Comment by DevX101 | Hacker News Reader

Comment by DevX101 | Hacker News Reader

Parent Full thread

I thought I'd take the liberty post a comment from a user in the techcrunch comments. I'm curious to hear reaction from those more familiar with telephony than I am.

<comment>

PhonoSDK (http://phono.com) is a far superior implementation of this same concept that has the advantage of being released almost a year ago - close to 10,000 developers are now using it.

The number of advantages that PhonoSDK has over this is pretty impressive:

- Phono supports enterprise-grade open standards for VoIP and IM (SIP and XMPP). Phono supports HD audio (much better caller experience).

- Phono supports in-browser IM functionality as well as in-browser phone functionality, using the same library.

- Phono is open source.

- There is no charge for Phono -> Phono calls or IMs.

Noticeably absent from this offering is the use of any of the generally accepted standards for telephony or communication. You can call this "VoIP" but but it is that in only the loosest sense of word.

The lack of open standards and the inability to interoperate with systems that suport them make this little more than a toy for developers to play with between real projects.

Enterprise customers won't touch this with a 10-foot pole.

</comment>

Hmm, so Phono is free and open source. It sounds like I'll have to do a lot more than sign up for Twilio, give them my CC#, and drop four lines of JS into my app.

Twilio is frequently aggressive with describing how much work integration is. Literally, yeah, you can ring a phone in one line of code but you really don't want to. It starts getting powerful when you can push full features - tested, with UI/UX, production-ready - in about a day, and you could write a telephony business in about a month, from scratch. Those aren't exaggerations in the slightest.

See my upcoming free guide on productizing Twilio apps for why using the sample code in a live environmwnt is almost always a vewwwwwwy bad idea.

Well now I'm curious. Just this weekend I had a great time texting myself with the official Python and Ruby libs from Twilio in only a few lines [1].

Sure there were rough edges like "which and how many phone numbers do I have access to" and "am I going to run out of credits and cost myself a jillion dollars accidentally" but on the whole it was a great starting experience.

Edit: I'm guessing you mean I need to build my own TwiML wrappers and scripts so that my customers can have a meaningful interaction with the computer rather than simply getting called by a useless one-line proof of concept that can't talk or listen to them?

[1] https://gist.github.com/1102938

Wait for the whole story later, but in general: what happens in terms of user experience if you input a wrong number? What happens if you input someone else's number? What happens if 4chan discovers your site and uses it to input a particular someone else's number 8,000 times? etc, etc

At least for the texting side of things, it's downright trivial. My fuzzer texts me every time it finds a crash, and it takes all of 3 lines of code to do this. Works like a dream.

Looks like PhonoSDK requires the use of PhoneGap for iOS clients. There are no details, but I hope that Twilio doesn't.

Our mobile SDKs are native and do not require another framework.

Reply on news.ycombinator.com