So basically, the folks developing the skills kit ...
... are without skills themselves?
Physician, heal thyself.
Developers trying to integrate their applications with the Amazon Echo have suffered headache after headache for the past week, thanks to mystery problems with Alexa APIs. For several days now, third-party apps have been unable to use Amazon's Skills Kit as expected, and thus are unable to receive voice commands via Alexa …
This post has been deleted by its author
Do you really like fucking about experimenting for hours on end instead of getting work done when the manufacturer already knows the answer and should just give you a changelog or new instructions as a matter of course.
I mean, send them an e-mail with three bullet pointed questions and they manage to answer none of them. Not speaking from experience in the past hour at all. Oh no.
I actually think this is a fair statement. You put heartbeats and alerts on the external APIs that you use. You probably own an Echo which you use for testing anyway. You may want to consider recording a series of statements and playing them into your test Echo daily and checking that you get the right responses back (or can see in your own system that the requests got to you).
Wow, 6 thumbs down [and counting]. Does this mean no one out there checks their software when it's deployed? Some of the projects I maintain use dozens of partner and third party API's and provide many APIs used by others. By not having automated and manual checks to ensure the system is 100% ok we'd have clients wanting to know why we we were caught with our pants down and some of those outages would be likely to be mentioned in a article on here.
Sure it's naff when a dependency is offline or changes but that is life and you need to be as proactive as possible. But someone complaining that this is something they shouldn't have to do staggers belief, even if you have an arranged level of service things can go wrong. Testing and health checks are essential if you want to be able to react BEFORE the pressure is on.
How many enterprise systems do you know that are self contained?
Maybe you are lucky you are to have a problem domain that does not need to interact the outside world. Possibly you work for a company that does it's own weather forecasting, banking facilities and handles currency conversion, provides it's own banking facilities, can plot travel routes taking into account current traffic conditions, checking stock levels at customer and distributor sites (all using different APIs), checking for fraud patterns, ...
This bug means that Echo is not sending your app input where it was before.
I would compare it to having an iOS or Android app that requires input from the OS's keyboard, and suddenly the keyboard isn't working anymore.
If you are saying that developers need to check their skills for such a problem every day, that would be the equivalent of mobile developers having to test any feature of the mobile OS on which they require.
When Apple or Android releases new features, at least they are communicating about it. In this case Amazon is changing the behavior of the Echo without any communication. Its a stretch to blame this on the developers.