About the Toolbox

Toolbox is the companion to Core Readings. Core Readings curates what to read and discuss; Toolbox catalogs what to consult and use — the descriptive and encoding standards, identifiers, classification and vocabulary systems, linked-data infrastructure, software, repository platforms, and library systems that do the actual work of organizing information.

It is the 2026 renewal of the 2001 Knowledge Structures Toolbox (University of Arizona), preserved in dLIST. It is selective, not comprehensive, and superseded systems are kept, dated and labeled, rather than erased.

Scope

The Toolbox includes resources people use to organize, describe, and discover information, and the communities through which those resources are developed, maintained, and applied. Software, standards, services, vocabularies, and professional groups are included according to their practical contribution, not simply their format. Experimental resources may be included when they make an important development visible and their limitations are clearly stated.

Placement rule for groups and communities

Groups are included according to what they contribute, not as a category of their own. A group earns inclusion through an identifiable contribution: specifications, implementation support, maintained vocabularies, shared infrastructure, sustained practitioner assistance, or historical stewardship of foundational knowledge-organization work.

This avoids treating every group as a standards organization, or duplicating a group’s website across several sections. Journals follow the same rule: they belong in Core Readings, not here.

A task index (“Start with a task”), fifteen sections, and a short list of professional organizations and lists.

Each entry carries three labels, shown as rounded pills:

CategoryStatusAccess & licensing

• Category (.tag): what kind of resource it is, sometimes followed by a qualifier. The first term comes from a fixed set — Standard, Conceptual model, Vocabulary, Classification scheme, Authority file, Identifier, Software, Service, Organization, Project, Community resource, Teaching resource — and the qualifier after the dot comes from a short list, used only where it separates entries within a type. Standard: content rules, encoding, linked data, metadata, protocol, e-learning, vocabulary construction. Vocabulary: subject headings, thesaurus, metadata, place names. Identifier: scheme, registry. Software: library system, repository, discovery, digital collections, metadata tool, metadata management, archival description, course-reader builder, AI feature. Service: cataloging, catalog, hosting, knowledge base, discovery, scholarly metadata, web tool. Standard also takes e-learning packaging. Conceptual model: ontology. The other types take no qualifier.
• Identifier and link: every entry and listed community has a stable slug (id), which is also its anchor on the page, so …/toolbox/#mods links straight to MODS. Ids are never reused; a withdrawn entry retires its id.
• Cross-references (see_also): the ids of related entries, shown on the page as a “See also” line.
• Task routes (tasks): the “Start with a task” routes an entry appears in. The routes themselves, with their order and step notes, are defined once in the page’s task index and in toolbox.json; an entry’s tasks value is derived from them.
• Status (.status): the resource’s current role in practice. Active: currently maintained, supported, or established for operational use. Emerging: still developing toward routine use, or its adoption in the practice this Toolbox covers is limited or unsettled; it may already be operational, and the annotation states whether the reason is development, deployment, or adoption. A new release, version transition, or AI feature does not by itself make an established resource Emerging. Legacy: superseded or no longer substantially developed, but still used or materially present in operational workflows. Historical: retained principally for historical understanding rather than as a current operational option. Old, being replaced, and discontinued are not synonyms; a discontinuation is a fact stated in the annotation.
• Access & licensing (.acc): describes the resource named in the entry, not automatically its provider, documentation, underlying software, or associated services. Free access does not necessarily imply permission to modify or redistribute. The values are Open source (software with a verified open-source license), Open data / content (stated reuse permissions), Free to access (reading or use is free; broader reuse rights are not implied), Freeware (free-to-use proprietary software), Paid / subscription (payment is required for the resource described), Source-available (the code can be inspected, but licensing restricts use or redistribution), Mixed (materially different free and paid components or access routes, which the annotation identifies; several licenses, or a license not yet confirmed, do not by themselves make a resource Mixed), Restricted (use is limited to an eligible community, institution, or approved participants), and Not applicable (an organization, program, concept, or unreleased tool for which an access label would mislead). A resource is labeled Open data / content only where the page records stated reuse permissions; otherwise a free resource is Free to access.

