What I find fascinating about node.js is that it is one of a long chain of products that represent the constant repackaging of old ideas - with better results.
Twisted has been around for a long time and was the best (IMO) platform for building real time apps. Many of the first great comet apps were built on twisted. But you really have to work at using twisted - the docs aren't great, the reactor / deferred patterns take some time getting used to, and finally and most importantly the community isn't effective in evangelizing their product (sorry Glyph)
Another issue seems to be timing. Low latency apps have a current buzz and you see people rediscovering old collaborative ideas like Push to talk (referenced in the article). I wish them luck - PTT has failed spectacularly in the past but maybe will have a strong second life in a world full of smartphones. Node.js seems to riding this wave - and the backers leveraging the buzz to drive interest in the platform.
I also wonder if location is an issue. The Twisted team had a strong core in Boston while the Node interest seems to be coming out of the valley - and leveraging that environment's idea pressure cooker.
It also seems that node.js would benefit from a good implementation of the generator pattern that would enable you to write asynchronous code in a (somewhat) iterative style. Unfortunately I don't think that V8 implements any of the ecmascript extensions pioneered by Mozilla that implement yield, list comprehensions, and other fun stuff.
Comments
What I find fascinating about node.js is that it is one of a long chain of products that represent the constant repackaging of old ideas - with better results.
Twisted has been around for a long time and was the best (IMO) platform for building real time apps. Many of the first great comet apps were built on twisted. But you really have to work at using twisted - the docs aren't great, the reactor / deferred patterns take some time getting used to, and finally and most importantly the community isn't effective in evangelizing their product (sorry Glyph)
Another issue seems to be timing. Low latency apps have a current buzz and you see people rediscovering old collaborative ideas like Push to talk (referenced in the article). I wish them luck - PTT has failed spectacularly in the past but maybe will have a strong second life in a world full of smartphones. Node.js seems to riding this wave - and the backers leveraging the buzz to drive interest in the platform.
I also wonder if location is an issue. The Twisted team had a strong core in Boston while the Node interest seems to be coming out of the valley - and leveraging that environment's idea pressure cooker.
It also seems that node.js would benefit from a good implementation of the generator pattern that would enable you to write asynchronous code in a (somewhat) iterative style. Unfortunately I don't think that V8 implements any of the ecmascript extensions pioneered by Mozilla that implement yield, list comprehensions, and other fun stuff.