MSIX at Scale: An Automated Packaging Study of 3,261 Windows Application Cases

Abstract

Background: MSIX is Microsoft’s package format for installing and managing Windows applications. This study evaluated whether an automated workflow could construct MSIX packages and perform initial validation across a large and heterogeneous set of existing Windows applications.

Methods: We analysed records from June 2026 for 3,261 application cases in which MSIX packaging was attempted using EtherApps Forge. Applications were identified through searches of the public Windows Package Manager catalogue. Each case was counted once, irrespective of outcome. Recorded smoke-test passes were verified against their associated saved test reports. Smoke testing assessed installation, start-up and cleanup, but did not evaluate application functionality comprehensively.

Results: Of 3,261 cases, 2,957 (90.68%) produced an MSIX package build, 2,919 (89.51%) completed installation and registration, 2,638 (80.90%) launched a selected program within the package, and 2,610 (80.04%) completed the defined smoke test successfully. Successful cases spanned eight installer and distribution types. All 2,610 reports associated with recorded smoke-test passes were consistent with the test fields examined during verification. Six pairs of cases shared identical reports, and two cases classified as passing launched a deactivation utility rather than a verified main application.

Conclusions: In this dataset, automated MSIX packaging and initial validation were completed successfully for a substantial proportion of attempted application cases. The findings demonstrate the feasibility of using EtherApps Forge to construct and perform initial checks on MSIX packages at scale and support further application-specific evaluation of MSIX as a packaging option. These results should not be interpreted as a general MSIX compatibility rate or evidence of production readiness. Further work should include application-function testing, more complete environment records and validation in representative deployment environments.

Research status: This company-authored preprint (Version 1) has not been peer reviewed. An accompanying analytical supplement provides sanitised case-level data, methodological notes and scripts for reproducing the reported calculations, but does not contain all materials required to repeat the original application tests.