[detaro]

Yes, it is a bit much (I hate that people nowadays use the "curated" label for lists that seem like they just collected everything they could find for a category. Which is fine, but not curated)

I don't think Alexa traffic rankings would correlate all that well to quality.

Grouping them by specialization instead of company name might be interesting, and possibly could be done using some company data set or with text analysis.

[rawnlq]

Score of past submissions on hn/reddit is a good signal.

For example to get sorted submissions by domain on reddit: https://www.reddit.com/domain/blog.codinghorror.com/top/?sor...

Similarly for HN (but not sorted): https://news.ycombinator.com/from?site=codinghorror.com

Maybe someone can automate this?

[pbnjay]

Yes! I was mainly trying to find some proxy for quality that could be easily automated, so these would be a much better start.

[jsnell]

The format Dan Luu used in http://danluu.com/programming-blogs/ is probably the way to go. A short explanation of what the blog is about, why he finds it interesting, and linking to a couple of individual posts as examples.

Though it probably requires a consistent authorial voice. So it'd only work for genuine "this is the stuff I read" curation, not for "send a pull request to get your own blog included".