Yahoo Pipes End-Of-life Announcement

One thing not mentioned in this thread is that Pipes 2.0 was a rewrite from the original Pipes which was written in Perl. We ported it to the same engine that powers YQL.

As for why shut down Pipes if it shares infrastructure with YQL? Probably because usage is mostly RSS crawlers that no one looks at anymore.

Argh! The UI was the innovative part. The whole point of something like this is that you don't have to learn a bunch of syntax - when you add a module, it tells you what sort of data it needs to know (such as a URL or...) and when you connect modules together it tells you what sort of inputs and outputs are available. This allows you to focus on the actual problem domain instead of the morphology of the query language, which is not important to the user.

I keep encountering this attitude among programmers - 'I type code all day, so why don't you want to do the same thing?' If people wanted to type stuff all the time the command line would still rule, but the majority of people prefer graphic user interfaces because they don't want to write code, they want to make selections from available options and use the computer to take care of the plumbing. This is a Good Thing. It's easier, more fun, and yields greater productivity, for the same reason that handing someone a box of Lego is better than handing them a few blocks of wood and a set of carving tools. The existence of Lego hasn't led to a collapse of sculpture or mechanical engineering. On the contrary, the simplicity, consistency, and interoperability of the different components has made a massive creative, educational, and commercial success.

I don't mean this as a hit on you personally, but I find comments like the above frustrating. Imagine you went into a police station to report the disappearance of your bicycle and the person behind the desk shrugged and said 'you're still able to run around on your legs, what's the problem? Your bicycle may have allowed you to go faster down the street but it's not like you could ride it up and down stairs so you should be happy with walking on your feet.' The problem is the loss of a tool that offered substantial efficiencies on some common tasks even though it wasn't ideal for every situation. You'd probably be equally pissed off if your favorite high level language went away and you were told you could just go back to writing assembler. I personally love doing things in assembler because it makes me feel smart but that doesn't mean it's better than high level languages.

I also thought that pipes queries were amazing, but slow, so it was hard to use them as part of an application.

I would love to read a blog post about how exactly it was mismanaged. It would be great for others to learn lessons from the mistakes.