The article analyzes an alternative hybrid methodology, Scrumban, in the context of modern IT project management. It examines the development of the agile Scrum approach and the visual Kanban method, together with their mutual influence. It reflects on the evolution of agile approaches to the formation of Scrumban as a hybrid methodology. An analysis of the advantages and disadvantages of Scrumban is provided.
The results of a study examining the pragmatic and creative aspects of onymization in multimodal discourse are presented, using titles of English-language drama films. Two online databases, Kinopoisk and the Internet Movie Database, were used as sources. Criteria for selecting film titles and an algorithm for analyzing their formation using an integrated methodology were developed. The resulting multimedia corpus of research material includes 193 titles of contemporary American and British drama films and over 420 of their multimodal representations in film posters and trailers. As a result of the study, primary structural-syntactic types of titles for the films under analysis have been established, and their lexical and semantic features have been identified and described. The conceptual and semantic transformations that appellatives undergo during their transition into film titles are systematic, serving primarily to capture the audience’s attention and communicate narrative content. It is demonstrated that onymization within multimodal discourse operates as a dynamic, creative process unfolding across three interconnected dimensions: linguistic, verbal–visual (exemplified in cinematographic posters), and verbal–audiovisual (manifested in trailers).
The paper represents the results of a comparative study on Internet users’ reviews on “The Snow Queen” fairy tale (translated into Russian in 1899), the cartoon “The Snow Queen” (produced in 2012 in the Russian language), and the movie The Secret of the Snow Queen (produced in 2015 in the Russian language). We analyzed 184 written reviews (three datasets) in Russian language published by users of the platforms Kinopoisk, Kino-Teatr, Otzovik, Litres and LiveLib by computational linguistics methods. In each dataset we analyzed words frequencies and topic-forming collocations, described trends in topics’ modeling. It was revealed that reviews of the fairy tale are distinguished by a description of the storyline and the subjective experience of its perception by recipients, while reviews of the cartoon and film mostly characterize the genre of the work, its creators and actors.
This paper investigates the functional and stylistic distribution of the colon (le deux-points) in French and Russian literary and authorial nonfiction corpora, evaluating how cross-lingual punctual asymmetries expose the cognitive boundaries of contemporary Large Language Models. Utilizing parallel subcorpora from the Frantext and the Russian National Corpus spanning over 7 million tokens and 21 000 parallel contexts, this study uncovers a profound discrepancy between human translation practices and algorithmic text generation. While monolingual frequencies indicate a higher baseline productivity of the colon in Russian (68 per 10 000 words) than in French (51 per 10 000 words), translational dynamics demonstrate an elastic, human-mediated restructuring: French-to-Russian translation yields a density increase from 55 to 71 colons per 10 000 words, whereas Russian-to-French translation exhibits a contraction from 97 to 58. By analyzing functional symmetry in speech introduction alongside deep asymmetries in asyndetic causal, explanatory, and “headlinese” constructions, it is argued that the colon acts as an empathetic marker of implicit prosody and pragmatic rhythm. Ultimately, these subtle, nonexplicit discourse relations represent a “blind spot” for Transformer-based architectures that rely on maximum-likelihood estimation and systematically regress toward explicit, high-frequency syntactic connectives.
The rapid growth of multilingual audiovisual business content (cinema, podcasts, TED Talks) necessitates efficient localization, yet the translation of culture-specific business idioms remains a critical challenge for large language models (LLMs). The paper experimentally tests four prompting strategies (zero-shot, few-shot, chain-of-thought (CoT), and contextual) across three LLMs (ChatGPT-4, Gemini 1.5 Pro, DeepSeek) to improve the adequacy of English-Russian translation of business idioms in audiovisual translation (AVT). A corpus of 100 business idioms extracted from “The Wolf of Wall Street”, “The Gentlemen”, business podcasts, and TED Talks was evaluated using an adapted LISA scale (1–5) for semantic adequacy, stylistic adaptation, and AVT constraints. The results show a clear positive dynamic: zero-shot is unsuitable (mean score 2.8), few-shot provides a significant quality leap (+1.2 points, mean 3.7), CoT delivers the highest linguistic accuracy (mean 4.1), and contextual prompting achieves the best overall performance (mean 4.3), with 78
