Romano, tu quoque :-). This is exactly the finger/...

2011-10-18T18:34:30.915+02:00

Romano, tu quoque :-).
This is exactly the finger/moon problem :-(, as this post has very little to do with checkout, and much to do with how we approach requirements in general. Still, I’ll take your comment as a good chance to explain a few points better:

- in a general case, we not necessarily have a reference system, like an existing supermarket. Therefore, mere observation / replication is not enough. However, we may recognize common sub-problems on a smaller scale (which is also why I suggest to go at a finer granularity than the usual “analysis patterns”).

- even if we have an existing system, we don’t necessarily know its behavior in boundary cases. I surely haven’t tried boundary cases at the supermarket. Could be fun now that self-scanning checkout is more popular.

- even if we have an existing system, and we can test boundary cases, it does not mean the corresponding behavior is right. It’s extremely common, for legacy systems, to handle corner cases (if they do at all) not in the most logical and valuable way, but in the only possible way. That is, corner cases are often discovered after implementation (exactly because analysis was lacking) and handled within the limits of the existing implementation (like: you don’t want [or don’t have time] to move from a linear to a combinatorial process, so you hack something together). This is the kind of problem I hinted about in my 1997 paper.

If you were really careful when you go shopping in...

2011-10-18T10:23:15.984+02:00

If you were really careful when you go shopping in a supermarket, some questions you might avoid them. Lesson: Common sense to remember the real experience.

Romano

Comments on Carlo Pescio - blog: You're solving the wrong problem

Romano, tu quoque :-). This is exactly the finger/...

If you were really careful when you go shopping in...