Projects:AverageCostRefactor/DevelopmentAndDesignNotes

Introduction

In this document basic notes and requirements are listed to be considered during project design and development.

Notes

• Extensible algorithm definition to manage costing. On first phase standard and average will be available. FIFO, LIFO and Specific Identification will be defined on a later stage.
• Asynchronous cost calculation, new infrastructure similar to accounting engine needed.
  • Background process to calculate costs.
  • Flag on documents / transaction table to check that cost has been calculated.
  • Ability to know the last day costs have been calculated.
  • Extensible with templates to implement different costing algorithms.
• Cost calculated based on several dimensions:
  • client is mandatory and each product can define another one if needed.
  • client, organization, warehouse
• Dimensions to calculate are set on a new generic configuration table. This table can be accessed from product if a specific configuration to it is set.
• New processes shall be java classes using DAL.
• New API to generate the cost entry by each algorithm.
  • Define needed parameters.
• New API to get the cost of each transaction.
  • Different for each transaction that can consume cost.
  • Transaction as a parameter.
• Pending to determine how to source of the price of a receipt/shipment (invoice, order or price list)
• Pending to define how to apply landed cost.
• The order of the transactions is determined by the created property of the M_Transaction table.
• Pending to estimate the effort to include not instanceable attribute sets as a costing dimension.
• FIFO / LIFO algorithms require a new table(s) to store each transaction and its price.
  • Each outgoing transaction from a storage bin should be related to an ingoing transaction.
    • It could be null if shipment is done before the receipt. Study inconveniences as unknown price of the product.
  • This table is populated when the cost is calculated. Each input has a record. Each output has at least one record related to an input.
• Each algorithm should provide a method to return the stock valuation of a warehouse as a datasource. To be filterable by the dimensions.