Software Project Estimation
Moderator: General Moderators
Software Project Estimation
Look a this page - http://www.tribalsmile.com/nvc/
Do you agree with its calculations?
Do you know where to find more quality links of information on this subject?
Do you know some kind of software for software project estimation?
Do you agree with its calculations?
Do you know where to find more quality links of information on this subject?
Do you know some kind of software for software project estimation?
There are 10 types of people in this world, those who understand binary and those who don't
-
alex.barylski
- DevNet Evangelist
- Posts: 6267
- Joined: Tue Dec 21, 2004 5:00 pm
- Location: Winnipeg
Re: Software Project Estimation
There are no hard and fast rules to project estimation...software isn't like cutting grass where you can estimate by calculating the area and dividing that by a constant time. Unless cutting grass was more dynamic, like you had assume the engine would break every 5 minutes and may take upto a week to fix it. Or the ground you worked on was actually moving and changing all the time - moutains popping up and lakes appearing, etc...
Obviously you know this, that was more for anyone who finds this thread and assumes software estimation is easy - cause it ain't!
Anyways, I run sloccount against my project daily and compare the estimates with my own daily journal (using COCOMO). The results are now off by about 3 weeks, but I estimate that those three weeks have been the amount of time I've spent re-factoring the source code, so if you ignore that fact (mostly spent fine tuning a framework) the estimates are pretty damn close - probably within a week or few days at best. Not bad!
As for a tool which will predict timelines...I have no idea. What would you feed it for input? Specs perhaps...I think that would be horribly inaccurate without more variable input...for example the time it takes to implement a single CRUD view w/ listings when using a framework is significantly less than doing everything from scratch. Likewise for the interface, etc...
Those are all variables which would need to be plugged into the final equation...all of which are subject to change based on your development environment, experience, etc...
Obviously you know this, that was more for anyone who finds this thread and assumes software estimation is easy - cause it ain't!
Anyways, I run sloccount against my project daily and compare the estimates with my own daily journal (using COCOMO). The results are now off by about 3 weeks, but I estimate that those three weeks have been the amount of time I've spent re-factoring the source code, so if you ignore that fact (mostly spent fine tuning a framework) the estimates are pretty damn close - probably within a week or few days at best. Not bad!
As for a tool which will predict timelines...I have no idea. What would you feed it for input? Specs perhaps...I think that would be horribly inaccurate without more variable input...for example the time it takes to implement a single CRUD view w/ listings when using a framework is significantly less than doing everything from scratch. Likewise for the interface, etc...
Those are all variables which would need to be plugged into the final equation...all of which are subject to change based on your development environment, experience, etc...
- Christopher
- Site Administrator
- Posts: 13596
- Joined: Wed Aug 25, 2004 7:54 pm
- Location: New York, NY, US
Re: Software Project Estimation
If I understand it correctly, "delete all users from the database," would have an N-Factor of 5 and a V-Factor of 100. So the NV-Product would be 500. According to the system:
- The number of lines of code would be 500,000
- The number of person-months would be 50,000
- The dollar amount that your company will spend will be $5,000,000
That sounds a little steep to "delete all users from the database,"
I think the basic noun/verb idea may have merit, but me thinks the multipliers may be a little off. Say you took the sentence given ... what multiplier might actually make sense? First how much more weight do verbs get than nouns (they say 10 times). And then what are realistic multipliers for what is probably more of a 1 day job.
- The number of lines of code would be 500,000
- The number of person-months would be 50,000
- The dollar amount that your company will spend will be $5,000,000
That sounds a little steep to "delete all users from the database,"
I think the basic noun/verb idea may have merit, but me thinks the multipliers may be a little off. Say you took the sentence given ... what multiplier might actually make sense? First how much more weight do verbs get than nouns (they say 10 times). And then what are realistic multipliers for what is probably more of a 1 day job.
(#10850)
Re: Software Project Estimation
$5,000,000
Muahahaa
Didn't make these calculations - but sounds really good
OK, seriously - when I have a project I estimate how much every module person-hours will take (by experience), then multiply it by 2 to cover the time for recoding due to bad/new user feedbacks/requirements.
