We’re shutting down our web component catalog
October 19, 2015
We launched this Component Kitchen site in April 2014 with a web component catalog as its centerpiece. Today we're shutting down that catalog so we can focus on web component consulting and our open web components projects, including the Gold Standard Checklist for Web Components and the general-purpose Basic Web Components library.
Running the component catalog was a great way to learn about building production applications using web components and Google's Polymer library. But we've come to feel that the catalog's utility is limited, and it no longer makes economic sense to continue it.
-
As in any online space, much of what is freely given publicly is, unfortunately, junk. Many registered components are "Hello, world!", or trivial wrappers of a third-party library, or utilities so specialized as to be useful to no one but their authors. That makes it hard to find something worth using.
-
Even potentially useful components are seldom written at a level of production quality. We tried to start a "component of the week" blog post series in late 2014. We would sift through a dozen components just to find one that even worked outside the narrow confines of its demo. This isn't necessarily the fault of the component authors — it's really hard to write a solid web component today. (That's one reason we believe that establishing a Gold Standard for web component quality is a vital project.) We also believe that the web user interface framework vendors, such as Polymer and X-Tag, need to invest more heavily in making it easy to create components that meet that standard.
-
Our catalog pages showed GitHub stars, but the direct GitHub metadata for projects like web components isn't all that useful in assessing a project's quality. A GitHub star doesn't tell you whether anyone has actually tried a project, whether they thought it was good, whether they used it, or whether they're still using it. We have some ideas for how to compile better indications of a web component project's utility, but they would require a larger investment of our resources than we can afford.
-
People want to see a component project's own documentation, not a third-party repackaging of that documentation. We think that's true for any online catalog, including successful ones like npm. When we find a package on npm, we never read the package details there, because we'll have to go to the actual GitHub repository to form an assessment of whether it's worth actually trying the package.
-
GitHub itself could easily improve its own search and ranking features to the point where external catalogs would struggle to add value. We'd rather not compete with them.
-
As a result of the above points, if you want to find a foobar web component, it's easier to google for "foobar web component" than to consult a purpose-built catalog. If you follow the search results to GitHub, you'll end up where you wanted to go anyway.
-
It turns out that a coherent collection of components designed and implemented together is more interesting than individual components from multiple authors. While we had plans to feature component collections, we didn't think the payoff would be high enough. Again, the collection creator is more interested in driving traffic to their collection's own site, and that's probably what the component customer is interested in seeing too.
Retiring the catalog lets us invest more time in the projects that we think matter more. People who want a component catalog can use customelements.io, which has seen many improvements lately. Also, as a service to people who have made deep links to our catalog's component pages, for the indefinite future we will continue to serve up tombstone pages at those URLs that offer a link to the corresponding repository on GitHub.
When we started our catalog, there were only about 40 publicly registered web components — now the number is in the thousands. We really appreciate all of the people who visited our catalog in the last year and a half, who built a great component we could feature, and who took the time to give us feedback or a shout-out on social sites. We're still excited to be working in this transformative web components space, and look forward to sharing our ongoing work here soon.