Matt, Ben, et al, On Tue, Apr 2, 2013 at 6:41 PM, Matt Mahoney <[email protected]>wrote:
> > People are really bad at predicting costs of major projects, > especially software. The way that most people estimate lines of code > is to take their available funding and divide by $100 per line. Maybe > you can do better if you compare your program to something of similar > complexity already written. For AGI, that would be your DNA. If you > compress it, and compare it to the compressed size of typical code, it > comes out to 300 million lines, or $30 > billion.<https://www.listbox.com/member/?&;> Software cost prediction has become one of my strong points. I have a sure fire method that has worked well for many projects, and has yet to fail. There are MANY ways of estimating software, e.g: 1. Comparing it to past projects for which the costs are known. 2. Guesstimating lines of code. 3. Expert developer's WAGs (Wild Ass Guesses). 4. etc... Statistics gained from SUCCESSFUL projects show that programming is ~1/7 of the entire project, with ~1/2 of the project being to clean it up after it seems to "work". My preferred method is to estimate the project three different ways, adjusting the numbers until all three ways match. There is a hidden trick in this, in that screwball guesstimates are almost always based on screwball presumptions, that if applied to a different way of guesstimating blows them completely apart. So far, despite my best misguided efforts, I have yet to be able to come up with THREE bad guesstimates that all agree with each other. One trick to keep people "honest" is to count programming as zero, and guesstimate the remainder of the project - and then add in another ~1/6 in for programming. For those who are transfixed by programming, I let them do their estimates, and then multiply by ~7, depending on their definition of "done", which then become one of my three guesstimates that must be made to match each other. I think Matt's numbers are too high, and Ben's numbers are too low. I am willing to work with anyone here to try to come up with a good guesstimate for their AI/AGI project. Steve ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
