blue_slonopotam - О стоимости ошибок. [entries|archive|friends|userinfo]
blue_slonopotam

[ userinfo | ljr userinfo ]
[ archive | journal archive ]

О стоимости ошибок. [Oct. 27th, 2009|12:49 pm]
Previous Entry Add to Memories Tell A Friend Next Entry
A general belief was that some accident or series of accidents involving loss of life, perhaps followed by an expensive suit for damages, would persuade software managers to consider the merits of program verification.

This never happened. When a bug occurred, like the one that crashed the maiden flight of the Ariane V spacecraft in 1996, the first response of the manager was to intensify the test regimes, on the reasonable grounds that if the erroneous code had been exercised on test, it would have been easily corrected before launch. And if the issue ever came to court, the defense of 'state-of-the-art' practice would always prevail. It was clearly a mistake to try to frighten people into changing their ways. Far more effective is the incentive of reduction in cost. A recent report from the U.S. Department of Commerce has suggested that the cost of programming error to the world economy is measured in tens of billions of dollars per year, most of it falling (in small but frequent doses) on the users of software rather than on the producers.


Некоторые думают, что относительное падение зарплат в CS/SE снизит количество идиотов, работающих в отрасли. С какой стати? Есть стоимость исправления ошибки и есть стоимость исправления последствий ошибки. За последствия платит покупатель. Стоимость исправления самой ошибки SE самая низкая среди всех возможных инженерных ошибок, т.е. выбор инженера-механика или электроника гораздо критичнее, чем программиста. Это довольно хорошее объяснение и отношения "если ты не можешь делать hardware, делай software" и результирующего снижения зарплат - выбирать "лучшего" не так важно, как в других отраслях. Снижение зарплат приведёт лишь к тому, что отрасль потеряет тех людей, кому есть из чего выбирать, или просто будет переполнена socially retarded.

Вдогонку. Выдержка со слашдота, не из самой статьи.
the highest-performing students ...are opting out of science and engineering ..., suggesting that the threat to American economic competitiveness comes ...from a lack of incentives that would make science and technology careers attractive
LinkLeave a comment