Talk:COCOMO

COCOMO II

I'd like to start a section on COCOMO II. If no one objects, I'll go ahead and do it.

Brian jaress 00:44, 5 May 2007 (UTC)

Yes, but not a whole new COCOMO II section

Good idea but the addition of a note about it that COCOMO is superseded by COCOMO II and why is enough. Everything else just needs to be updates to COCOMO II.--TW Burger (talk) 00:18, 20 March 2012 (UTC)

Moved above comment from a separate topic to this topic, as it should be 108.67.200.56 (talk) 22:56, 3 October 2015 (UTC)

Last 2 paragraphs

I think the last two paragraphs of this article are badly written (although the rest is great, albeit brief):

"One of the most important observations in the model is that personnel motivation overwhelms all other parameters. This would suggest that leadership and teamsmanship are the most important skills of all, but this point was largely ignored. Researchers would rather create tools.

Personnel motivation is not part of the model. The single most important driver is software complexity, followed by personnel attributes (capability and experience, not motivation)."

The first sentance of the first paragraphs says the observation from the model is that motivation is most important but 2nd paragraph states that motivation is not part of the model! Also, who are researchers? And who "suggest[s] that leadership and teamsmanship are the most important skills"?

It seems to me that the first paragraph is either 1) an observation noted in COCOMO (therefor it surely is then a part of the model), 2) an observation made by others (in which case it should state - with sources - who has made this observation), or (to a lesser probability) 3) original research by the editor (and should not be included). Canderra 16:36, 9 May 2006 (UTC)

NO cost driver is more important that any other. All effort adjustment factors are evenly weighted. DSParillo 18:50, 16 January 2007 (UTC)

Is this method considered accurate?

I mean, does industry actually accept the estimates that this method produces? For small projects? For the project I'm currently working on, even using the most generous estimates plausible with the intermediate method, it reckons it should have taken us 3 1/2 years, and we're approaching release quality now after 3 months. What gives? Why is there an order of magnitude+ discrepancy between the results I'm getting and the results this model predicts? JulesH 14:10, 25 September 2006 (UTC)

Because it is Software Engineering article. And those folks make either sh1tload of money or help their self-reliance by publishing their theories. Amount of weasel words is absurd. Generally, one should not look at precision but get the idea behind this:

Work is divided equivalently among people and lines of code.

Which just has to be off. For example refactoring "steals" work but saves some later, a.s.n. 86.61.232.26 (talk) 12:23, 16 April 2009 (UTC)

Appropriate detail?

This post is related (sort of) to the post made by User:JulesH - people want to know fairly high level information about the subject before going into a detailed description of cost and effort modifiers that are really only useful when you are trying to create an estimate of your own. They are documented more completely in the appriopriate COCOMO model manuals.

All the articles describing parametric software cost estimation should have similar sections:

1. Background / History
2. Theory / Mathematics
3. Appropriate use
4. Controversy of parametric models (perhaps this should be a page of it's own)
5. References!

Applies (at least) to COCOMO, Putnam_model, and SEER-SEM

DSParillo 18:39, 16 January 2007 (UTC)

Accurate

Tho I know this might not be the place to criticize this method, and the article is good. However I agree with others here that the result that comes from the estimation is really odd. We pushed out a project in three man months that is calculated to have taken a whopping 6 years to be developed. The project now counts 30k lines of code (freshly written, yes that is 10k lines a month). We consider this a very small project.

You might consider learning how the model works before criticising. I ran 30K of simple code with no CSCI integration required and got 3.4 months. The factors you choose *heavily* influence the estimate (as they should, obviously...) 214.4.238.180 (talk) 19:25, 13 September 2010 (UTC)

As someone who is briefly studying COCOMO I have to say, this all seems a little absurd. This might best be shown in the quantification of "Team Cohesion." I'm trying to figure out how you put a number on drama that happens within the group. (It's a joke... I know it's more about how well the group knows each other and can communicate, but the same thing applies) —Preceding unsigned comment added by Rotorius.kool (talk • contribs) 16:27, 17 November 2008 (UTC)

Just as there should be a Wiki page on alchemy, so should there be one on COCOMO, and one should try to follow its logic as best as possible.

With that in mind, dimensional analysis fails for this work. If Effort is man-months, then plugging into TDEV formula, the scale factor would have to be in months^1-x per man^x, for an x that is unknown but approximately 1/3. I don't believe this can be science.

The exponents for the Effort and Development Time are more or less 1 and 1/3, respectively, so this entire magilla boils down to: a developer writes 10 lines of code a day, but easily as much as 100 and as little as 1 (according to the assignment of quasi-subjective factors), and development time grows as the cube root of SLOC. 108.67.200.56 (talk) 23:11, 3 October 2015 (UTC)

(By xxx)

What's with all the (by someone) lines in this article? They don't make sense. If it's a note from the original editor, it shouldn't be there and it should be erased. If it actually means something, it should be blended into the text in a less intrusive and more explanatory way. —Preceding unsigned comment added by 75.51.126.182 (talk) 16:08, 19 September 2008 (UTC)

Contested deletion

This page should not be speedy deleted as an unambiguous copyright infringement, because... (your reason here) --115.108.54.16 (talk) 09:25, 10 May 2012 (UTC)

Business, IT and Management students need to know what a COCOMO is, especially if they missed class and the teacher and TA are unavailable116.75.41.49 (talk) 12:09, 29 October 2016 (UTC)