Ask HN: How do i determine how much software development will cost? | Hacker News Reader

Ask HN: How do i determine how much software development will cost?

4 pointspoof13 comments

I'm starting a business to create a software/hardware product and i need to estimate how much the software will cost to develop. I dont have any experience coding to know how to determine this and no idea where to get this information. Any ideas?

View on HN

You might try shopping your idea around some place where programmers congregate -- like, ohhh, HN? -- and get feedback on the specifics.

Don't worry about anyone stealing your idea. Anyone good enough to pull it off is already too busy with other projects.

I'm trying to build a product like autodesk's flame but a bit different. They sell this for $300k/unit, so i'm wondering how much it actually cost them to make it.

http://usa.autodesk.com/adsk/servlet/pc/index?siteID=123112&...

Edit: To elaborate, flame is pretty much like an adobe after effects + premiere hybrid.

Whoof. I support any effort that competes with AutoDesk -- they're an abysmal company that go out of their way to encourage people to compete with them -- but you're picking quite an ambitious project here.

If you want to produce a closed-source product, your programmers will be somewhat limited in the various open source projects that they can use as an effective foundation for this.

I'd say, purely from a software development standpoint, to produce a version 1.0 of something that might have a chance of competing with Flame, you're gonna need: a good UI guy, a brilliant 3D hacker, a pretty sharp math guy (fortunately they're cheap these days :-), a good programmer, and an experienced project manager with some respectable coding chops just to keep the project moving along. So, say a minimum of $250k combined yearly salaries. If you're a strong businessman, and if you get the right team together, you might be able to get a beta released in 12 months.

That estimate is high or low depending on how much higher or lower your goals are in terms of a first release, but it should be kind of ballpark.

Thanks for the estimate. I figured it would take a team of 50 programmers @ $100k/yr salary 2 years to complete it (which is how long i want to give it), which would equal a total cost of $10m, which i imagine would be reasonable considering it wouldnt take much to recoup.

Seriously, it would only take like 5 people to write a program like flame from scratch?

Well, you can do it with only 1. But it would him/her several years and a LOT of effort. I think parent was saying that to get a "functional" team, a minimum of 5 would be needed.

Even with the 5 people, I think the estimate of 1 year is a bit extreme. Unless you get brilliant programmers that can work together really well, I'd imagine they would need more time than that.

I don't know what the product is, but to be able to compete with a mature product(even a bad one) takes a lot of time and energy. It all depends on what you want version 1.0 to do.

It's extremely difficult to estimate the amount of work a software project is going to take. If you're working with an experienced developer, and have a good specification of the software you're building, they will probably be able to give you a ballpark estimate.

No matter how good the ballpark estimate is, it won't be right on, either in terms of time or money. Every piece of software you build is, by it's nature, unique. If it wasn't, you'd just license the existing software.

I'm not working with a developer at the moment, im just in the planning stages. If i were to look for a developer to consult me on this, what sort of qualifications would i look for?

I think the main question would be do you currently have access to any resources that you plan on using to create this software? Do they have any experience in this field, or is this going to be a learning experience for them, just like it would be for you.

Estimation is very hard, my suggestion would be to try and find what is the minimum viable product that you want to produce, because in all likelihood you aren't going to be able to duplicate every feature that Autodesk has, and find some developers who have worked on this type of work to give you some idea of time. Even then figure that their estimates are going to be off and add some buffer. But still somewhat of a crap shoot because stuff always pops up.

Derek

For a big application like that, I would hire someone to write up a draft spec so that programmers can have an idea of how much effort it would take.

Writing a draft spec wouldn't cost you that much(a few k perhaps) and you would have to do it anyway. I think of it as a feasibility study.

Where and whom do i look for to write a draft spec?

If you don't know any freelancer local to your area, perhaps post here asking for help and contract it out for the person you "click" with.(most probably this person will be your next project manager)

I would say that you need an experienced programmer or a project manager type(with good programming skills). Someone who understands how to write code but is experienced enough to communicate with anybody else on a higher level.

I wouldn't personally recommend any of the project bidding sites out there for this project(although I've used them personally as seller and buyer before).

No experience, no knowledge. Sounds like a perfect plan.

Clearly it will cost more than the retail price. One would think at least x10. So therefore $3M

maybe COCOMO II or COSYSMO to give you an effort estimate. They're meant to be calibrated to your environment though.

Or a book like this: DeMarco, Tom. Controlling Software Projects: Management, Measurement and Estimation. ISBN 0-13-171711-1.

Reply on news.ycombinator.com