These labels are dated editorial assessments, not permanent properties. When one set of labels would conceal an important difference between resources, those resources receive separate entries or an explicit qualification, and the page splits entries on that basis (for example Wikidata and Wikibase, LRMoo and FRBRoo, Omeka.net and the installable Omeka software).

The access label is explicit on purpose

Access is stated outright rather than assumed, including where it sits uncomfortably beside a “critical and open organization of information” framing. RDA Toolkit, WebDewey, and Classification Web carry paid access because the field’s own required standards genuinely sit behind paywalls. Every entry carries an access label; none uses a dash.

Section VIII lists AI-assisted cataloging features from OCLC (the AI Enrichment macro in the Connexion client) and Ex Libris (the Alma AI Metadata Assistant and the Alma Specto Digital Metadata Assistant), next to Alma’s rule-based normalization rules. For these entries the vendor’s own documentation is the primary source for how the workflow and its safeguards are described. Documentation is not evidence of how accurate or unbiased the output is; that is a question for the reader, the library, and the field.

One working distinction runs through Section IV and Section VIII: WebDewey and Classification Web support human classification decisions, while the AI features propose metadata for a human to evaluate.

Besides this README, the Toolbox is published in several forms. The teaching note explains what each one is for.

• index.html: the authoritative edition, with a schema.org JSON-LD graph (a CollectionPage, the maintainer as a Person, and an ItemList of every entry and listed community, and an ItemList of the task routes). A “Last updated” line in the header is the human-readable version anchor.
• toolbox.csv: one row per entry and per listed community, in page order, with the columns id (a stable slug, also the entry’s page anchor), section, subsection, entry, category, status, access_licensing, annotation, url, additional_urls, tasks, and see_also (lists separated by “ | ”). The professional organizations and lists carry no labels, since they are not badged on the page. The file is UTF-8 and every field is quoted. Spreadsheet programs sometimes turn text such as “MARC 21” into a date; to avoid that, import the file (in Excel, Data > From Text/CSV) and set the columns to Text. The id column is the stable key to use when comparing versions.
• toolbox.json: the same entries as structured data, with the two label vocabularies, the scope statement, and the task routes (each with its order and step notes), for scripts and programs.
• toolbox-sample-dctap.csv: a sample DCTAP application profile for a digital collection item, to open in a spreadsheet and adapt. Checked with dctap-python, whose only warning is that the shape identifier item is not a URI — a useful first fix for students.
• ai-linked-data.html: a teaching note with class activities on one developer’s metadata lifecycle (Eunomia’s Labors, Crossref, Pinakes, Aletheia) and on knowledge graphs as research notes in other disciplines.
• toolbox.md: a plain-text copy for pasting into syllabi, course sites, notes, or AI tools.
• CITATION.cff: citation metadata for tools that read it. Like Core Readings, the Toolbox keeps its own file in its own folder, so it describes this work alone. Codeberg does not show a “Cite this repository” button from it; readers on Codeberg see the citation line below.

All of these are generated from the page, so the page is always the source: when an entry changes, regenerate the files rather than editing them.

Toolbox shares the Infophilia design tokens with Core Readings and the Festschriften page exactly: Lora, Playfair Display, and Source Sans 3; the brass and teal palette; and CSS custom properties in :root. There are no new design decisions. The stylesheet was extracted from Core Readings.

• The Cataloging Lab entry was removed from Core Readings’ Section III and now lives only in Toolbox Section XV (Critical & reparative cataloging). Core Readings keeps its scholarly readings.
• Toolbox’s professional organizations list deliberately excludes ALA Core itself, BARTOC, the ISKO Encyclopedia, and the LC/PCC portal links. Those stay on Core Readings’ “Professional resources (selective)” section. The Toolbox does point to the Core interest groups and committees most relevant to cataloging and metadata, and explains where the former ALCTS groups went.
• Journals (for example LRTS, CCQ, and the Journal of Information and Knowledge) are reading venues and belong in Core Readings under the placement rule above.
• LibGuides was evaluated and excluded: its API and XML export is a proprietary CMS content model (guides, pages, boxes, subject tags) with no metadata-standards footprint — no Dublin Core, no OAI-PMH, no schema.org crosswalk — so it is outside this tool class, not merely a weak member of it.

