Sign up

The Greendizer Blog

News and thoughts by the people behind Greendizer

Open Domain Directory

03/12/2012 8:57 p.m.

We're one of those businesses where the sign up form cannot be sized down to a couple of fields. We decided that if we couldn't act on the number, we would try to make it smarter and try to guess as many values as possible from the context.

We were surprised by how much information could be found in a simple email address. For example: is very likely to be someone working at Springshield, a France-located business. With a few lines of javascript, we've been able to reduce the sign up effort of our users by 1/3. 

Ironically, a big chunk of the information needed to sign up on Greendizer can be found publicly - every business has its name and full address available somewhere out there on the web. The problem is the information is not complete, accurate, verified or standardized in a single format which can be used by anyone.

Of course, we could have used an API like LinkedIn's but we didn't want to make the process even longer by asking the user to authorize an intermediary service to access his profile.

We also considered the WHOIS records, but spammers and marketers forced the registrars to protect the privacy of the domain owners.

We came up with a simple solution and reached out to our community to get some feedback and eventually adopt it. We'd like to improve the web with a free, open, and decentralized domain-based business directory, where every domain-owner - aka business - remains in control of the data he chooses to make public.

We think it's achievable with a simple manifest file located in the root of every domain (, with a bunch of information about its owner.

"desc":"Invoicing designed for developers",
"address":"15 rue de Chambéry",
  • Businesses would no longer have to update their profiles on every online service they registered in;
  • Developers can speed up sign up processes by using a free, open, verified and up-to-date source of information;
  • Many developers would no longer have to deal with logo uploading, organization profile management, etc.;
  • Caching mechanisms of the web are more effective as everyone would be linking to the same URLs (especially logos).

To make it accessible to browser-based applications, we suggest to make the responses compliant with th W3C CORS specification.

The special Access-Control-Allow-Origin header required for CORS

JSONP is excluded because it's only viable if you are 100% sure that the source source is trustworthy. In an  open environment, there is no way to know if a manifest contains the expected JSON data or a malicious script that your browser would blindly execute.

The CORS specification is supported by all the major browsers (IE8+, FF 3.5+, Safari 4+, Chrome 3+), and doesn't require the resource to be altered to become accessible. Websites like make it very easy to configure a server to return the appropriate headers.

We like the extreme simplicity of the idea. Favicons work pretty much the same way, and proved this kind of approaches can be viable. The question is, how would you improve it?

Update (03/20/2012):
  • The humans.txt initiative proposes a similar idea to reference the "people" behind a website (thanks @hbou)
  • OpenGraph does not offer a solution to the problem we're trying to solve. A metadata is only useful when you've been able to locate the resource its describing.
Mohamed Attahri
Published by Mohamed Attahri
03/12/2012 8:57 p.m.
domain, favicon, directory, whois, odd, cors, business

Live from twitter®

Follow Greendizer on Twitter

Join us on Google Plus®

Find us on Facebook®