[TanglyEagle7718]

Add dataset

Removed non-compliant documents

[jarulraj]

Thanks for contributing these test materials. Having representative course documents is useful, but we should not merge these PDFs until their redistribution rights are documented.

A few specific concerns:

• The two chapter extracts appear to come from Database System Concepts and together contain 100 PDF pages. The PR does not include permission or a license allowing us to redistribute these chapters.
• ComPRE.pdf (Complete Removal of Redundant Expressions, PLDI 1998) includes a first-page copyright notice permitting personal/classroom copying under conditions, while requiring additional permission for republication or posting on servers. A newer applicable reuse license may exist, but none is documented here.
• The two course-slide bundles and the USENIX paper also need source links and applicable redistribution terms. Availability on a course website or free access to a paper does not, by itself, establish permission to include a copy in a public repository. TokenSmith’s Apache-2.0 license cannot grant rights to third-party documents.

Could you replace these files with documents explicitly licensed for public redistribution, such as CC BY, CC0, or material with a verified public-domain status? Please add a tests/dataset/README.md or manifest documenting each replacement:

1. Document title and author/rightsholder.
2. A link to the authoritative original source.
3. The exact license name and version, plus a link to the license and the source’s statement applying it to that document.
4. Required attribution and notices, and any changes made to the original.

Please retain any required notices in the documents and confirm that the terms allow inclusion in this public repository. For useful materials without redistribution permission, we can link to the original source and let users supply their own local copies instead of bundling the PDFs.

Educational use alone does not automatically establish fair use; the U.S. Copyright Office explains that it depends on the circumstances and four statutory factors. Using clearly licensed test documents will make this dataset easier for everyone to reuse.

Thanks again for helping expand the test coverage and for addressing the provenance and licensing details.