ScholarXIV Journal
An open preprint server for every field. Post your work the day it's ready — no fees, no waiting for review.
Today
2 papers · November 3A Quality-Constrained Accounting Model for Localizing a Next.js Application into Ten Languages: A Critical Review
Mohammed Ibrahim, Linguana Team
Translation usage charges alone cannot determine the cost of releasing and maintaining a localized application. This paper reviews selected localization and post-editing evidence and derives an analytical comparison of hand-managed resources, API-based workflows, and localization platforms. Eight core sources and two accounting frameworks inform the distinction between translation productivity, application integration, and acceptable release quality. The resulting model includes common internationalization work, workflow-specific setup, recurring language-dependent labor, and external charges. It derives a break-even condition for recurring releases and shows algebraically why a reduction in generation price can be outweighed by additional review effort. The contribution is a quality-constrained accounting model, not a measured Next.js case study. No vendor prices, observed engineering hours, original cost savings, or Linguana performance figures are reported. The analysis identifies the conditions under which an economic comparison is interpretable and the measurements needed to instantiate it.
Translation Correctness in React Applications: A Standards-Grounded Failure Taxonomy and Evaluation Model
Mohammed Ibrahim, Linguana Team
Translation quality in a web application includes more than the linguistic quality of resource text. Messages interact with parsers, runtime values, locale-dependent selection, document direction, and constrained components. This paper synthesizes ten core research and authoritative technical sources plus a versioned MQM specification to derive a layered account of translation correctness for React applications. It distinguishes resource validity, message semantics, rendering behavior, and task meaning, then organizes failures by both observable symptom and attributable cause. The analytical result is a release-readiness model in which linguistic adequacy cannot compensate for broken technical contracts, and technical validity cannot establish correct meaning. The paper also identifies denominator and attribution choices that affect reported failure rates. The taxonomy is proposed for application and validation; no React translation benchmark, prevalence estimate, or product comparison is reported.
Preprints are not peer reviewed. They are posted as the authors wrote them and may report findings that have not been checked by anyone else.