Scratch

I love this idea, but I think that any algorithm trying to find "most relevant projects" not by the creator would still have to take popularity into account because:

a. Higher-quality projects tend to have more views. This correlation doesn't always hold true, but in general I'd like to think that worthy Scratchers almost always get noticed sooner or later.

b. People want to see higher-quality projects. Using an algorithm to dig through the database of projects to find, say, the projects with the most tags in common with the current project will only worsen the problem of tag-spam (and I think JTxt's FaveTags proposal addresses this to some degree). People don't necessarily want to see the project that, according to a computer, most closely resembles the one they're on -- they're going to want to see projects that are higher-quality. Unfortunately the easiest way to rate projects for quality is to rank them by views, love-its, and downloads, which brings us back to our original problem. I think that expanding the curator program might help to an extend, and FaveTags would be a fantastic addition to the Scratch portal.

Anyway, I totally support this idea, I just like to rant.

I support