Evidence
Comparisons and evidence
OfficeMaker will not publish performance tables that were not measured. Comparison URLs exist only when there is a documented workflow, honest caveats, and a date. Pages without that evidence stay unpublished.
Published comparisons
OfficeMaker vs PptxGenJS / Python code-driven PowerPoint generation is published. It restates a 15 April 2026 workflow estimate with methodology and limitations. It is not a laboratory benchmark.
Not published yet
Comparisons against python-docx, openpyxl or Microsoft Copilot are not published here because this repository does not contain a dated, repeatable test pack for those products. When a test exists — input, environment, run count, success criteria, output characteristics, limitations and date — a page can be added under /compare/…
What a future evidence page must include
- Test input (what was generated)
- Environment (tools and models involved)
- Number of runs or a clear statement that the figure is a reasoned estimate
- Success criteria
- Output characteristics (native Office file versus script-plus-screenshots)
- Limitations and date tested
Questions
- Does OfficeMaker publish benchmarks against every Office library?
- No. Only comparisons with a documented workflow are published. python-docx, openpyxl and Microsoft Copilot comparison URLs are reserved until measured evidence exists.