Skip to content


Repository files navigation

Element Specification

Build Status

ElementSpecification is a small library that assists with writing automated UI tests. It provides a simple, easy to read way to specify a particular element in the interface. For example identifying a button to click to submit a form or identifying a table row to check the data presented is correct. Using ElementSpecification not only makes those selectors simpler to read, it automatically avoids many of the most common mistakes that can lead to intermittent failures in tests.


Quick Start

The exposed API is a single class ElementSpecification which provides a number of static methods as starting points for selecting elements. You can then refine the selection by chaining further method calls. Using static imports is strongly recommended to avoid repeating ElementSpecification. for every selector.

Literally any element on the page.
Any element of the specified type, e.g. table.
Select an element with a given classname. Equivalent to the css `.classname`.
Select an element by ID. Equivalent to the css `#id`.

There are two categories of methods that can be used to refine the selection:

  • with methods add additional requirements for the current element. Most commonly used:
    • withId(id)
    • withClass(classname)
    • withAttribute(attributeName)
    • withAttributeValue(attributeName, attributeValue)
    • withAttributeContaining(attributeName, attributeValueSubstring)
  • thatContains methods move to a descendent element. Most commonly used:
    • thatContainsA(tagName)
    • thatContainsAnElementWithClass(classname)
    • thatContainsAnElementWithId(id)

Element specification selectors only ever move down the DOM tree, never back up or across. This is a deliberate restriction to keep selectors simple and reduce brittleness. More complex selectors tend to depend on the layout of the page rather than the semantic structure leading to more brittle tests.

Avoiding Intermittency

ElementSpecification will automatically avoid a number of potential causes of intermittency, for example ID attributes that happen to start with a number can fail to match CSS selectors in older versions of FireFox so:

anElementWithId("3ADF8") is equivalent to *[id="3ADF8"] because #3ADF8 would not match.

Similarly when using XPath, most people would translate anElementWithClass("valid") as //*[contains(@class, 'valid')] but that would also match elements with the classname invalid. Instead ElementSpecification generates contains(concat(' ', @class, ' '), ' valid ').

There are a number of little details like that built into ElementSpecification which avoid the need for every person on your team to be constantly aware of every possible pitfall while writing tests.


Note that generally the CSS version is shorter but the ElementSpecification syntax aims to be clearer and readable without having to recall all the details of CSS.

anElementWithId("login-form").thatContainsAnElementWithClass("username") is equivalent to #login-form .username

anElementWithId("data-table").thatContainsA("tr").withAttributeValue("data-id", "78").thatContainsAnElementWithClass("name") is equivalent to #data-table tr[data-id='78'] .name

anElement().withAnyOfTheseClasses("buy-price", "sell-price") is impossible in CSS so ElementSpecification automatically switches to XPath and generates:
//*[contains(concat(' ', @class, ' '), ' buy-price ') or contains(concat(' ', @class, ' '), ' sell-price ')]

Output Formats

ElementSpecification will automatically select the simplest (and generally fastest) format that a selector can be accurately represented in. The three available formats (in order of preference) are:

  • Simple ID (ie:
  • CSS
  • XPath

The selector can be generated as either a selenium 1 style locator (String) using asSeleniumLocator or a WebDriver locator (By) using asWebDriverLocator.


A small library that assists with writing selenium or WebDriver UI tests.







No packages published