Glossary

This glossary explains the terms used in DocView Capture and in this help. Terms are in alphabetical order.

A–C

Archived: The state of a batch after export has finished. Archived batches stay in the Batch Manager and can be moved back to any step. They are deleted after a number of days (30 by default). See Archive and retention.

Audit log: A record of administrator and operator actions, including field changes with old value, new value, user and time. See Audit log.

Batch: A group of pages or files captured together, for example one scan session or one upload. A batch contains documents, and documents contain pages.

Batch field: A field that applies to the whole batch, such as a batch reference or cost centre, rather than to one document.

Batch Manager: The screen that lists all batches, with filters and search. Supervisors can open, reprioritise, move, suspend, resume, unlock, retry and delete batches there. See Batch Manager.

Batch Setup: The complete configuration for one kind of batch: workflow, image cleanup, separation, document classes, fields, finders, validation, event scripts and export. Known as a "batch class" in some other products. See Batch Setups overview.

Classification: Deciding which document class a document belongs to, using rules (keywords, patterns, barcodes, anchor text) or a trained classifier.

Classifier training: Teaching the machine-learning classifier from sample documents (Document classes & fields › Samples & training). Every training is kept as a numbered model with an accuracy report, so models can be compared and a version can use any of them. See Samples and classifier training.

Confidence: How sure DocView Capture is about a class or a value. Below the Threshold, a document or field is sent to an operator.

Correction: A manual step where operators fix one field at a time across the whole batch, using a clip of the page. See Correction.

D–F

Document: One logical document in a batch, such as one invoice. It has a document class, fields and one or more pages.

Document class: A type of document, such as Invoice or Credit note, with its own fields, finders and classification rules. See Document classes and classification.

Classification: A manual step where operators confirm uncertain document classes and check how pages were split, before extraction. See Classification.

Double-key verification: A second operator keys selected fields again without seeing the first entry. Mismatches are flagged. This is the Verify step. See Verify.

Draft: A Batch Setup version that is still being edited. Operators can't use it until it is released.

Event script: A JavaScript script that runs when something happens, for example when a field changes or a document is completed. See Events and scripts.

Export destination: Where exported files and data are delivered: a folder, SFTP, SQL Server, REST web service or C# script. Known as an "export connector" in some products. See Output and export.

Field: A piece of data captured from a document, such as an invoice number or total. Also called an index field.

Finder: A way of locating a value on a document: by zone, label, pattern, barcode, table, database lookup or script. Fields take their value from finders. See Finders and zones.

Formatter: A named rule that tidies a value, for example trimming spaces or changing case.

Four-eyes rule: A setting on a workflow step that requires a different person from whoever worked on the batch in another step.

I–O

Import source: An automatic way for documents to arrive: a watched folder, an email mailbox or the REST API. See Import sources.

Index / Validate: The main manual step where operators check and correct field values next to the page image. See Index and validate.

Lock: While an operator or DocView Worker has a batch open, nobody else can work on it. Operator locks expire after a period of inactivity, and supervisors can unlock batches.

Lookup: A database query used to check values, fill in fields or offer a pick list. See Lookups.

Module: A kind of workflow step, such as Scan, Index / Validate or PDF generation & export.

Notification: A message DocView Capture showed you (a toast in the corner of the screen) or news about your scheduled reports and batches. All are kept in the bell at the top of the screen and on the Notifications page. See Notifications.

One-touch scanning: A way of scanning where the operator presses a document type button (or key) before feeding each document, so a new document of that type starts. No automatic separation or classification is needed. See Scan documents.

OCR (optical character recognition): Reading text from page images. DocView Capture uses Tesseract by default, or a cloud engine. See OCR engines and languages.

OMR (optical mark recognition): Reading tick boxes and signature boxes on forms. See Forms, tick boxes and signatures.

P–S

Patch code: A special barcode pattern printed on a separator sheet, used to split documents.

PDF/A: A PDF format intended for long-term archiving.

Pinned: A Batch Setup whose operations use one chosen version, rather than Auto (the latest released version). See Versions and publishing.

Queue: The list of batches waiting in a workflow step.

Release: For a batch, sending it on to the next workflow step. For a Batch Setup version, making it available to operators.

Scan Agent: The small Windows program that connects a TWAIN or WIA scanner to the browser. See Scan Agent.

Scan profile: Saved scanner settings, such as resolution, colour, duplex and paper size. See Scan profiles.

Screen designer: Designs the form operators see in a workflow step, per document class. See Screen designer.

Sheet: One piece of paper. Scanned on both sides (duplex), a sheet gives two images, its front and back. Many scanners report the sheet and side, and the Scan screen shows them, for example Sheet 3 · back.

Searchable PDF: A PDF with the page image plus an invisible OCR text layer, so you can search and copy the text.

Secret: A password, key or connection string stored encrypted. Settings refer to it by name.

Separation: Splitting the pages of a batch into documents, by page count, blank pages, barcodes, patch codes, keywords or classification. See Image cleanup and separation.

Straight-through processing: When a document is classified and all its fields are confident and valid, manual steps are skipped and it goes straight to export.

Suspend: Put a batch on hold. It stays where it is until someone resumes it.

T–Z

Table field: A field with rows and columns, such as invoice line items.

TWAIN / WIA: Standard Windows interfaces for scanners. The Scan Agent supports both.

Validator: A named check on a value, such as a pattern, range, date, list, or a UK VAT, IBAN, Luhn or EAN-13 check digit. See Validation rules.

Version: Each release of a Batch Setup is a version. Batches keep the version they started with.

Watched folder: A folder that DocView Capture checks regularly and imports new files from. Also called a hot folder.

DocView Worker: A background program that does the automatic work: OCR, separation, classification, extraction, export and import polling. Several DocView Workers can share the queue. See DocView Workers.

Worker profile: Settings for a DocView Worker by name: which modules it runs, in which priority, how many batches at once, and for which Batch Setups.

Workflow: The ordered steps a batch goes through, from Scan to PDF generation & export. See Workflow.

Zone: An area drawn on a sample page where a value is read. See Finders and zones.