Invisible XML (iXML) Cheatsheet

Invisible XML (iXML) Cheatsheet

Data Just Wants to Be Format-Neutral

Wow! Invisible XML is a dream come true!

Problem

You have some (unstructured) text input:

Russell Watson

and you need a structured form of it?

XML:

<Name>
  <FirstName>Russell</FirstName>
  <LastName>Watson</LastName>
</Name>

or JSON?

{
  "Name": {
    "FirstName": "Russell",
    "LastName": "Watson"
  }
}

Fed up of string manipulation in FileMaker?

Never quite got the hang of parsers?

Solution

Invisible XML (iXML) to the rescue!

{My first ixml grammar to parse & structure a name}
Name: FirstName, ' ', LastName.
FirstName: letter+.
LastName: letter+.
-letter: ["A"-"Z"] ; ["a"-"z"].

Applying this iXML grammar to the input text Russell Watson (using an implementation of an ixml parser) produces the desired XML output:

<Name>
  <FirstName>Russell</FirstName>
  <LastName>Watson</LastName>
</Name>

FIXME LINKS HERE

Here is a cheatsheet to get you started with iXML.

Invisible XML (iXML) Cheat Sheet

A quick reference guide to the Invisible XML (iXML) standard, based on the W3C Community Group Final Report (2023-12-12).

Purpose

Invisible XML allows you to parse non-XML data and transform it into clean XML using a readable grammar.

Basic Grammar Structure

Grammar Rules

Syntax Meaning
rule: expression. Defiition of a rule. The . terminator is required.
rule = expression. Alternative delimiter =

Each rule defines how to match input.

  • The expression defines what to match
  • The rule name becomes an element in the output XML
  • The first rule is the root element of the output XML
  • Rules are case-sensitive

Expressions

Syntax Meaning
'text' Literal string: match the string text
"A's" Literal string: match the string A's
["A"-"Z"] Character range: match any character from A to B
[L] Character class: match a character in the unicode class
rule Reference: Match the named rule
A, B Sequence: match A followed by B
A; B Choice: match A or B
A \| B Choice: match A or B
A? Quantifier: match zero or one A (optional A)
A* Quantifier: match zero or more As (optional, multiple As)
A+ Quantifier: match one or more As (multiple As)
A**B Quantifier: match zero or more As, delimited by B
A++B Quantifier: match one or more As, delimited by B
(A, B) Group: treat the elements inside the ()s as a group
(A, B)* Group quantifier: applies to the whole group
{comment} A comment: ignored(*)
{WS} (*) Some special built-in rules are hidden in comments

Markup Control

Markup vs Non-Markup

Use the - prefix to suppress a rule’s XML element or supress matched content:

Syntax Meaning
rule: 'A' Normal = markup and output: Output <rule>A</rule>
-rule: 'A' Supress markup: Output A
rule: -'A' Supress content: Output <rule></rule>
-rule: -'A', 'B' Supress markup and content:Output B
  • Normal rule → produces XML element:
word: [L]+.
```→
  
```xml
<word>abc</word>
  • Non-markup rule → no XML output:
-ws: ' '+.
→ whitespace is ignored in output

⚙️ Special Constructs

Grouping

You can group elements using parantheses

rule1, (rule2, rule3)? → optional group

For example:

GNU-stands-for: "GNU's", (' ', 'Not', ' ', 'Unix!')+.

Character Classes 

["A"-"Z"] {→ character range}
[L] {→ unicode character class}

Combine character ranges using the or syntax:

letter: [A-Z]; [a-z]; [0-9].

Built-in Rules

Rule Description
{WS} Whitespace: space, tab, CR, LF
{NL} Newline characters
{Char} Any valid Unicode character
{NameStartChar} Start char for XML names
{NameChar} Valid char for XML names

Use them directly in grammars — no need to redefine.

Trapping errors

If an input matches the grammar, the XML is created, however, if the input does not match an error ixml element is output, describing where the error occurred and what was expected:

<ixml xmlns:ixml="http://invisiblexml.org/NS" ixml:state="failed">
Failure at line 1 column 9
Given ' ' (codepoint 32). Expecting one of:
   [L] {#4: LastName: [L]+.}
Input:
   Russell  Watson
           ^
</ixml>

To test for failiure you can check the XPath (FIXME)

/ixml:ixml/ixml:state='failed'

where ixml=”http://invisiblexml.org/NS”

🚫 What iXML Does Not Include

  • No actions or embedded code (unlike YACC)
  • No semantic value computation
  • Designed purely to transform input into XML

📖 Further Reading

✍️ Tips

  • Beware ambiguity: ensure your grammar can only match one way.
    • Don’t use ?as well as ‘*’.
    • the following is ambiguous:
    name: ["a"-"z"]+, optional-index?.
-optional-index: [0-9]*.
  • Use non-markup rules (-rule) to strip punctuation, whitespace, etc.
  • Use {WS} for optional spacing between elements.
  • Rule names = XML element names. Be consistent and semantic!