Pipe Dream

[anavish]

Well fuck me gently: it's here. The programmable web. I mean, come on, everybody knows the web is a series of tubes pipes.

For all you skeptics from the last post: this is what RSS and tagging allows us to do, OK?
Yes, this is a first and hesitant step, but it doesn't take much to see where it leads.

Comments

Is this supposed to be news?

[talash]

How is this fundamentally different from, say, Greasemonkey? And then, greasemonkey isn't something new either. It has been predated by Proxomitron by many years. Not that the website isn't neat, but the idea of on-the-fly rewriting/programming of websites isn't new.

Re: Is this supposed to be news?

[anavish]

My, are we anti today.
Yes, this is supposed to be news. Of course Greasemonkey was before (and Proximitron before that), but it dealt primarily with form, not content. Though it's possible to build a content manipulator in Greasemonkey, it's hard and requires intricate coding knowledge (I couldn't easily build you a greasemonkey script that combines two feeds, for example; and I'm a coder).
As opposed to that, Pipes leverages the metalanguage (RSS) in order to manipulate the content of feeds directly, regardless of their form. The closest we had to that was Feedburner, and you can't even begin to compare.

On second thought

[talash]

You're right.
However, I can't get rid of the notion that I've seen something like this before -- and I'm referring to somewhere on the web (and not unix pipes of course). A strong one.

Perhaps it is a sign that the idea is indeed brilliant.

I thought this was around for a while

[ninjarabbi]

Is this different from any other aggregator because it's easier to create and modify? I couldn't be bothered to make a yahoo ID to find out.

Re: I thought this was around for a while

[anavish]

It's not an aggregator; it lets you program with feeds. You can actually do something like "take this feed here, ask the user for a location and filter it based on that, then pipe it through this for-loop which performs a google search for each entry's title, filter away duplicate titles, and give me what came out sorted by date, as an RSS which I'll add to my favorite aggregator". And all this is done by connecting pipes on a GUI. It's quite different from everything else I've seen.

Re: I thought this was around for a while

[ninjarabbi]

Oh, ok, so you basically use the output from one data feed as input into a second data feed, just like pipes in Unix.

Neat.

But what use is it to my grandma?

Re: I thought this was around for a while

[anavish]

"Search for apartments on craigslist in a place I specify and show me flickr photos of the neighborhood." Look at the examples.
I don't really know, but since it's open and you can use any website as a feed (not just the "standard" ones, but anything that supports parameterized GETs) people will start coming up with ideas pretty soon.
...or not. Office 2007 isn't of much use to your grandma either, but it's still neat.