Skip to content

History / BPMS adapters

Revisions

  • Rewrite thirteen shorter pages in plain English Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lxo4MXbp6VPoDkUrTBNXJ1

    @stephanpelikan stephanpelikan committed Oct 11, 2026
  • docs: the adapter page points at the ladder instead of retelling it The levels are explained on the workflow module page. The extension section told the same story a third time, so it names the section now and keeps only what is the extension's own: it reads both halves of the ladder, which is where the eight positions come from, and the level is asked before the adapter inside it. Checked against ExtensionSettingsResolutionTest and SettingsLevel in adapter-platform-integration: same resolution as an adapter setting, a level beats the adapter section above it, and an adapter section beats the general section of its own level. Story 601. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0168Fox39crp8dfWkZ9W84Ub

    Claude committed Sep 25, 2026
  • docs: one word for the reader, and eight places for a setting The glossary now names both words for the same thing: the core calls it an extension, its own documentation calls it an adapter sitting next to the BPMS adapter. The migration page says adapter and points at the glossary. The configuration paragraph said two places. There are eight: four levels, each of them again for one adapter. Checked against MigrationAdapterProperties#resolveForExtension, migration-adapter/README.md and decision 53. It also says now that the core reads from the general towards the specific, and that an extension without a section of its own has nothing to do for it, which leaves the Business Cockpit's own section as the exception it is. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RYPiyZ32jQQhqYKx44cw23

    @stephanpelikan stephanpelikan committed Sep 24, 2026
  • docs: the flattening rule links on to what each BPMS makes of it Every adapter wiki now carries a page "What becomes of a shared value", answering the same four questions in the same order. This wiki keeps the rule which is the same everywhere and points at the three pages instead of repeating them. Workflow-aggregates: the sharing rule links on to the three pages, and the serialization-format link of the Camunda 7 adapter follows the chapter which moved out of its configuration page. Choosing-a-BPMS gains the row a reader comparing two BPMS was missing, what becomes of a value the engine has no type of its own for, and BPMS-adapters names the new page among what an adapter wiki carries. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RYPiyZ32jQQhqYKx44cw23

    @stephanpelikan stephanpelikan committed Sep 18, 2026
  • Say that the Business Cockpit's keys stayed where version 1 had them The migration page told applications that every cockpit key had moved into the extension section. It did not stay that way: the cockpit reads `vanillabp.cockpit.*` and the `cockpit` sections of a workflow module, of a single workflow and of one of its user tasks, which is where version 1 read them, so an upgrade is a dependency and not a rewritten configuration file. The page about extensions gains what makes that possible. The two locations the core reserves are still the ones an extension uses by default, and an extension which owns a section outside them binds that section itself, with the same overlay a BPMS adapter uses for keys of its own. VanillaBP neither reads those keys nor knows them, and on Quarkus the overlay is what keeps a key below `vanillabp` from ending the startup. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01YU73krA4jKHuWYB155np2X

    @stephanpelikan stephanpelikan committed Sep 13, 2026
  • Document extensions and the Business Cockpit upgrade The extensions section of the BPMS adapter page now says what an extension does from an application's point of view: when it joins the deployment, that it may bring annotations of its own, that its reports travel through the outbox, and the two configuration locations the core reserves for it. The Business Cockpit is named as the extension to look at, with its three BPMS repositories and the platform-neutral half in the cockpit's own repository. The migration page gains a Business Cockpit section: the renamed artifacts, the keys moving into the extension's section, the refused version attribute, reports arriving after the commit, and the two things which differ per BPMS - a new process version on Camunda 8 and a history level on Camunda 7. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01YU73krA4jKHuWYB155np2X

    @stephanpelikan stephanpelikan committed Sep 13, 2026
  • docs: the guide is linked, the boundary is named Two answers existed and were not reachable from where the question comes up. Somebody weighing an adapter for another BPMS found the adapter guide only by browsing the repository, so the adapter page and the glossary link it and say who it is written for. And no page said who may complete a user task or read the viewer API, which is a question VanillaBP does not answer at all: it authorizes nothing and publishes no endpoint, the connection an adapter opens to its BPMS being the one security surface it does own. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XjxJNwTRcKzTmXCwLU7Wi4

    @stephanpelikan stephanpelikan committed Aug 31, 2026
  • docs: every page opens with what to do and carries a table of contents A reader who has never used VanillaBP met the reasoning for a rule before the rule itself, and had no way of telling how far to read before writing a first workflow. The pages now start with the sentence somebody can act on and keep the background where it belongs, further down. Nothing was deleted; what explains a rule moved behind it. Every page gained a numbered table of contents in the shape Home and the Spring Boot page already used, so a reader sees the sections and what each of them answers before scrolling. Where a first heading only repeated the title GitHub shows above the page, it is gone and the page starts with its first sentence. Home is now ordered the way a newcomer needs it, and what VanillaBP is sized for became a page of its own: it answers a question somebody asks before they start, not one they trip over in the middle of the introduction. The architecture picture from the repository's diagrams folder is embedded there, linking to its Excalidraw source. The horizontal rules of the workflow-module pages are gone, the heading is the separator everywhere else. Hints are written as sentences rather than marked with a label, headings sit at one depth per page, and the paragraphs written by hand over the last months were brought to the tone of the rest without changing a single statement they make. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

    @stephanpelikan stephanpelikan committed Aug 31, 2026
  • docs: a rule says where its decision is written down Five decision logs hold 61 entries, the wikis linked none of them, and 22 of the 42 objections an outside review raised were already answered in one. A reader who wondered why a rule is what it is had no way there short of reading source repositories. So a rule carries a (why) link now, pointing at the numbered entry which decided it. Home says once what the links are and where the logs live, and the form is the same on every page of every wiki: an inline link at the end of the sentence stating the rule. Numbers are handed out once and never reused, so the links keep resolving. The rules the review named are linked first - the outbox on Camunda 7, probing per operation, the two retentions, sharing aggregate values, the Camunda 8 correlation key - and the remaining entries follow wherever a page states their rule. BPMS-adapters also says in one sentence what the module migration-adapter is, which it linked without explaining. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CvgyhjwTEh834mwbptZZxh

