Text-to-speech isn't new but new cloud API services promise to make the transcription of speech to text faster. But how accurate...
The argument for a signaling standardChris Koehncke
I was tickled pink that Google elected for WebRTC not to incorporate a signaling standard (per my previous article). It would only be a matter of time before the telco types started yapping about mapping SS7 to WebRTC. I’ve also learned thru the years that all roads about voice eventually lead to a theoretical discussion about lawful intercept (we all want to be James Bond I guess). Historically, dragging old technology forward is painful and usually short lived (anyone remember SNA over IP?), but the conversations persist.
The lack of a signaling standard means the developer has to create one themselves. It also means they kinda of have to have a clue about what they wanna do (the easy bit) but the hard part is ensuring they’ve thought through all the use cases least they have to go back and re-invent the signaling again down the road.
We, as the proponents of WebRTC, need to think hard about how to expand our circle to include the ROW. The VoIP world is a small secular group with great ponderings such as whether to send DTMF inband or out of band (and eventually elected to do both). Thus the onus is on us to make this easy and accessible to Google’s dream for every modest HTML developer to have some skills with WebRTC. We are indeed making progress, witness the number of #webrtc tweets by those I would consider outside the club (BTW I’m a proponent of just killing off DTMF in it’s entirety, invented in 1963 – it’s had a good ride and deserves a proper funeral).
WebRTC needs a default signaling standard. Not one that some small group in a cheesy hotel bar decreed we must use. But one that the entire community says – “I’m OK using that one“. Clearly those who want to go off reservation should be free to do so. WebRTC thus needs libraries that are freely available, contributed to by the community, time proven to work and for those less inclined, no idea how they really work.
Today, Github is showing 493 repositories that reference #webrtc (jQuery has > 33,000) so we have a little more progress we can make. Perhaps Google should sponsor a code challenge for the creation of support libraries, a token amount for those selected, the honor of being selected would propel all sorts of late night development. A $100,000 might well create hundreds of new libraries for an applications developer to choose from. Who knows WebRTC might be as popular as jQuery at some point?
While the low level libraries do some heavy lifting, it’s not where the real magic is. The real magic is in what the actual application is doing and that’s exactly where we want the app developer to focus. Less time on grunt work, more time on things that move us forward.