Skip to main content

CDSL Newsletter — July 2026

Here is the July 2026 edition of the Cologne Digital Sanskrit Dictionaries newsletter, covering the activity of the past month. Significant work this period: a broad preface digitisation sweep, a new dictionary added to the online search, security hardening across all the server-side code, and the first release of the new structured API for Monier-Williams.

Prefaces digitised for sixteen more dictionaries​

A systematic effort to OCR and transcribe the front matter of every CDSL dictionary has been under way throughout June. Sixteen dictionary repositories now have a prefaces/ folder with the scanned preface text, along with English and Russian translations where the original is in German or French:

Böhtlingk–Roth (PW, 36 pages), Macdonell (MD), Böhtlingk kürzere Fassung (PWK), Cappeller (CCS), Stchoupak Sanscrit-Français (STC), Benfey (BEN), Schmidt (SCH), Shabda-Sagara (SHS), Vācaspatyam (VCP), Wilson (WIL), Monier-Williams 1872 (MW72, complete front matter), Indian Epigraphical Glossary (INM), Vedic Index (VEI), Böhtlingk Indische Sprüche (BOP), and Grassmann / Macdonell (already included in the June relaunch).

The goal is to make the scholarly apparatus of each dictionary — the editor's introduction, scope note, abbreviation conventions, and acknowledgements — accessible online for the first time. A summary page will be added to the guides site as the sweep completes.

New dictionary: NMMB (Nāmamālā)​

The Nāmamālā dictionary (NMMB) has been integrated into the online search infrastructure: the XML generation pipeline, the web display, the offline app, and the headword normalisation layers. It will appear in the search interface once the data preparation passes validation. This is the first dictionary addition since the original batch.

Security hardening across all server-side endpoints​

A Semgrep SAST (static analysis) scan was run across every PHP and Python codebase in the organisation. All reflected-XSS vulnerabilities in JSONP callback parameters have been patched (csl-apidev, csl-websanlexicon, mw-dev, GRA, CORRECTIONS). Database queries in the correction submission handler and the main search API have been converted to prepared statements. Dependabot is now configured on every active repository and pull requests are auto-merged when they pass CI.

These patches are live on the CDSL web server. No user data was at risk (the site does not have accounts or user-submitted content), but the input validation gaps have been closed.

Salt API Phase 1: structured endpoint for Monier-Williams​

A new api1/ layer is now deployed in csl-apidev, offering three structured endpoints for MW lookups: api1/entries (full XML entry by key), api1/ids (headword-to-internal-ID mapping), and api1/graphql (prototype structured query). This is the foundation for the forthcoming semantic linking work (TEI/OntoLex export, GRA coordinate citations, and the Benfey English sense extractor). The csl-standards repository reached v1.0.0 as the validation workbench for these export formats.

Developer documentation will follow in the guides site.

Text corrections: 43 batches closed in June​

The daily correction workflow continued through the month. Forty-three correction issues were closed in June across MW, SHS, SKD, INM, SCH, LRV, WIL, and AE — sourced from proofreading against the original print editions. The cologne-stardict offline dictionaries were regenerated automatically after each batch (82 commits, tracking every correction commit in csl-orig).

To report an error or suggest a correction, open an issue at github.com/sanskrit-lexicon/CORRECTIONS.

sanskrit-util: shared transcoding library launched​

A new repository, sanskrit-util, provides the canonical IAST/SLP1/Devanāgarī transcoding and normalisation functions as a dual Python + JavaScript library. The two implementations are behaviour-identical and share a cross-language test suite. A browser global build (window.SanskritUtil) is available for direct use in web pages without a bundler. This consolidates the transcoding code that was previously duplicated across several repos.

To receive future editions by email, subscribe here.