Talk:Function point
Bold text'''
Bold text
<nowiki>Insert non-formatted text here</nowiki>Other alternative methods should be mentioned here or otherwise in Software metric
Alternatives: Lines of Code, Mark II, COSMIC FFP, 3D, FP+
merge with Software metrics
Function points are a little past their prime. They were bigger in the 80's and 90's but the frequency of mention in literature is almost vanished. I would make this one of the historical methods used for the software metrics article.BillGosset 01:27, 11 June 2007 (UTC):In comparison to which methods?
- in comparison to having experienced project managers provide subjective (but detailed) estimates. Seriously, they aren't much more accurate. SPend your time measuring the benefits (those are much more uncertain, anyway)BillGosset 02:16, 29 June 2007 (UTC)
- Project managers? I think the metric is much more a tool for the supplier of the software to indicate to the buyer how much the software will cost. It has the same purpose as the KLOC method.
- Van der Hoorn 12:22, 29 June 2007 (UTC)
- What do you mean "Project managers?". You haven't heard of them or you don't know why it is relevant here? The measure of any tool like this is to compare it to experienced human judges - project managers being one of that set. If it predicts real outcomes better than experienced human judges, it adds value. If not, it doesn't - regardless of who uses it and why. And any line of code based method has the same rules. If you are counting lines of code to estimate costs, it has to be better than other methods to be an improvement. Furthermore, KLOC only applies after you've written the code so comparing it to FP is apples and oranges. The only questions you have to ask is why you need to know (what decision will this information change), does it reduce error compared to human judgement, and what is the cost of the proposed estimation effort (in this case, a FP counting effort)?. I did a study for a client in the UK in 1999. I found that FP counts provide no better estimates on average than their best project managers subjective estimates.BillGosset 20:04, 29 June 2007 (UTC)
I did a Google "view:timeline" search (go to google labs, experimental searches) at {http://www.google.com/views?q=%22function+points%22+view%3Atimeline}. I found that references to "function points" rise in the 1990's and peak at 2000, but by 2004 shrunk to a fraction of that. IFPUG must not be marketing very well or, perhaps, people are finding function points just aren't worth the effort. I know its been a few years since I've seen anyone use it. Hubbardaie 20:13, 29 June 2007 (UTC)
Some of this text is lifted (probably with no permission) from the SEI text on function points (http://www.sei.cmu.edu/str/descriptions/fpa_body.html). This page should be rewritten.~~~~