Muahahaa
OK, seriously - when I have a project I estimate how much every module person-hours will take (by experience), then multiply it by 2 to cover the time for recoding due to bad/new user feedbacks/requirements.
There are 10 types of people in this world, those who understand binary and those who don't
- Christopher
- Site Administrator
- Posts: 13596
- Joined: Wed Aug 25, 2004 7:54 pm
- Location: New York, NY, US
Re: Software Project Estimation
I do something similar, but I do thing the noun/verb thing is interesting. It could potentially catch complexity that was not apparent.
(#10850)
Re: Software Project Estimation
In fact, I was searching for object hierarchy modelling software, which extracts the information from the user requirements in the same manner this article describes. But it really looks interesting - the problem of estimating the cost of a project, quickly and relatively simply, is difficult to solve.arborint wrote:I do something similar, but I do thing the noun/verb thing is interesting. It could potentially catch complexity that was not apparent.
There are 10 types of people in this world, those who understand binary and those who don't
- Ollie Saunders
- DevNet Master
- Posts: 3179
- Joined: Tue May 24, 2005 6:01 pm
- Location: UK
Re: Software Project Estimation
Anything based on that stuff is usually rubbish. Why? Because you can verb a noun and you can noun a verb. So are we talking browse or browser? Running or runner? Front controller development or development of front controller?but I do thing the noun/verb thing is interesting
Re: Software Project Estimation
uh, just to point this out...
Maybe my sarcasm detector is broken, but you all sounded pretty serious. I'm pretty sure this article is meant as a joke. I mean, c'mon... I about died laughing when I read
The Article Linked wrote:The NVC Method was first presented to the "Software Project Planning and Management" course at Oakland University, Rochester Michigan, in the Spring of 1998. It was then posted to the comp.software-eng newsgroup on April Fools Day, prompting numerous replies from experts and interested persons throughout the industry.
Maybe my sarcasm detector is broken, but you all sounded pretty serious. I'm pretty sure this article is meant as a joke. I mean, c'mon... I about died laughing when I read
How much more absurd could you get?The Advanced NVC Method picks up where the Basic NVC Method leaves off, by factoring in adverbs and adjectives. For example, if the requirements document says to "hack into the competition's secret database quickly," you'd have one adjective (secret) and one adverb (quickly).
-
alex.barylski
- DevNet Evangelist
- Posts: 6267
- Joined: Tue Dec 21, 2004 5:00 pm
- Location: Winnipeg
Re: Software Project Estimation
I wasn't being sarcastic, nope... 
- Ollie Saunders
- DevNet Master
- Posts: 3179
- Joined: Tue May 24, 2005 6:01 pm
- Location: UK
Re: Software Project Estimation
I never read it. Although I do now feel silly for bookmarking it in my del.icio.us which is published on my facebook profile. 
Re: Software Project Estimation
I've found this site: http://www.itprojectestimation.com/
Indeed, I can't agree that the N-Factor and the V-Factor are chosen properly, but it sounds good.Humour (or is it?)
The NVC Method of Software Project Estimation is a radical new approach that takes considerably less time and is probably just as accurate as many more sophisticated techniques.
There are 10 types of people in this world, those who understand binary and those who don't
- Christopher
- Site Administrator
- Posts: 13596
- Joined: Wed Aug 25, 2004 7:54 pm
- Location: New York, NY, US
Re: Software Project Estimation
Yes, it is a joke. Just look at the date -- April Fools Day.Chalks wrote:Maybe my sarcasm detector is broken, but you all sounded pretty serious. I'm pretty sure this article is meant as a joke. I mean, c'mon... I about died laughing when I read
(#10850)
Re: Software Project Estimation
PS: It's not a joke.
PPS: I couldn't find any evidence for it in Google, so I suppose it's just here, in Bulgaria
There are 10 types of people in this world, those who understand binary and those who don't