Hello! + Announcing First(?) Concurrent WebPipe Block

[Steve Phillips]

WebPipers,

Intro + Vision

Greetings!  I discovered WebPipes just a few days ago and love the idea.  Why?  A few reasons, including:

1. I'm a big believer in machine-readable schemas moving the web forward by making way for much more automated automation, with humans spending far less time digging through docs then manually gluing stuff together.  I should be able to simply tell a web app about another one (by URL or perhaps, eventually, by name), and they should immediately...

    (i) read each other's schemas;

    (ii) auto-generate the types/classes to be used to internally represent each other's schemas as data structures in the programming language in which it is written (perhaps only for efficiency);

    (iii) (perhaps) auto-generate or download each other's documentation;

    (iv) (perhaps) perform diagnostics, by either (a) providing each other sample inputs and outputs (basically externally-verified, programming language-agnostic integration tests) and running those against each other, or (b) auto-generating test requests based upon the schemas and make sure each response is well-formed;

    (v) (perhaps) notify all other web apps they've ever communicated with about each other, and/or adding each other to every registry/directory listing of these sorts of intelligent web apps;

    (vi) (perhaps) add each other's schema, examples, and any other relevant docs to its own human-readable interface, so any technical or non-technical person trying to do something with it knows the full breadth of functionality he/she has at his/her fingertips

2. I've also thought that it'd be awesome to somehow create programming language-agnostic libraries.  WebPipes can totally play this role, albeit in a relatively low-performance way, but this will often be good enough... and if the desired WebPipe is open source, it could be hosted on localhost (or a machine in the cluster), which would speed up the socket communication a bunch.

Two New WebPipe Blocks

Three days ago I created my first WebPipe block, HTTP Status Code Retriever (link).  Today I finished my second one, Web Latency Comparinator, which is only slightly less boring, but at least shows off Go's concurrency primitives a bit.  Both have been added to the registry.

Soon I'll send out my thoughts/opinions/ideas on the spec based upon my experiences writing these apps and relevant ideas I've seen expressed elsewhere.

Hope to talk to you all soon!

--Steve / elimisteve

P.S.  Thank you to Matthew (Hudson) for being responsive and stoked... and accepting my pull requests :-)

P.P.S.  ajvb and I are in #webpipes on Freenode