Adding an entry touches these places in index.html:

1. The entry card, in the right section or subsection, using the article class="entry" id="…" pattern: a new, never-used id; category, status, and access & licensing labels (chosen from the vocabularies above); citation; annotation; a “See also” line if it has cross-references; link.
2. The task index, if the entry belongs on a route; keep routes to three to six entries.
3. The JSON-LD ItemList: add a ListItem, renumber position, and update numberOfItems.
4. The count in “On this edition” near the end of the page, the “Last updated” line in the header, and then regenerate toolbox.csv, toolbox.json, and toolbox.md.

The count follows one rule: annotated entries are the cards in the fifteen sections, and the professional organizations and lists are counted separately. A card names one resource, or a closely related group when the same labels accurately describe all its members; where the labels would conceal an important difference, resources receive separate entries or an explicit qualification. Annotations should describe what a resource is and does, drawn from the resource’s own documentation or another cited source. Do not invent claims or opinions. Section anchors: #tasks #descriptive #encoding #identifiers #classification #vocab #authority #semweb #cattools #display #repos #ils #discovery #protocols #orgs #critical #professional #provenance; subsection anchors: #metadata-management #archival-environments #preservation-chain #preservation-transfer #library-discovery #scholarly-discovery #vocab-to-validation; route anchors: #task-describe and so on. New material goes into subsections where it can, so section numbers and anchors stay put. If sections are inserted, reconcile the numbering and anchors together.

18 October 2026: task index and additions

A second pass the same day, still version 1.0, after which the Toolbox is frozen until the course ends.

• Start with a task. A task index of ten routes through existing entries: describe a resource; build or apply a vocabulary; reconcile names and identifiers; define a local metadata profile; validate and transform records; publish a collection; preserve and transfer files; provide discovery; document rights and provenance; and prepare a book’s metadata for discovery and reuse, whose theme is AI and linked data. That last route recommends no particular cataloging service and carries a one-line maintainer disclosure. Routes suggest starting points and useful connections, not mandatory workflows: some entries are alternatives, others work together. Rights and provenance is an unordered group; Publish a collection branches into a digital-collection path and a course-reader path; Provide discovery has an optional scholarly-publishing branch.
• Stable links. Every entry and listed community now has its own anchor, using the existing id slugs, and the data files gain subsection, tasks, and see_also fields.
• Scope and placement. A scope statement, a placement rule for groups, and an expanded contribution criterion (see What this is).
• Retitled: Section VII, Semantic web & knowledge graphs; Section VIII, Cataloging, metadata management & quality tools; Section XII, Discovery interfaces & services.
• New subsections: metadata management, quality & provenance (VIII); course readers & resource assembly and archival description environments (IX); preservation transfer & services (X); library discovery layers & connectors and scholarly discovery & exploration (XII).
• Added (22 entries): MODS, DCTAP, BagIt, and Common Cartridge (II); Eunomia’s Labors (III); MADS (VI); SHACL and LINCS (VII); OpenMetadata (VIII); Aletheia, ArchivesSpace, and AtoM (IX); DART and CLOCKSS (X); Yale Library AI Connector, OpenAlex, Scholia, Open Knowledge Maps, and Pinakes (XII); RightsStatements.org and SHACL Playground (XIV); Local Contexts (XV). DART was chosen over Bagger, whose repository says its future is under review. ResearchRabbit was considered and left out: its citation-network role is covered by Pinakes and Scholia, and its data is less inspectable.
• Professional list: added the Madras Library Association, the Sarada Ranganathan Endowment for Library Science, and the CILIP Metadata & Discovery Group; the ALA Core entry now explains where the former ALCTS groups went.
• Teaching aids: a “From vocabulary to validation” table (Section II), a preservation table (Section X), and toolbox-sample-dctap.csv.
• Category qualifiers added: Software · metadata management, Software · archival description, Software · course-reader builder, Service · discovery, Service · scholarly metadata, Standard · e-learning packaging.
• New access value, Restricted: use is limited to an eligible community, institution, or approved participants. The Yale Library AI Connector moves from Not applicable to Restricted: access does apply, and the restriction is the information.
• Label corrections: the M2R project moves from Mixed to Free to access (an unconfirmed code license is not a paid component); the Alma AI Metadata Assistant moves from Emerging to Active (Phase I describes scope, and a roadmap does not by itself make a resource Emerging); Pinakes’s Emerging status now rests on its documented development rather than its date.
• Grouping rule restated: the edition note had reversed the rule; an entry names one resource, or a closely related group only when the same labels accurately describe all its members.
• Teaching note added: AI and linked data, unleashed, with five class activities.
• Corrections: the schema.org entry now points to validation tools in Section XIV (not XIII); a repeated phrase was removed from the Z39.50 entry; two section comments in the page source were renumbered.

