In COCOMO II effort is expressed as Person-Months (PM). A person month is the amount of time one person spends working on the software development project for one month.
COCOMO2 (and COCOMO, for that matter) give output in man-months. One man-year = 12 man-months (obviously). Two men for one year = 24 man-months.
To convert that to man-hours, multiply by your organization's number of usable man-hours per man-month. Your accounting department can give you the number. (It is usually HIGHLY proprietary, as it is key to your estimating process and your corporate operating efficiency.)
In the United States, the nominal number of available non-overtime man-hours per year is given by 2080 - (vacation days + holidays + sick leave days)*8. That's 52 weeks * 40 hours/week, minus vacation, holiday, and sick leave. You may or may not want to include a few days for Natural Disaster contingency in that. Discuss it with your accounting department. If 12*your usable man-hours per month is significantly greater than the nominal number of available non-overtime man-hours per year, you are being set up to assume significant unplanned, unbudgeted, uncompensated overtime in your project estimate. (Yes, I have seen places that do this. Smart ones don't.)
It should also be noted that COCOMO and COCOMO2 assume that one month is about the same as another. This is DANGEROUS if you are getting close to the end of the calendar year, or if you are trying to estimate (and schedule) a short, small project. In the United States, 1 Nov - 31 Dec is not much more than one month, in terms of amount of work getting done.
Assuming you have chosen your various factors correctly, your math appears correct. It looks as though you are calculating 9530 person-months, or a little under 80 man-years, for your 100,000 lines of code project.
发布评论
评论（2）
来自 COCOMO II.2000.0 模型手册：
那篇论文很长，描述了很多不同的因素。在维基百科页面上
给出了一种名为 Intermediate COCOMO 的表格，其中有 15 个成本驱动因素，也许那个更简单。在 http://csse.usc.edu/csse/research/COCOMOII/cocomo_downloads。 htm 有一个程序可以下载来计算（未经我测试）。
此外，还可以通过 COCOMO Estimation 在线完成项目估算。正在使用 3 个模型 - COCOMO、COSYSMO 和 COCOMO。科夸莫。
From the COCOMO II.2000.0 Model Manual:
That paper is long and describes a lot of different factors. On the wikipedia page
a form called Intermediate COCOMO is given which has 15 cost drivers, maybe that one is simpler. At http://csse.usc.edu/csse/research/COCOMOII/cocomo_downloads.htm there is a program you can download to calculate (not tested by me).
Also, estimation of projects can be done online at COCOMO Estimation. There are 3 models being used - COCOMO, COSYSMO & COQUALMO.
COCOMO2（以及 COCOMO，就此而言）给出以人月为单位的输出。 1 人年 = 12 人月（显然）。两个人一年 = 24 个人月。
要将其转换为工时，请乘以组织每个人月的可用工时数。您的会计部门可以给您该号码。 （它通常是高度专有的，因为它对您的估算流程和公司运营效率至关重要。）
在美国，每年可用的非加班工时的名义数量为 2080 年 -（休假日 + 节假日） + 病假天数）*8。即 52 周 * 40 小时/周，减去假期、假期和病假。您可能想也可能不想在其中包含几天的自然灾害应急时间。与您的会计部门讨论。如果 12*您每月的可用工时明显大于每年可用的非加班工时的标称数量，则您将在项目估算中假设大量计划外、预算外、无补偿的加班。 （是的，我见过这样做的地方。聪明的人不会这样做。）
还应该指出的是，COCOMO 和 COCOMO2 假设一个月与另一个月大致相同。如果您接近日历年年底，或者您正在尝试估计（并安排）一个短期的小项目，那么这是危险的。在美国，从完成的工作量来看，11 月 1 日至 12 月 31 日并不超过一个月。
假设您正确选择了各种因素，您的数学结果似乎是正确的。看起来您正在为 100,000 行代码项目计算 9530 人月，或者略低于 80 人年。
COCOMO2 (and COCOMO, for that matter) give output in man-months. One man-year = 12 man-months (obviously). Two men for one year = 24 man-months.
To convert that to man-hours, multiply by your organization's number of usable man-hours per man-month. Your accounting department can give you the number. (It is usually HIGHLY proprietary, as it is key to your estimating process and your corporate operating efficiency.)
In the United States, the nominal number of available non-overtime man-hours per year is given by 2080 - (vacation days + holidays + sick leave days)*8. That's 52 weeks * 40 hours/week, minus vacation, holiday, and sick leave. You may or may not want to include a few days for Natural Disaster contingency in that. Discuss it with your accounting department. If 12*your usable man-hours per month is significantly greater than the nominal number of available non-overtime man-hours per year, you are being set up to assume significant unplanned, unbudgeted, uncompensated overtime in your project estimate. (Yes, I have seen places that do this. Smart ones don't.)
It should also be noted that COCOMO and COCOMO2 assume that one month is about the same as another. This is DANGEROUS if you are getting close to the end of the calendar year, or if you are trying to estimate (and schedule) a short, small project. In the United States, 1 Nov - 31 Dec is not much more than one month, in terms of amount of work getting done.
Assuming you have chosen your various factors correctly, your math appears correct. It looks as though you are calculating 9530 person-months, or a little under 80 man-years, for your 100,000 lines of code project.