The digital landscape for the Rohingya language has long faced a unique structural obstacle stemming from its dual active writing systems. While speakers of the language are often fluent in either Hanifi Rohingya—a distinct right-to-left script officially integrated into Unicode in 2018 within the U+10D00–U+10D3F block—or Rohingyalish, the adaptation of the language into the Latin alphabet, transitioning between the two has traditionally required cumbersome manual intervention. Many community members read Rohingyalish comfortably while struggling with Hanifi, and vice versa. To mitigate this accessibility barrier, developers at RohingyaLanguage.org have engineered Rohingya Reader, a novel Chrome browser extension designed to perform real-time, in-place transliteration of Hanifi text directly on active web pages.
Unlike traditional translation tools that alter the underlying semantics of a message, Rohingya Reader executes strict transliteration, converting the writing system while preserving the original Rohingya terminology. Although the extension awaits final review on the Chrome Web Store, its foundational architecture builds directly upon an existing online script converter hosted by the organization, shifting the paradigm from static webpage copying to dynamic, automated text transformation.
Background Context and Linguistic Evolution
The necessity for such a tool arises from the complex sociolinguistic history of the Rohingya people, an ethnic minority primarily originating from Myanmar’s Rakhine State. Decades of displacement, systemic marginalization, and migration have scattered the Rohingya diaspora globally, leading to fractured educational opportunities and divergent literacy practices. Within diaspora communities in countries like Malaysia, Bangladesh, Saudi Arabia, and various Western nations, younger generations and those educated in Western or secular systems frequently find the Latin-based Rohingyalish much easier to navigate. Conversely, traditional scholars, religious educators, and older generations retain a strong affinity for Hanifi Rohingya, the native script developed in the mid-20th century to provide the language with a dedicated typographic identity.
This bifurcation has created a digital divide within the community itself. Websites, community forums, and social media feeds featuring Rohingya content often utilize one script exclusively, alienating readers literate only in the alternative format. While digital rights advocates and language preservationists have increasingly published content online, the friction of manually copying text into external web converters has hindered seamless consumption, prompting the development of automated browser-level solutions.
Engineering Challenges in Real-Time Transliteration
Developing a browser extension capable of modifying live Document Object Model (DOM) trees without degrading user experience or breaking website functionality required overcoming significant technical hurdles. The engineering team detailed several critical obstacles encountered during development and the methodologies deployed to resolve them.
Detecting Hanifi Script via Code Points
A primary challenge in automated script conversion is accurate text identification. Web pages rarely incorporate proper language metadata, and variations in user-installed fonts further complicate text parsing. Relying on raw HTML lang attributes proved entirely unreliable. Consequently, the extension bypasses metadata reliance, identifying Hanifi text strictly through Unicode code point detection utilizing regular expressions mapped directly to the U+10D00–U+10D3F block.
To maintain efficiency, text nodes are dynamically split into maximal runs of Hanifi code points. Only these specific sequences are routed to the conversion algorithm, ensuring that embedded English, Arabic, Bengali scripts, punctuation marks, and emojis remain entirely untouched. Extensive property testing involving thousands of randomized, mixed-string datasets confirmed that the extension’s output matches the established baseline of the organization’s primary web-based converter.
DOM Integrity and Performance Optimization
A common pitfall in browser extension development is the reckless manipulation of the DOM, particularly through the rewriting of innerHTML. Such practices routinely destroy active event listeners, destabilize framework states, and disrupt element identity. Rohingya Reader strictly limits its modifications to the data property of existing text nodes, completely avoiding the creation, removal, or structural relocation of HTML elements.
To prevent infinite observation loops—where an extension reacts to its own modifications—the tool records both the original source text and the precise output written to the node. Furthermore, memory management is tightly controlled using WeakMap collections and sets of WeakRef references, ensuring that elements removed dynamically by host pages remain fully eligible for garbage collection without causing memory leaks.
Handling Dynamic Content and Rendering Jank
Modern web architectures heavily rely on infinite scrolling, asynchronous comment sections, and dynamic feeds, rendering a single, static page conversion inadequate. Rohingya Reader employs a MutationObserver instance to monitor DOM updates, queuing exclusively newly added or modified nodes rather than rescanning entire documents.
By executing computational workloads in time-sliced increments of approximately 12 milliseconds, the extension prevents the main browser thread from locking up, eliminating visual stutters or "jank." In performance benchmarks simulating high-density content delivery, a webpage containing 5,000 distinct paragraphs underwent complete script conversion in roughly 0.4 seconds, registering zero long tasks. Certain sensitive regions—including code blocks, text inputs, textareas, contenteditable zones, and specialized editors such as CodeMirror, Monaco, and ProseMirror—are systematically excluded to protect user inputs and active cursor interactions.
Typography, Bidirectional Layouts, and Text Fragmentation
Linguistic nuances presented further complications, particularly regarding Hanifi’s reliance on combining diacritics for tone and length. Web publishing platforms frequently fragment words across multiple HTML elements, wrapping specific characters or tone marks in isolated <span> tags. Because the transliteration engine relies on contextual character lookaheads, processing isolated text nodes independently produces inaccurate outputs.
To resolve this, adjacent inline text nodes are evaluated and converted collectively. If a safe morphological boundary cannot be verified, the text defaults to its original Hanifi script rather than risking mistranslation.
Additionally, the integration of Hanifi (right-to-left) with Rohingyalish (left-to-right) introduced severe layout challenges governed by the Unicode bidirectional (bidi) algorithm. Standard text insertion frequently triggers visual reordering anomalies around punctuation and hyperlinks. Rather than modifying layout attributes via dir parameters, the extension embeds invisible Unicode isolate characters—specifically Left-to-Right Isolates (U+2066) and Pop Directional Isolates (U+2069)—directly into the text nodes. This ensures word order remains structurally intact across complex elements like hyperlinks without disrupting overarching paragraph alignment.
Security, Permissions, and Dataset Integration
Balancing functional capability with user trust dictated a minimalist permissions model. The extension requests foundational permissions including activeTab, scripting, and storage, accompanied by optional host permissions granted strictly on a per-site basis. A dedicated background service worker continuously synchronizes saved user preferences, active permissions, and script registrations across browser restarts, updates, and permission alterations.
Beyond the browser extension, the underlying transliteration engine and comprehensive lexicon power an open-access dataset hosted on Hugging Face. The Rohingya Hanifi–Rohingyalish–English Lexicon, released under a Creative Commons Attribution 4.0 license, provides researchers, linguists, and developers with standardized resources for low-resource natural language processing, search normalization, and computational linguistics experiments.
Broader Implications and Future Outlook
The deployment of tools like Rohingya Reader highlights an innovative approach to digital inclusivity for marginalized linguistic groups navigating multi-script environments. By automating the technical friction of script conversion without compromising web performance or data integrity, the project bridges a persistent divide within the global Rohingya diaspora.
As the extension transitions from its review phase toward public availability on the Chrome Web Store, developers are actively soliciting feedback from bilingual community members and website administrators to refine rule-based approximation models. Through continued refinement of lexical round-trips and community-driven reviews, technological infrastructure continues to adapt to the preservation and digital flourishing of the Rohingya language.