18 October 2026

• The edition is now version 1.0. toolbox.json, toolbox.md, and a CITATION.cff file were added beside toolbox.csv.
• Added what-changed.html, a short account of the audit-driven revision written for students, including what each file format is for.
• The footer’s Repository link now points to the Toolbox folder of the repository, where the file history shows each revision.
• Consolidated category qualifiers (see Entry model).

17 October 2026

Revision following an entry-by-entry audit of the earlier version.

• Labels defined. A “How to read the labels” note now defines Category, Status, and Access & licensing. The third label was renamed from “Access model.”
• Access vocabulary replaced with eight values (Open source, Open data / content, Free to access, Freeware, Paid / subscription, Source-available, Mixed, Not applicable). “Open Access” and “Commercial/Proprietary” were retired because each combined questions about cost, reuse rights, and business model. The dash used on five cards and the ad hoc “Free for small orgs / Commercial” label were retired as well.
• Categories normalized. The more than forty free-form category strings of the earlier version became twelve primary types plus a short, fixed list of qualifiers (35 distinct labels in all).
• Entries split where one set of labels could not describe every named resource: LRMoo and FRBRoo; Wikidata and Wikibase; Omeka Classic/S and Omeka.net; GeoNames and GEOnet Names Server; OCLC Connexion and WorldShare Record Manager; NOTIS & Dynix and Horizon; NISO Z39.19 and ISO 25964; LibraryThing, LibraryThing for Libraries, and TinyCat; ZineCat, the Zine Subject Thesaurus, and CollectiveAccess; VuFind and Blacklight; Creative Commons and License Chooser.
• Corrections and softened claims. The British Library cyberattack is dated October 2023. Horizon is no longer described as discontinued. EAD is described with EAD4 as the current version. WordNet is Legacy (no longer developed). MarcEdit is Free to use, not Open Source. Intentional Arrangement SKOS is Source-available, with its size-based license described as such. Claims that could not be confirmed against a current source were removed or reworded (CIP eligibility and funding, DSpace and EPrints release details, “dominant” and “superseded” market claims, and similar).
• Status changes and definitions. Status definitions were tightened: old, being replaced, and discontinued are no longer treated as synonyms. RDA, MARC 21, the OCLC AI Enrichment macro, and the Alma Specto Digital Metadata Assistant are Active; AACR2, FRBRoo, Voyager, and Aleph are Legacy; Colon Classification is Historical (retained for teaching, with no sourced evidence of continuing use). The Alma AI Metadata Assistant, BIBFRAME, the M2R project, Intentional Arrangement SKOS, and ErgoCat stay Emerging, and each annotation says whether the reason is development, deployment, or adoption.
• The standalone Knowledge Graph card became a see-also note on the Wikidata entry, keeping the WorldCat Entities link.
• Version text reconciled. The “Version 1 scope note” was removed and a counting rule for entries was added. Later the same day the edition was named version 1.0 (see 18 October).

