Yahoo Pipes could do better
by Paul Browne
Like many other people, I've been playing a bit with Yahoo pipes : Yahoo Pipes in 10 easy steps. It's a very good example of a Web 2 tool . It's beta, and does what it says on the tin ; It allows you to combine / filter / clone and edit RSS streams. In the same way SQL allows you to query a database, pipes allows you to query XML (or to be more precise RSS) streams for the data you want.
Yahoo Pipes is worth checking out for the following reasons:
- The user interface (finally) puts Gmail to shame. Just how do they generate the dynamic / curvy pipes linking the boxes?
- It's completely graphic. Users with at a 'power user of Excel' level can generate streams that would previously have taken an experienced programmer a number of days.
- It's another piece on the Web2 infrastructure. All other desktop apps have migrated to the web. It was just a matter of time before developer tools did as well. Does it make sense for you? Your call.
Pipes, for the reasons below, is not yet going to displace teams of Java people who do nothing but code RSS streams all day. Before, the choice on many IT projects was Build , Buy or use Open Source (or various combinations of those three). Online Web 2 apps (of which pipes is only one example) gives a fourth option to put into the mix.
So what does Yahoo pipes need to overcome the 'toy' label and become a 'serious' option for IT projects?
- The problem is, it's free.How do Yahoo intend making money out of Pipes? More accurately , will they make enough money so that my project can still use it in 3 years time.
- You're stuck with Yahoo.Meaning, if you build against pipes, you're stuck with them. Even in the database world, it is possible, if expensive to switch product supplier. I'd love Yahoo to open source pipes to solve this dilemma, and allow them to build a business around the 'pipes hosting' part.
- It's completely graphic. This is mainly a good thing (see above), but no doubt you'd still like the option (like most developers) to see / edit the generated code.
- It's hard to extend. If there is a way of extending it with my own 'widgets' , then I missed it. I can host RSS-generating code on my own server, but this detracts from pipes overall ease of use.
What do you think - will Yahoo Pipes become more than a toy to Java Developers?
I completly agree. Here what I wrote in my blog:
1 - Sure the UI is great, but wouldn't a script language be more appropriate and possibly richer?
2 - Why is there so much buzz around this and far less around Amazon's incredible web services?
3 - Meta-services like Yahoo pipes have little chance to succeed: there exist very little successful meta-services in the electronic world (is there a single very successful meta-search engine?) as in the physical world. As human beings, we usually prefer to handle our workflows ourselves, or, our workflows are so personal that they are difficult to generalize and create a service from.
4 - Meta-services make most sense in the enterprise market but then you need the full power of workflow engines.
Yahoo! Pipes exists for several reasons...
@David - I'd solve a script language (SQL for RSS?). Amazon's stuff is great (and yes , better than pipes), but don't give the immediate *visual* hit that pipes does. Yes, that's style over substance.
|Can't Type: That's 'love' a script language, not 'solve' ...|
The main issues is you can't edit the code. Why couldn't they just add a simple option similar to wordpress new post to actually view the code...