    @stephanpelikan stephanpelikan committed Aug 27, 2026
  • docs: a page for readers who have to pick an engine first The comparison a newcomer needs existed twice, in the contributor skills and in the review guide, and neither is published. Choosing-a-BPMS carries it now: one row per property, one column per adapter, plus what the review found missing - when the shared values reach the BPMS and what a crash between commit and command leaves behind, what identifies a delivery, whether the age of an open task is measured at all, and whether the adapter can tell that a workflow is unknown to it or has to guess. Three of those rows decide how much an application has to carry itself, so they are repeated below the table in prose. Underneath it, one short section per question somebody arrives with: a Camunda 7 in production, a Camunda 8 cluster, or nothing to run an engine on yet. Home links it from the adapters entry and BPMS-adapters from the top, both in the words a reader without a BPMS would use. Home also gets the glossary the wiki was missing. It defines the terms a user meets while reading it - adapter id, election, the outbox, the delivery log - and stays out of the contributor vocabulary, which the repositories document. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CvgyhjwTEh834mwbptZZxh

    @stephanpelikan stephanpelikan committed Aug 27, 2026
  • docs: what Quarkus reports when the core extension itself is missing Story 81's message about a missing BPMS adapter comes from the core extension, so an application without that extension gets Quarkus' unsatisfied-dependency error for 'ProcessService' instead, and nothing of VanillaBP can reach the reader: the support library is a plain JAR on purpose, because a workflow module stays free of platform infrastructure and Quarkus extensions belong to the application. Both pages now name that error, so searching for it finds which dependency belongs where. Story 82. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Jss55UjjvxYcLQPVFUFe25

    @stephanpelikan stephanpelikan committed Aug 18, 2026
  • docs: the artifacts of every adapter, and where the messages point The startup message about a missing BPMS adapter no longer lists the adapters - it links here, so this page has to carry the list: Spring Boot artifact and Quarkus extension per adapter, and that the version is not managed by the VanillaBP BOM. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Jss55UjjvxYcLQPVFUFe25

    @stephanpelikan stephanpelikan committed Aug 18, 2026
  • docs: what happens if you forget the BPMS adapter Story 81 gives that case a message of VanillaBP's own, naming the artifacts to add. The page a newcomer lands on now carries it, plus the note that an application without a workflow module is not reported. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Jss55UjjvxYcLQPVFUFe25

    @stephanpelikan stephanpelikan committed Aug 18, 2026
  • docs: a phase-two failure repeating cannot fix blocks the entry at once

    @stephanpelikan stephanpelikan committed Aug 17, 2026
  • Story 55: point at what an adapter wiki still carries An adapter wiki is now configuration plus deviations, so the pages delegating BPMS specifics have to name the page which answers them: the wiring conventions and correlation mechanics sit on the adapter's Configuration page, what a BPMS cannot serve on its Deviations page. "What is documented in the adapter's wiki?" listed six kinds of content, four of which the adapters no longer keep. Also says that Camunda 7's table-prefix does not start an engine, where the version-1 migration table offers it as the successor of camunda.bpm.*.

    @stephanpelikan stephanpelikan committed Aug 13, 2026
  • Unify the wording of all pages Two passes over the wiki: 1. The pages written earlier (workflow modules, workflow aggregates, the platform integration intros) were aligned with the wording of the newer ones: grammar and typo fixes, consistent terminology (process = the BPMN model, workflow = one execution; message names instead of event names) and active voice instead of impersonal constructions. 2. All pages were revised to read less like generated text: em dashes in prose replaced by proper punctuation, promotional phrasing and filler removed, boldface and inline-header lists reduced to where they aid scanning. Two corrections found on the way: the Quarkus workflow-module page pointed at the "main Spring Boot application module", and the property precedence lists of the two platform pages are sorted in opposite directions, which is now stated (Spring Boot: highest first, Quarkus: lowest first, matching the config source ordinals 250/251/255/256). No configuration key, behavior or example was changed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

    @stephanpelikan stephanpelikan committed Aug 10, 2026
  • Extensions belong to contributor documentation (story 29 follow-up) The Extensions page documented a mechanism only someone BUILDING an integration needs - that is contributor documentation and now lives in migration-adapter/README.md. BPMS-adapters gained a short section instead: besides the adapter, an extension may take part in deploying workflow modules (the VanillaBP Business Cockpit as the example), users only add its dependency, and building one is documented in the migration adapter's README. The Home entry is gone with the page.

    @stephanpelikan stephanpelikan committed Aug 7, 2026
  • Document the viewing API (story 26) New page 'Viewing workflows': the three read-only ProcessService methods, the call-activity drill-down via the secondary history context, the opaque (adapter-namespaced) process definition ids and the guiding errors - linked from Home.

    @stephanpelikan stephanpelikan committed Aug 6, 2026
  • Move BPMS-specific content into the adapter wikis The main wiki now documents only what is identical across all BPMS. Everything specific to Camunda 7, Camunda 8 or the Process-Engine-API moved into the respective adapter wiki. Where such content stood, the pages now carry a list of the available adapters linking directly to the matching page of the adapter's wiki. The new page 'Available BPMS adapters' holds the full list and is linked from Home. A custom footer with the VanillaBP logo and the license notice was added. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

    @stephanpelikan stephanpelikan committed Aug 5, 2026