Community Involvement for an Open Source Project?

pfleming asks: "Several months ago I began a maintenance fork of some niche software. Essentially, these are PHP/MySQL scripts for real estate offices. The original developer moved on to an incompatible version to what I was using. Upgrading for me and many other users was not the easiest option. Luckily the software is GPL'd and so continued work on the fork is not a big deal. I have set up a site, made it available for download, announced the availability of the fork on Freshmeat and the forums for the original software. Now I have a few people subscribed to the project on Freshmeat, and a few on a mailman list set up for the project. This project has been listed on the GNU Website and other mirror sites but doesn't get much discussion on the mailman list and nothing from the Freshmeat subscribers. There is usually an increase in interest (indicated by a short term increase in site hits) when new releases are announced but this fades back to regular traffic of ~40 visits per day as measured by webalizer after a short period of time. Is this an anomaly? Should I be thankful that there aren't tons of bug reports and feature requests?"

"More questions for you to chew on:

• Is there more interest in a new project vs. one that is more or less mature?
• Is the project too narrow to attract an audience?
• Could the underlying business (real estate) just be too saturated with web sites?

Just what are the secrets to a successful (measured by lots of contributors, etc) project...or am I just not defining success correctly?

What other thoughts does Slashdot have on this subject?"

Comments

look at samba-tng versus samba

[netmask]

Samba-tng forked off of samba quite a while ago. The user base using tng is still very small. There is a flurry of people checking it out each time a new release comes out (Which lately has been due to security problems in both code bases).

The user lists are fairly slow, and there are a few developers on the dev lists. The development is still highly active, but the purpose of TNG isn't as important to most people as the functionality and features of Samba it self. The people who need to the changes made in TNG, will go to TNG. However, the vast majority of people don't need anything beyond what the base Samba 2.x or 3.x code has.

Then again.. I would also say, most people haven't checked out the rad features included in rpcclient with tng.. which makes pen testing windows extremely easy.. Oh wait, so does dcom. :)

It's like an engine

[Rosco P. Coltrane]

I too maintain niche projects, some bigger than others, some more popular than others. Here's how I understand the dynamics of a community around a projects. You have 3 cases :

1) your project is too specialized, you have a smallish community of people who use it, few bug reports now and then, and you end up doing all the work on your project.

2) your project is interesting enough that the community around it grows to a point where most of the improvements come from patches, bug reports ... i.e. bits of work done not by you, but you still end up integrating the changes and act as the only maintainer of the project.

3) your project is very interesting and the community around it grows exponentially. The improvements / bug reports flood you and, essentially, your own bandwidth is not enough to maintain the project. You have to delegate and trust other people, in which case A) you're a shitty project manager and someone else who has that talent eventually makes a code fork and takes it over, or B) you become a successful OSS project maintainer, the extreme case of which, for example, is Linus.

The added fun is that, if you code well while you start the project, it can go from a shitty thing to something of interest, just because the look-n-feel that detracted people from trying it before now attracts more people. That's where all the interest is, see how you can "prime the pump" and build a community around your ideas by doing the initial work, then watch the improvements come already made.

I personally choose to create/maintain projects that I reckon will fall in or near category 2), because I don't want to maintain big projects anymore, with the flood of patches, suggestions and hate mail that comes with it, but I don't want to end up having my name associated with a shitty tarball that nobody cares about either.

Re:Surprising

[W. Justice Black]

No. Because they are PHP/MySQL scripts for real estate offices. Calculate number of real estate offices in the US. Substract those that have a meaningful IT infrastructure beyond a few PCs to type and print contracts.

No kidding. The fact is that big shops will write their own (or implement a large CRM, or...), while the small shops will follow the path of least resistance (like not doing computer-based anything, or keeping it to a minimum). Your best bet would be to set up an ASP (Application Service Provider) to host their data so they can get at it over the web, or set up appliance machines to do this task (or both).

This is one case where GPLed software has little value by itself (because there's few people to implement it in its target market), but you could add lots of value by implementing it (and make money in the process).

Usually...

[SharpFang]

projects smaller than "large" usually consist of one-three maintainers, some 3-4 "minor contributors" who supply hints, bug reports, small handy hacks and good ideas on irregular basis, several "fans" who look for updates, sometimes report bugs or help newcomers by answering questions, and besides that, small, regular flux of visitors who come, maybe ask a question or two, look, eventually download and go usually without ever saying thank you. I've seen that with several projects I participated in, as such "fan" or "minor contributor". From time to time some fan or minor contributor leaves, sometimes a new one finds it and stays. If the maintainer leaves the project though, it dies quite quickly, unless someone else decides to take over and continue the work. That doesn't happen often though.

Look at this from positive side. 1) 40 visitors a day, means maybe 1-4 new sites using your software. 2) No bug reports - probably no bugs so that's very good, isn't it? :) 3) No suggestions, ideas, patches - probably the design is so good that nobody feels need for these.

(of course it could be opposite, after first look people discard it and never think about it again, but... :)

One of good ideas to "exist" on the market is to package your stuff for some major distributions and try to include it - even if not in core of one, then at least in official software archives. So crazy people like me, who look through all packages dselect displays get to notice it :)