Okay, so going forward with the project and asking for feedback was absolutely the right thing to do. I've met some really awesome people and there's a lot of discussion going on.
So, a very important observation is the following:
- You can't copy Ravelry 1:1 - it just won't work.
- forums, groups and pages (they're intuitive to use, not too fancy and the traditional format simply works without stressing people out)
- projects (giving people the ability to upload pictures, write a bunch of notes and record data about their projects is pretty universal)
- profiles that let you link your blog feed
- Implementing these is not the hard problem that needs to be solved in order to get a "Ravelry for sewing".
- If you start this way, you'll be wasting a lot of time building all these things and then, after months of battle, you face the hard problem. So, now you're at a point where you poured all your heart and soul into building something and into reaching out for people to use the site - but it didn't take off yet because it's not much more than just another forums/groups/projects site. People simply aren't as impressed as you hoped them to be. You become disillusioned and the project fades away... It's a trap.
- The fact that people enter and share craft-related data freely and improve the entries in the database makes it possible to successfully search for things that would, otherwise, be nearly impossible to find. You can easily find "green yarns that are made of alpaca and polyester". You can easily get a huge list of patterns for "seamless, fingerless gloves", or, for that matter, lots of other much more involved queries.
- You can organize your stash and assign stashed yarns to your projects. This helps you keep track of what you're working on and what items are available to use for new projects.
- Through the database, Ravelry is able to give you smart suggestions about patterns others have used with a given yarn that you have stashed, or suggest yarns from your stash that could be used with the pattern you're looking at.
- And Ravelry does so much more on top of that.
Okay, so why I'm saying all this: I'm telling you what the hard problem is. It's building something that makes it possible to search successfully, that lets you keep track of what you're working on and what items are available to use for new projects. It's something that gives you smart suggestions.
And in this point, any "Ravelry for XY"-project will differ substantially from Ravelry. To enable smart suggestions, you need to build a system that understands the data by structuring it in the same way people think about the data. There's a word for that which is floating through the web since years: "Semantic Web".
And now comes the insight: People think about fabric in a completely different way than they think about yarn. So when you build a system that's supposed to give smart suggestions and answers, you engineer the database for that system with the questions in mind that people want answered:
- "What fabric from my stash can I use with this pattern?" (Here, the system must check which stashed fabrics have sufficient yardage to cover the fabric requirements of the pattern.) The converse "Which patterns from my stash can I use with this fabric?" works in a similar fashion.
- "What fabrics have others used when they made projects using this pattern?" "... and how happy were they with the result?"
- "Wow, I like this fabric, has this fabric recently been sold by a fabric store in my vicinity?"
- "Where is the location closest to me where I can take sewing lessons?"
- "Where can I shop a blue seersucker fabric that is 100% cotton?"
- There are countless more.