Prev | Current Page 137 | Next

Peter Farrell-Vinay

"Manage Software Testing"

None has any experience of .NET
though management believes their expensive knowledge of J2EE will ???carry across.???
??? Complexity: although the AMS worked, it did so inconsistently. A look at the code reveals why.
There is a risk that the new version will be as badly designed.
??? Lack of requirements: the old AMS was itself a kludge of a bought-in source from a nowdefunct
Alerts company. There are consequently no requirements for the new AMS other than
???it should do what the old one did.???
??? Voice input is much touted and terribly fashionable, but not yet proof against good folks from
Aberdeen and the Adirondacks and their challenging speech. Thus either user HR has to be
kept in the loop to ensure that staff pronunciation is of an acceptable level (not a core skill of
HR departments), or there could be problems.
??? No definition of throughput: The existing system handles twenty simultaneous users. The
proposed system must handle 2000. No work has been done to show that the database could
handle this level of throughput let alone the rest of the system.
??? Age: The current development hardware is 5 years old and must be replaced.
Note that these are
threats
not
risks
. The lack of developer expertise (for example) is not in itself
a risk but a threat. No one will refuse to buy your software because none of your programmers
has a PhD. They might if your programmers are a bunch of cowboys. The risk is simply that they
36
Manage Software Testing
don??™t buy
because
the product doesn??™t work, because your developers are a bunch of cowboys.


Pages:
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149