It is a common scenario: someone has found that performance becomes a serious problem at some day, so the goal is set to the team: "Improve the performance at least x percentage before y date" since then. Then relative objectives and milestones has been inserted into the schedule, followed by teams of experts and sprints been involved in. Suddenly but not surprisingly, "Performance improvement" becomes the most popular word and highest priority in every meetings and reasons.
Consider the familiar story: A grand company has found there is problem in its financial statements. Thus a director has been hired and assigned by the board to reach (or secure) the goal of company's profit. The following is just like a standard solution: cut the cost by firing employees, shutdown branches, and reduce perks. Possibly the goal then be reached when it met the deadline, or not; perhaps another new CEO has been hired in two or three years later, along with new goals, or not.
It is quite straightforward: cutting the cost is the only thing that could be sure on it's way toward to the profit goal under the pressure. The business directions, market strategies, or the tailored advisement are all gambles still, with possibilities, or risks. It is no doubt an available and effective option to improve the performance and hit the goal immediately, but only true if the price been understood clearly.
Similarly to the software case, cutting features, removing error handling, and crashing the flexibilities are predictable to reach the performance goal. Not to say cheating to the profiler, refusing the new functions, or sacrifice on system design. The end justifies the means. Still it is definitely an available solution when the cost is acceptable, such as hot fixes when performance kills.
If performance considerations is involved regularly in every single point at every aspects. The price of hot fixes would be aware, the cheating should be found, and the broken structure could be healed. And yes, it takes more time, but it make much clearer on every cost of decisions on chasing performance. It is the alternative, make the performance improvement be the culture.
Consider the familiar story: A grand company has found there is problem in its financial statements. Thus a director has been hired and assigned by the board to reach (or secure) the goal of company's profit. The following is just like a standard solution: cut the cost by firing employees, shutdown branches, and reduce perks. Possibly the goal then be reached when it met the deadline, or not; perhaps another new CEO has been hired in two or three years later, along with new goals, or not.
It is quite straightforward: cutting the cost is the only thing that could be sure on it's way toward to the profit goal under the pressure. The business directions, market strategies, or the tailored advisement are all gambles still, with possibilities, or risks. It is no doubt an available and effective option to improve the performance and hit the goal immediately, but only true if the price been understood clearly.
Similarly to the software case, cutting features, removing error handling, and crashing the flexibilities are predictable to reach the performance goal. Not to say cheating to the profiler, refusing the new functions, or sacrifice on system design. The end justifies the means. Still it is definitely an available solution when the cost is acceptable, such as hot fixes when performance kills.
If performance considerations is involved regularly in every single point at every aspects. The price of hot fixes would be aware, the cheating should be found, and the broken structure could be healed. And yes, it takes more time, but it make much clearer on every cost of decisions on chasing performance. It is the alternative, make the performance improvement be the culture.
留言
張貼留言