|
老友圈|VIP
飘雪 - 帖子
- 262
- 精华
- 17
- 积分
- 15
- 书币
- 57135 枚
- 注册时间
- 06-12-20
- 最后登录
- 08-3-12
|
1#
大 中
小 发表于 06-12-27 15:02 显示全部帖子
THE COGNITIVE DYNAMICS OF COMPUTER SCIENCE
THE COGNITIVE DYNAMICS OF COMPUTER SCIENCE | | 收藏 | | | 作者 | Szabolcs Michael de Gyurky | | ＩＳＢＮ | 0-471-97047-6 | | 出版日期/版本 | 2006 | | 格式 | ＰＤＦ Adobe Acrobat |
主题内容 | | Introduction
Computer software development has become very expensive in these United States,
and there are a number of issues that high cost has brought with it. I suppose the
most important of these is that the programming jobs are going overseas, along with
most of our other manufacturing jobs. We can claim that we still do high-technology
design, and the critical jobs of testing and integration. But these jobs, even if we did
them well (which we don’t always do), will also go overseas, along with the making
of our beloved Levi jeans.
There is also a silent acceptance within the computer science community that the
quality of American software is poor, to put it mildly. This is well known, even to
our major publishing houses. They note that we write fewer and fewer ‘‘readable’’
books that our young and not so young practitioners can use. We forget that the
academic preparation of young Americans is the foundation of their professional
lives. It is hoped that a long and serious apprenticeship lies ahead of them, until
they, too, become masters of their professional fields. A large portion of our publications
and books is written just to publish, and not to teach the young professional.
When I pick up foreign textbooks, I’m astonished at the pains their
authors took to make them readable and truly useful to the reader. Foreign books
are not written solely for one’s academic peers; they are written to teach the reader.
We Americans are the great innovators and inventors, of course, but such jobs
generally are restricted to a limited number of research houses, facilities, and
universities. The earning of our daily bread as software professionals is becoming
more and more difficult as the jobs are going away.
Yet, we are able to produce software of the highest quality, and at a lower cost,
here in this country than anywhere else in the world. At present, based on my
experience, the Federal Government pays between $350 and $800 per commented
line of code for ground systems (the command and control software for aerospace
and military applications). And for on-board flight systems—the software that
is used in missiles, satellites, aircraft, spacecraft, instruments, and smart-guided
weapons—the costs are much higher, ranging from $450 to $1,200 per line of commented
code. Such systems are the very reasons for our global military superiority.
In contrast to these staggering costs, my colleagues and I here at the Jet Propulsion
Laboratory have been able to reduce the cost of ground systems to about $10 per
commented line of code on our most recent systems. During the past 20 years, my
teams and I have produced high-quality software at a cost ranging from $10 to $50
per line of commented code—depending on the mathematical complexity of the
application, the availability of the technology in the marketplace, and the schedule
to which we had to work. |
附件: 您的会员级别暂时无法下载附件，请注册成为会员或者提升论坛权限，相关信息请参考新手指南！
|
