The time and money invested creating a website makes testing it
properly essential. Last month we tested a site for the Department
of Communities and Local Government (DCLG). We checked its
accessibility, standards compliance, functionality, branding and
content across a range of browser.
Before testing any project we draw up a test plan from original
project documentation. This can include specifications, statements
of requirement, standards the site is designed to comply with and
branding guidelines. Without a plan there is a danger that testers
will miss areas of a site or functionality that need testing.
When we draw up a test plan we included a number of separate,
independent tests that can be performed. Each test confirms if the
site complies with one requirement from the specification. Some
tests, such as compliance with html standards, are independent of
web browsers. Others, like testing site navigation works, need to
be carried out across a range of browsers to ensure a feature is
available to all visitors. Some are best carried out by testers
with specialist expertise, such as a designer checking a site’s
styling complies with branding guidelines.
Once we have a test plan testers can carry out the testing. A
Pass or Fail is recorded for every test to ensure that every test
is run at least once. While carrying out the test plan testers
record any issues outside the test plan they comes across. A
written explanation for each failure or issue explains what is
wrong.
With a detailed audit trail recording outcomes and who carried
out each test we can support the conclusions of our testing. This
is important as it builds confidence for the client we are testing
for and for the site’s creators. A summary list of failures would
be quicker to produce and provides a list of issues requiring
attention. A summary list doesn’t build confidence in the
effectiveness of the testing.
The DCLG testing was based on the site specification and
branding guidelines. A list of web browsers the site had to work
with was identified by the DCLG. We drew up the list of tests and
identified which tests needed to be carried out across different
browsers and which were independent of the browsers. The test plan
included over 500 separate tests.
With the test plan written we went through it step by step. For
each test we decided if the site passed or failed and recorded the
outcome. A summary and a report of the test results was compiled
and sent to the DCLG the next day.