Pipes!

Wednesday, 7 February 2007

Yahoo! (finally!) released Pipes as a beta today; congrats to the very talented team that put this together.

Niall gives the geeks-eye view, and to be clear, this is not going to be the next great consumer Web site; your grandmother is not going to go out and build pipes.

However, I do think it’s going to be a big wake-up call for the “Enterprise” software industry.

This tool does more to deliver on their promises of non-programmers slicing and dicing data and wiring together workflows than the big software vendors have in the past five years, and it does it using the actual Web, not the lip-service Web that’s in Web Services.

To my eye, the difference — and the power — in Pipes is that the data model isn’t an XML Infoset, it’s an Atom or RSS feed (usually; most of the modules push you in this direction). That gives you more structure and semantics to grab onto and use in the modules, building more value into standard components, rather than having to go and re-invent the wheel for each application, because they all have different formats.

This is critical; using standard rather than application-specific formats gives you a lot of leverage, and is the missing part of many “RESTful” interfaces IME.

URIs are first-class types in the Pipes world as well; they identify things in the system because the data that Pipes works with is on the Web; it has to be.

I’m really excited by the potential here; there’s a lot of work to do, but it is a much more productive, Web friendly tool than anything I’ve seen come out of the big vendors for a long time.

8 Comments

Stefan Tilkov said:

Neat. It seems to me the “visual” aspect draws attention to the wrong thing, though - as if the innovation were the UI. I suspect a “textual” DSL would work much better in the long run, even though it wouldn’t be as cool …

Friday, February 9 2007 at 4:56 AM

pasha sadri said:

Hi Mark,

Thanks for the your enthusiasm for Pipes and your technical and moral support during the project. You sure are right in that there is a lot of work left to do.

Pasha

Friday, February 9 2007 at 5:18 AM

Geoff Halprin said:

Lots of discussion on Pipes at the moment. I’m still processing what it might mean. Thanks for the link.

Is it just me, or does this look like Lego Mind Storms for RSS?

I agree with the need to devolve web programming and mashups to end users (much as spreadsheets did two decades ago). I’m not convinced that this will be it, but then again, it’s an interesting IDE concept for end users.

Apart from cute home pages, I’m not sure how big a system end users will be able to build with it. Then again, the same would have been said of spreadsheets.

Definitely one to follow.

-g.

Friday, February 9 2007 at 11:08 AM

Bill Seitz said:

Does it recognize MicroFormats?

Friday, February 9 2007 at 12:44 PM

Bill de hOra said:

Nice; I monikered pipes as “BPML2.0” in del.icio.us this morning. It’ll stop people trash talking Y! developers for a while.

But you need a followup post, talking about DSLs instead:

“This tool does more to deliver on their promises of domain specific languages than the big software and services vendors have in the past five years, and it does it using the Web of Resources as its domain, not the lip-service object oriented domains that’s present in most enterprise systems.”

Friday, February 9 2007 at 12:59 PM

Dilip said:

Did you know of this post that talked about a similar concept way back in 2003? http://friends.newtelligence.net/clemensv/PermaLink,guid,154.aspx

Saturday, February 10 2007 at 3:53 AM

Hans Gerwitz said:

Beautifully said. I think Pipes is a tool that’s existed in a lot of minds but has never been tangible. To see and touch it makes it hard to ignore that the organic market of the web has done a better job of homogenizing data formats than OurXML committees can ever hope to.

Friday, February 16 2007 at 10:13 AM

www.datagrep.com said:

seems yahoo is now getting serious into web2.0 technologies and giving google a good run for the money

~BALA

Sunday, March 25 2007 at 7:22 AM

mark nottingham

other HTTP APIs posts