Well, obviously allowing anonymous posting is not a good idea, as it significantly lowers your signal to noise ratio. In fact, there should be some sort of cost associated with creating each online persona, to discourage people from creating throwaway accounts just for the purpose of harrasing others. Although why anybody considers it worth their time to troll is beyond me...

One of the main purposes of an online community is to establish your credibility over time with consistently insightful comments. Thus an archive where you can lookup up the complete history of a subject or online persona is necessary.

I think the best way to proceed is to just start a community, attract whoever you can, and then let the community itself guide (and hopefully help implement) any enhancements.

All of the tech details are important, but don't lose sight of the fact that you need something to attract people to the site. You can't just build out momentum. Sites that only get one or two posts a day get dismissed rather quickly by us fickle geek types as lame. The idea/need for the community should come first with the tech build around it to support it. Good luck!

First, find your community, then start a site for that community.

My wife and I were going through something personal, and visited a well-known community board to discuss the subject with other people with a similar problem. The people who used the board were terrific, but the organization running the board were absolutely awful; one of those communities that tries to fiercely restrict and censor the ongoing discussions based on arbitrary rules that are constantly in flux, while (we suspect) harvesting the email addresses of their subscribers for spam.

So, we started our own board, and a large number of people (nearly 100 in the first few days) joined our community. We've grown slowly ever since.

Why? Well, no unholy restrictions; we simply move poorly-placed posts to the appropriate locations and ask people to reconsider certain material that might emotionally disturb others (it's a sensitive subject that's being discussed).

Oh, and we're not trying to make any money, not even to cover our expenses (which are minimal). In short, we're part of the community we're trying to serve, and since we're not motivated by profit (direct or indirect) this community site has been a very easy and satisfying thing to create and maintain.

Second, use open source software with a good reputation and active development.

We use phpBB, and were up and running within a few hours of deciding to start the board. Once in a while we apply a security patch, and our web host does daily database backups, so it's been smooth sailing for quite a while.

Good luck.

OK, maybe this isn't helpful at all, but it's meant to be.

First things first: You don't want to develop a site; you want to develop a business model that doesn't cost a fortune to set up and has some vague, distant hope of making you money. Lesson 1: A community site is a tool, not a means to an end, unless you're in the religion, politics or *nix-is-better-than-windows evangelical categories.

OK, so a site's a tool. Next thing to consider: Why do you want to pay all the people you mention? You brought up hosing (OK, you'll likely have to pay for that), but also artistic and web-design help, as well as DB management. There are scores of portal-style, slashdot-style and blog-style software packages out there -- go play with 'em a little. Most of 'em have a range of skins/themes you can download that might not be perfect, but they're a start. And do you REALLY want to build a custom app before you even figure out if your community is going to take off? Lesson 2: Quit overthinking -- if you're building a community site, then get something out there and put your energy into promoting it. If there's a market for it, the community will tell you how to make the site perfect because (repeat after me) you'll ask them every chance you get.

The greatest horror a new site -- or any new business -- faces is usually cash flow. Do you have enough money to pay the hosting bills? How long, in your worst nightmare, will it take the site to start making money? (Hint: Take that number, double it, and go up by one order of magnitude.) What are the quick ways to make money from the user base, what are ways to get additional money from users who have been around a while and what are the ways you can get long-term, sustainable income from the site? If you don't have multiple ideas for each category, you are going to hit a cash-flow problem. Lesson 3: Your great ideas are just that -- great ideas. But the thing is, great ideas are a dime a dozen; great execution is what makes people money. Have an execution plan.

This probably sounds discouraging, but it's meant to be just the opposite -- if you can do a little up-front planning, can resist the (very common) urge to overfret the technical details and over-buy from vendors/consultants, and can know in advance how you plan to convert eyeballs to money, then you'll likely see some success. These things are basic; it's just shocking how few people follow the basics. Disclaimer: I get paid to offer advice like this [greg-brooks.com].

You've gotten lots of advice already so I'll just stick to the questions no one else has answered so far. You asked what kind of people to go to. There are hundreds of web design companies around the country (I assume you're US-based) and probably thousands of individual web designers. They charge money to set up sites, so it might be worthwhile to get two or three quotes and see what kind of bucks you'll need to lay out. You can just cold-call some people who advertise in your local newspaper, or you can go to one of the various tech networking events around town; just stand up and announce you're looking for a web designer and they will flock to you like flies to honey. I agree with those who recommend open source, apache/mysql/php based solutions that you can easily modify and admin without paying someone.

If you're really looking to save bucks and start small, put up an ad at the local community college or university in the computer or web design office and get a student. They're eager to add more sites to their portfolio and will probably work for either free or minimal bucks. Many students have grown up with computers and are pretty adept, but make sure you interview them thoroughly and check out some of their existing work. If you find a rising star, it's a win-win situation.

There are lots of sub-$10 a month hosting companies out there and you should choose one with a wide range of bandwidths and colocation options, in case you grow quickly (a nice problem to have). Don't start with a tiny ISP that can't handle your growing pains.

As others in this forum have mentioned, you should focus on the business plan (whether or not this is a for-profit operation) and make sure you have a way to spread the word and draw people to your site. I disagree with those who think there's no more room for new online communities; if you have a great idea, go for it and ignore the naysayers. Their ancestors also said the world was flat and men would never fly. Good luck!

These were the questions I asked. I intended to setup a website for Australian mobile phone users - since no such website really existed.

We spent several weeks researching hosts and experimenting with content management systems. This is what we ended up with:

Hosting - Shared hosting at Hostony.net [hostony.net]. For around US$8/month, we got 1000M of space and around 35G of bandwidth, as well as the services we needed - such as MySQL databases, SSH, FTP, etc.

Choosing hosting was difficult - as not everyone offered all the features we needed - common problems were limitations on SQL databases, small space quotas, as well as setup fees when only on short term bill (we wanted preferably monthly payments so we could pull out whenever we needed if necessary. In the end we settled on 3 monthly payments).

Unfortunately the hosting service was unreliable - with many server reboots, timeouts, and complete outages. After complaining we were moved to a better server, which has delivered better performance (but still not up to scratch).

Content Management - we chose Postnuke and PhpBB (via the PNphpBB2 module) for Postnuke. Postnuke tended to offer the facilities we needed - although did have problems with things such as forum avatars and login sessions.

Domain name - this was easy - US domain name .COM only cost $7.95/year (with GoDaddy [godaddy.com]), or $16.95/year if using registering via proxy (to avoid having to include home address, etc).

The result : http://www.ausmobile.com [ausmobile.com]

My advice? Don't spend up big, as the project may fail, and you will lose money.