HN Explorer

Ask HN: Can you start an original, open source, project without writing a line of code?

3 points by nickmolnar ↗ HN
I'm wondering if you think it can be done. What would be the pitfalls? How would the founder provide value in the early development stages? What properties would the idea require to appeal to developers' self-interest? Has it been done before?

I've been asking around about this and have some ideas about how it could work, and how it could fail, but I'd love to hear what HN has to say about it.

11 comments

[ 3.5 ms ] story [ 32.4 ms ] thread
lethain
Could you be a bit more clear? Are you asking if an individual can come up with some idea, write a blog post about it, and have other people build it?

Yes, that can definitely happen, but only if the original person has capital. Most developers who do opensource work have a bunch of ideas and projects they'd like to be working on, so you're going to have to pay them to work on your project instead. Payment doesn't have to be money, it can be reputation or an exchange of services, but basically you have to have something to contribute.

Otherwise you're asking them to do something for you for goodness sake without actually put any effort into the project yourself. In particular if you want to be the legal or cultural head of the project, people will sniff that funk immediately.

nickmolnar
That's about what I was imagining.

I've had people tell me that setting up the SVN, moderating discussions in a mailing list, twittering project updates, and orchestrating the cultural/public facing side of the organization are the most productive early tasks.

My question is: will people start working on a project without even a line of code, or will there be paralysis until there is something to work off of?

lethain
Although it depends on individuals, I imagine most of the crowd here, myself included, doesn't have any hesitations about designing systems ground up from scratch. Some times you screw up and have to redesign later, but that's part of the process of learning.

That said, it would take exceptional circumstances for me to work on such a project. Programmer time has a high market value, and--personally--I'm still having a hard time not looking at your idea as asking programmers to work on your idea for free.

Project management that your discussing is a non-issue until the project grows considerably. I think for many programmers it would feel like you were just taking credit for their work, rather than providing a valued service. People become project leaders by contributing publicly recognized effort and talent to the project, and I just don't believe the efforts you are suggesting would be viewed as valuable.

nickmolnar
That brings up another important question. How do you find the ideal person to take a leadership role for development? That is where the credit should go, but it's a tough spot to fill.
nickmolnar
Another idea I have been thinking about is the necessity of having plans in place, from the beginning, for if, and when, the project begins to have commercial merit. If everyone knows that 90% of any ensuing corporation will go to the top 6 developers on the project, they are much more likely to commit valuable resources.
anamax
> If everyone knows that 90% of any ensuing corporation will go to the top 6 developers on the project, they are much more likely to commit valuable resources.

You're confusing "my plan will work if other people behave this way" with "other people will behave this way".

Moreover, it's an open source project, so you can't say anything about "any ensuing corporation". You can only make committments for corporations that you control; you can't stop other corporations from getting most of the benefits.

And, who's to say that 6 is the right number? What if a given project takes 10 people to do reasonably? Why will those last four people contribute?

nickmolnar
Fair point. Let's assume that the only goal is to make the project go forward. All compensation and credit should be structured to further this goal.

I guess 100% is the best way to structure it. Possibly even with a disclaimer saying the founder would not create any company out of the source code - but could be hired by anyone who does. The founder's role would then, simply, be an evangelist for the project(s).

I think the last four people would participate in the hopes of entering the top six, but a more open plan would probably encourage more participation. I think the last four would participate with the goal of entering the top 6, just like the dozens of applicants for the winner-take-all Ansari X-Prize.

nickmolnar
Maybe all that is needed is a non-binding proposal that the top 6 developers would, after the product reaches a particular milestone, band together and start a company, dividing shares however they choose.

I just think that having any plan for the future would make developers more likely to invest their scarce time.

anamax
> I've had people tell me that setting up the SVN, moderating discussions in a mailing list, twittering project updates, and orchestrating the cultural/public facing side of the organization are the most productive early tasks.

If there's no code, what "project updates" are you twittering?

Note that all of those things are fairly low cost, so doing them isn't much of a contribution. They may add more value than the same amount of time spent implementing, but there will be far more implementation time.

nickmolnar
I think there is news to report every step of the way. It would start with what needs to be done, and why the project is valuable, then move on to what is being done and what milestones are being hit.

All of these are definitely low value activities; the goal would simply be to reduce administrative overhead (time, not money). That way, contributors would spend more time doing the high value work and less time on BS. The goal would also be to foster a sense of community and make it easier for new members to join and get caught up on the big-picture goals of the project.

there
i see this quite often and it's usually a good indicator that the project is going to go nowhere.

non-developers post on a users mailing list about an idea, some other non-developer joins in with "hey i'll setup a wiki!" and someone else chimes in with "i'll make a logo and a website!" in the end you have all this planning and meta-development for nothing, and no code ever gets written.

look on sourceforge and see how many projects are in pre-alpha/planning stages with no code uploaded and how long they've been there.

a common phrase in the openbsd developer community is "shut up and hack." a developer's time is much better spent writing code and showing it to someone than writing up a technical spec about how something should quite possibly maybe some day be if we're lucky and follow the roadmap exactly.