??“ The site behaves: does the site load many client-side scripts, how fast do pages load, is the
database a bottleneck, does the site require a lot of ftp downloads?
2
No, no, let me guess.
72
Manage Software Testing
You need to see the relationship between targets and measures: Marketing will want as many hits
as possible to convert into sales. However as response times (latency) grows users will abandon
the site. The cost of abandonment has to be balanced against the cost of response times.
See Chapter 16
.
5.5 Web Tests
The diagram in Figure 5.3 (based on Figure 5.1) shows the various areas of web functionality and how
they can be tested. This is a suboptimal approach. One might well ask,
???why not begin with a requirements
specification as in any other mission-critical system????
One answer is simply that they are sometimes missing
or incomplete, and that one has to deal with the world as it is. Another is that websites are frequently
assembled from legacy and other systems and, since they evolve a lot, keeping track of baselines is a
major, if necessary, task. The baselines of each of the bits of the systems are shown below.
??? The website system test (sometimes referred to as
large scale integration test)
covers all the site
features, its performance, usability, and security. It will usually involve one or more tester??™s
machine(s) containing a number of test tools (see Appendix D). The baseline of the test will be
the overall website requirements specification (however expressed).
Pages:
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213