Design preview · Available for review

Technical platform / explained

One foundation.
More than one way to run a business.

Behind the everyday screens are record identities, module boundaries, access rules, provider connections and release decisions. This guide follows those relationships: what owns each record, what can change it, and which connections are implemented or still being designed.

Same foundation. Different operating scale.

Choose an operating context to see how record ownership and access boundaries change.

For the person doing everything

Start with a clear record owner.

One owner can begin with customer, quote, job and invoice records in a single context. Additional modules can reference the same source records without competing copies.

Modular starting point

Businesses, locations and brands are not interchangeable.

A brand is a customer-facing identity; a location is where work, people or stock are situated; a separate legal business can require its own accounting authority, access boundaries and contractual identity. Larger or multi-entity arrangements need an explicit implementation design—never an assumption that sharing a logo means sharing a ledger.

How the technical pieces fit together.

The Feature Atlas covers individual functions. Here we focus on data models, event handling, permissions and system boundaries.

01 / Module and runtime boundaries

Shared record identity, distinct duties.

Modules can show different views of the same customer, job, product or transaction. Explicit identifiers and source links allow records to connect while retaining distinct owners and update rules.

  • Define which module owns and may change each canonical record
  • Separate local desktop state from authoritative connected services
  • Join records by stable IDs instead of duplicating information
  • Verify permissions and implementation status per workflow

Commercial agreements / CO · CA · POL

A price tells you what.
An agreement tells you why.

A negotiated rate, a customer exception or an accepted set of terms shouldn't become an unexplained number on an order. This is the commercial layer we're building towards: a traceable relationship between an offer, the agreement behind it and the work it governs.

01 / CO

Commercial Offer

Propose prices, quantities, scope and conditions before acceptance. Preserve what was actually offered.

02 / CA

Commercial Agreement

Keep the accepted version and its parties, dates, terms, exceptions and applicable scope connected to the customer.

03 / POL

Operational Policies

Define reusable rules, eligibility and boundaries for how the agreed terms should apply to transactions.

WHY THIS MATTERS IN A REAL BUSINESS

Imagine a repeat customer with their own agreed pricing.

They receive an offer for a run of manufactured products. The accepted agreement might apply only to that customer, product range, quantity or time window. On the next quote or order, the system should be able to show the catalogue price, agreed price and which accepted clause explains the difference—without changing the price for everybody else.

Implementation status: This describes the planned commercial architecture, not a claim that every CO / CA / POL rule is already released or automatic. Check specific capability stages in the Feature Atlas.

How the records can move together.

A record chain shows relationships, not a mandatory operational path.

Customer startsBooking, website order or quote
Terms & contextSource price, agreements and conditions
Work happensJob, stock, purchasing and QA as needed
Customer is billedInvoice and applicable payment record
History remainsActual costs, analytics and traceability

Feature catalogue ≠ technical architecture.

The Feature Atlas is the right place to search for detailed actions. The Workflows demonstration shows connected journeys. This guide explains modularity, boundaries and deployment. Each has a different job rather than repeating the same wall of text.

Technical foundations / accounting & stock

Immutable ledgers.
History that stays history.

Accounting and physical inventory need an explainable record of what happened, not just the number the dashboard shows today.

THE RULE Post a change. Don't silently rewrite the event that came before it.

Once an authorised entry is finalised, its original values and provenance are retained. Corrections appear as linked reversals or adjustment entries so the full sequence can still be audited. Drafts and unconfirmed movements can follow different rules.

01 / FINANCIAL HISTORY

Accounting ledger & journals

Preserve accepted invoice figures, GST treatment, payment postings and the entries that explain their effect on the accounts. A later pricing change must not recalculate a previous sale.

  • Capture the original source transaction and accepted commercial snapshot
  • Link journal entries to their source, effective date and authorised actor
  • Use reversals or correcting entries instead of silently editing posted history
  • Keep account and GST reporting reconcilable back to the source records
02 / PHYSICAL HISTORY

Stock movement journal

Receipts, transfers, picks, production consumption and adjustments form a traceable movement history. The on-hand balance is the result of those confirmed movements, not a number with its explanation discarded.

  • Record item, quantity, unit, source and destination where relevant
  • Preserve which job, order, receipt or count triggered the movement
  • Retain actor, time, reason and a linked correction or void note
  • Keep reservations distinct from physical movements until stock actually moves
ILLUSTRATIVE RECORD SEQUENCEThe original event remains visible after a correction.
01Goods receivedReceipt R-108 · 24 units entered+24
02Correction recordedCorrection C-109 · linked to R-108−2
03Reconciled balanceBoth source and correction remain traceable22

This is the accounting and stock-record integrity principle, not a claim that every journal, permission gate, reversal path or automated posting is production-complete across every DD configuration. Final availability depends on the module and its verified workflow.

Customer applications / deployment architecture

Built for your business.
Updated for your business.

Not one universal desktop binary with a new logo painted on. The delivery and update architecture is designed around individually identified customer builds.

Customer-specific build model · planned rollout

One proven foundation.
Every deployment has an identity.

Each customer's application starts from the Denver's Desk base, then is assembled and integrated for the needs, modules, data connections and working environment of that customer. Those customer builds are maintained independently — changes to the common base are selectively carried into the builds that need them.

See customer onboarding ↗
EXAMPLE BUILD RECORD / NOT A LIVE DOWNLOAD Customer 014
BUILD / RELEASE SERIESCustomer 014 · Version 1.3.2
DISTRIBUTIONAssigned domain + customer bucket
UPDATE ELIGIBILITYTargeted to this customer

Numbers are illustrative, not a promised version format.

01 / ASSEMBLE

Custom application

Start from the shared foundation, configure the requested modules, and build for that customer's operation and approved integrations.

02 / ASSIGN

Independent identity

Allocate a customer/build identifier and its own version sequence. Record the package, release history and customer-facing change notes separately.

03 / DISTRIBUTE

Customer-specific delivery

Plan authorised distribution from the customer's assigned web-domain path and storage bucket from initial provisioning, rather than one generic public installer.

04 / UPDATE

Targeted releases

Changes are selected, adapted, tested and approved for the specific customer build. Another customer's release must not silently update this one.

THE RELEASE PIPELINE WE ARE BUILDING TOWARD

A customer release is a controlled event.

Applications ledger → customer-specific build → verification and checksum → assigned distribution path → authorised update manifest. The source base can evolve without forcing a one-size-fits-all update onto every installation.

Moving to Denver's Desk

You don't have to work out the implementation on your own.

We want the technical work to be supported by people who understand what real operating records mean—not just a list of import columns.