Why to study or not study node.js?

[McLaren]

I have revencly experimented with backbone, but did not use in real project yet. WHen I have oportunity I will try at work.

Now considering what to study next. I hear lot of node.js is cool. But really don't understand why should I learn it.

I tried googling, reading stackoverflow and variuos blogs about why node is good. But still don't understand.

One thing - the apache is slow because uses lot of memory because creates new thread for each connection. Node has only one thread.

But then - it cannot do CPU heavy tasks ,because its one thread, so all other requests will have to wait till the high CPU request finishes. Its maybe posible to use another process for CPU heavy task, but I am not sure how it would be done.

ALso with databse. Most of tasks use database. So even if node.js will be fast - the database will slow it down anyway. The user will have to wait till the database query will get results.

Also my co-worker read the book about node.js and experimented with it. But he is not using it at real projects so far - its few months or more since he fisnihed reading node.js book.

Only where we are using the node - for web sockets. This is great thing. But since there is lot of node tutorials, I assume node can do much more than web sockets messages.

So studing and not using it does not motivate me to study, especailly when I saw experienced programer not using it.

Also - when I look at job offers, I don't see the requirement node.js.

Also to replace slow apache - its good nginx.

So the conclution is - not to learn it and wait till it will be used widely or die?

[McLaren]

read more:

"Very nice, but why should I use it?"

One reason is efficiency. In a web application, your main response time cost is usually the sum of time it takes to execute all your database queries. With node, you can execute all your queries at once, reducing the response time to the duration it takes to execute the slowest query.

I think for one request - there mostly are database queries which need to be executed after getting results.

For example:

query1: find then id of manager
query2: get the shops by manager id.

So query 2 cannot be executer without query 1 results and this is most of the time.

Another thing - even if we need those different quries - how the database will make them get faster? It still has only one hard drive which is the limit of speed. I don't know how databse server handles multiple requests, but lets say it returns one request and then returns another request. So anyway - it has to wait till both request finish.

Or I am talking nonsense?:)