Building a more open internet

praise
1

Saw this tweet this morning:

Happy to see people sharing knowledge in places that are:

  1. Open to anyone
  2. Referenced by a URL
  3. Referenced by a URL that is human readable
  4. Open for comment and discussion
  5. Based on Open Source technology
  6. Can be exported

e.g.

  1. Not just in a Slack walled garden interleaved with seven other conversations
  2. Not just in an email with limited recipients
  3. Not just on a mailing list with a random URL
  4. Not just on GitHub or Reddit with a single company in control of the data

But nice that it’s easy to share it there too once it’s posted on Discourse.

This project and others that share it’s characteristics have a real impact on the world. Thanks for everyone who has continued to lead its development, contribute, and adopt it for their communities.

17 Likes
2

I think Google already does this in searches and I suspect this is how they built the dataset:

  1. User enters misspelled word in search
  2. User corrects mistaken word in next search
  3. Google learns to associate word #1 with word #2 over time as this action is repeated

You could test this by searching for variants of the misspelled words, above.

2 Likes