages. This research is the answer to this problem.

This research presents Hindi (Hindi is the national language of India) poem authoring system named "KAVYA-SAMVAD", which helps native Hindi students to learn the art of poetry. We classified Hindi poems under four categories based on their complexity level concerning vocabulary, linguistic features and writing styles. The poem classification scheme, along with the working of the "KAVYA-SAMVAD" system and its reference architecture, is discussed in detail in this research paper. This research further discusses issues and problems associated with the creation of the Hindi poem authoring system. Then the samples of the Hindi poems generated using the "KAVYA-SAMVAD" Hindi poem authoring system are presented along with the proper explanation. We believe that this system helps new Hindi poets to learn the art of poetry effectively. We further believe that software engineers and educators use the reference architecture of the "KAVYA-SAMVAD" system proposed in this research for designing similar poem authoring systems for other Indian and foreign languages.

Keywords:

Hindi Poem, Reference Architecture, Linguistics, Poem Architecture.