Awesome summary of the many steps involved in the publishing process. In particular I liked the part about multiple attempts to explain concepts... trying things until finding the one that works for beginners.

In case you're looking for a link to Joe's 2013 "Prags" book. Here it is: http://www.amazon.com/Programming-Erlang-Concurrent-Pragmati...

That link doesn't seem to work. Either way, Isn't it this one? http://pragprog.com/book/jaerlang2/programming-erlang

I was looking for the book, but your link didn't work for me, this one did:

http://www.amazon.com/Programming-Erlang-Concurrent-Pragmati...

Interesting post, but I feel it might do with a rewrite. It has quite a few minor mistakes and odd formulations -- it feels a bit like a first (or second) draft -- not a final post.

Now, for a normal blog post that isn't such a big deal (in the sense that publishing an interesting idea usually trumps simply not publishing it, due to never getting it "done") -- but in this case it doesn't reflect very well on the author -- and that reflects negatively on the book (unfairly, I think -- AFAIK Armstrong usually writes well).

Or am I the only one that gets that impression from the post?

Actually, Joe Armstrong is known to write quite raw and his posts tend to get edited as time goes on. Joe is a gift to the programming world. Just read his writings and make the mental substitutions as needed. Or submit errors to him.

I could swear I've read all of that before. I think this is a rewrite. =X

Unrelated perhaps, but wanted to say that I look up to Joe as a role model and an example of a good programmer.

At his age a lot of people have switched to doing something else. But he is still programming, exploring, learning, is active in the community.

He would post questions to Erlang mailing lists like "how do I do consistent hashing" or "anyone know of why websockets do this or that", I think that is pretty cool.

I was surprised by that too, when I first subscribed to the erlang-questions ML. Recently he started a thread, "How would you implement a blob store?," and I thought it was so cool that the "BDFL" (so to speak) of the language was kicking off cool practical CS discussions like that. Maybe I shouldn't be surprised, it's just not something I see a lot of other BDFLs doing.

I find him to be like a child in all of the best ways. So many master programmers never seem to stray far enough outside of their comfort zones to have to ask basic questions about anything. Armstrong always seems to be operating at the edges, trying to come up with something that's never been done before, but definitely doing something that he personally hasn't done before.

All of the old Ericsson Erlang guys seem a bit like that, though:) The right people were in the same place at the same time.