16 October 2026

• Removed the top-of-page banner.
• Section VIII: added the OCLC AI Enrichment macro, the Alma AI Metadata Assistant, the Alma Specto Digital Metadata Assistant, and Alma normalization rules with Browse Shelf Listing.
• Section VI: added SNAC (Social Networks and Archival Contexts).
• Section I: added FRBRoo / LRMoo. Section II: added the MARC21 to LRM/RDA/RDF Mapping and Conversion Project (M2R).
• Section IV: added Colon Classification, BISAC, and OCLC Classify (listed as discontinued 31 January 2024); added the LC Classification Outline and DDC Summaries links to the LCC and DDC entries.
• Professional list: added ALA Core interest groups and committees, the LD4 community, the LD4 Wikidata Affinity Group, the Wikimedia and Libraries User Group, RDA Toolkit regions, the SAA Description Section, and SNACSchool.
• Reordered the JSON-LD list to follow the page (Intentional Arrangement SKOS moved to Section V).

Before version 1.0 (earlier versions numbered 0.1)

• ErgoCat (Casey Mullin, Cataloger in the Loop) added to Section VIII. Emerging: an AI-assisted cataloging interface chronicled openly in a newsletter, not yet an installable tool.
• ZineCat (Zine Union Catalog) added to Section XV. Active, Open Source: a union catalog for zines built on CollectiveAccess, with a companion Zine Subject Thesaurus built because zines resist LCSH.
• CIP (Cataloging in Publication) added to Section III after LCCN/PCN. Active, Open Access. Distinguished from PCN, which assigns only the control number.
• SCORM, xAPI/cmi5, and IEEE LOM added to Section II between PBCore and schema.org. IEEE LOM is Legacy: its rich discovery ambitions never fully materialized and it survives mainly inside SCORM.
• Fourth status value, Emerging, adopted (see Entry model).
• First publication, 6 September 2026.

• Deposit the Toolbox in Zenodo together with Core Readings, and revisit CITATION.cff at that point so both works are cited consistently.
• Confirm sources for labels that rest on a single page or were not re-read in this revision: the LC Classification Outline link; the NISO Z39.19 page and its current status and access terms; the M2R code license; GeoNames data terms and GEOnet Names Server availability; the Omeka.net, CollectiveAccess, and LibraryThing family links and terms; and the reuse terms of ZineCat and the Zine Subject Thesaurus.
• Confirm current release and migration details for DSpace and EPrints, and sunset dates for Voyager and Aleph, with the vendors or projects.
• Check status and access labels against current vendor or organization documentation whenever an entry is added or revised, and re-check vendor AI entries, whose documentation changes quickly.
• Find exact document URLs where an entry currently links only to an organization’s home page.
• Confirm terms not checked in the 18 October additions: the DCTAP specification’s license, the reuse terms of RightsStatements.org’s statements, Pinakes’s code license, Local Contexts Hub account terms, and the Yale Library AI Connector’s software license. Pinakes’s article and journal counts differ between its home page and its developer’s October post; recheck before citing them. Re-check Scholia’s data access after the Wikidata Query Service changes, and the Yale connector’s beta status.

After the course ends. Selection questions, not commitments:

• Rights and culturally appropriate reuse: which practical resources help someone record and communicate what may be accessed, reused, or shared, and whose authority informs that decision?
• Mapping and transformation: an instructional note and worked example on evaluating a crosswalk (what is preserved, lost, or inferred, and whether the transformation is reversible).
• Community maintenance: re-sort the professional list into discussion venues, standards-development communities, and groups maintaining practical outputs.
• Readings for the book-metadata route: reports and proposals on books, AI training data, and ebooks belong in Core Readings, to pair with the “Prepare a book’s metadata” route; verify sources first.
• Careers: roles students meet in this field, such as linked data librarian and metadata manager, perhaps linked from the routes they draw on.
• Broaden the archival subsection to collection description environments, including museum tools.

Coleman, Anita S. (2026). Toolbox: Standards, Systems & Tools for Organizing Information (Version 1.0). Infophilia Tools, in Infophilia, A Positive Psychology of Information (ISSN 3069-6526). https://infophilia.codeberg.page/infophilia-tools/toolbox/. Source and version history: codeberg.org/infophilia/infophilia-tools.

Licensed under CC BY-NC-SA 4.0.