Linear 1D formats

Linear symbols are encoded as a one-module-high BitMatrix. The vertical size is a rendering decision: use the renderer's barHeight option when the symbol needs to be printed or displayed at a useful height.

Registry entries

Label	id	Write	Generic image read	Notes
EAN-13	ean13	✅	✅	Twelve input digits receive a check digit; thirteen are verified.
EAN-8	ean8	✅	✅	Seven input digits receive a check digit; eight are verified.
UPC-A	upca	✅	✅	Encoded through the EAN/UPC family rules.
UPC-E	upce	✅	✅	Compact UPC form with expansion/check validation.
ISBN (Bookland)	isbn	✅	✅	ISBN-10/ISBN-13 is emitted as Bookland EAN-13.
Code 128	code128	✅	✅	Automatic code-set selection with checksum validation.
GS1-128	gs1128	✅	✅	Code 128 with GS1 FNC1 semantics and parsed metadata.
Code 39	code39	✅	✅	Optional modulo-43 check character and Full ASCII writer mode.
Code 93	code93	✅	✅	Checksum and start/stop grammar are validated.
ITF	itf	✅	✅	Interleaved 2 of 5; the generic reader rejects very short ambiguous reads.
ITF-14	itf14	✅	✅	Fixed-length ITF-14 writer; the shared reader reports the base itf format.
Codabar	codabar	✅	✅	Optional A/B/C/D start and stop characters.
Code 11	code11	✅	✅	Optional check-digit validation, enabled by default in the writer.
MSI Plessey	msi	✅	✅	Optional modulo-10 check digit and scanline reader.
Pharmacode	pharmacode	✅	—	Writer-only by design; unsafe for unrestricted generic autodetection.

The runtime registry is the source of truth for these flags. ITF-14 and Bookland ISBN are meaningful application profiles over their base symbol grammar, so a shared decoder can return itf or ean13 while preserving the decoded payload. This is not a data-loss claim; it is the current result-format contract.

Writing examples

The format-specific functions are available from the oned subpath and the root facade:

import {
  encodeCode39,
  encodeCode128,
  encodeEAN13,
  encodePharmacode,
} from '@sythos/js_barcode_universal/oned';

const retail = encodeEAN13('590123412345'); // check digit is appended
const code39 = encodeCode39('A-123', { checkDigit: true });
const code128 = encodeCode128('ABC-123');
const pharmacode = encodePharmacode(12345);

The generic dispatcher is useful when the format is selected at runtime:

import { encode, toImageData } from '@sythos/js_barcode_universal';

const matrix = encode('1234567890123', { format: 'itf14' });
const image = toImageData(matrix, {
  scale: 4,
  margin: 10,
  barHeight: 96,
});

encodeCode128() chooses a legal Code 128 representation from the payload. The writer does not expose a separate public codeSet: 'A' | 'B' | 'C' switch in the current API; callers should pass a valid payload and let the encoder select the efficient set transitions. gs1128 is the explicit GS1 dispatcher entry and is not just a label applied after arbitrary Code 128 data.

Reading images

The root reader accepts an RGBA image object. For a linear format, it samples multiple horizontal rows, measures run widths, checks the format grammar and validates checksums where the format defines them:

import { decode } from '@sythos/js_barcode_universal';

const results = decode(imageDataLike, {
  formats: ['ean13', 'code128', 'code39'],
  profile: 'camera',
  tryHarder: true,
});

if (results[0]) {
  console.log(results[0].format, results[0].text);
}

The camera profile is deliberately stricter than a quick scanline probe. It expects a coherent, quiet-zone-qualified read and never turns a partial run or an uncertain character into application data. An empty array is the expected answer for a blank, noisy or structurally inconsistent frame.

The current camera rotation policy tries the fixed eight in-plane angles in 45-degree steps where the format detector path supports them. It does not claim arbitrary projective distortion, curved labels, multiple overlapping symbols or severe glare. The generic reader also remains useful for a clean module-aligned image and for an application that already extracted a suitable scanline.

Pharmacode is deliberately write-only

Pharmacode is present in listFormats() so an application can generate a symbol, but it reports canRead: false. The one-track narrow/wide grammar has no strong finder frame for unrestricted image autodetection. Guessing a value from an arbitrary row would create false positives, so the generic image path returns no Pharmacode result instead.

Use the writer only when the application already owns the input value and has a separate, trusted reading strategy:

import { encodePharmacode } from '@sythos/js_barcode_universal/oned';
import { toSVG } from '@sythos/js_barcode_universal';

const matrix = encodePharmacode(12345); // legal range: 3..131070
const svg = toSVG(matrix, { scale: 4, margin: 10, barHeight: 80 });

This document does not suggest that a downstream OCR or scanner result should be fed back into the generic decoder without its own validation boundary.

Licensing and naming

The runtime code is MIT-licensed original Sythos work. The format names are descriptive; they are not a certification or endorsement. The engineering inventory and review labels live in LICENSE, NOTICE.md, and the matching files in licenses/.