question

>sample utterance: >myIntent the date is {month}{day}{year}. Yes, no spaces between slots is not going to work. In general, two slots together with no fixed text between them does not work well with Alexa. It is, however, highly awkward to put text between most of the time. You would end up with: [code]myIntent the date is the {day} of {month} in the year of our lord {year}.[/code] The alternative, in having a discursive workflow, is almost as bad: A: What is the month? U: April A: I heard the month 'April'. If that is correct, say the day of the month. If not, say no. U: Twenty seven A: I heard the day of month '20'. If that is correct, say the year. If not, say no. U: No. A: OK. Please say the day of the month. U: Twenty seven etc. None of these is idea. But it's what you have to work with. Personally, I would just go with AMAZON.DATE. If they are a frequent user of Alexa skills, they will eventually learn how Alexa expects you to say the date and conform to it. Sure, you want to make things as usable as you can. But, ultimately, it's just another computer interface and users learn to adapt to what it expects. [For Starlanes, there are many alternatives to each command. But after experimentation, people mostly end up using the single word invocations. Alexa isn't quite up to true conversation yet.]

Have you tried creating custom slots for days (first, second, third), months (january, february, etc) and years? Then the utterance samples will look like: yourIntent do something {january|Month} {first|Day} {two thousand and fifteen|Year} jjaquinta is right that the parsing works best if there's some (standard) text between slots. But, if you use pretty confined custom slots, it's not so bad. You can then do the conversion to the numeric date format you like on your end. Steve

Sorry... Your utterance samples would look like: yourIntent {Day} {Month} {Year} yourIntent {Month} {Day} {Year} As long as the custom slots are constrained, and there's no overlap in the values, you might get pretty far with this approach. Steve

thanks for the replies. I had tried separating the values with commas -i.e. {month},{day},{year}, and it wasn't happy - I got the same error as with {month}{day}{year}. I guess I didn't try spaces. I'll give that a shot.

No, no, no. Putting commas or spaces won't make any difference. (Commas will cause an error, in fact.) You are missing the point. You need to put something [i]spoken[/i] between so Alexa can work out which slot is which. If you say "blah blah blah", for a "{one} {two} {three}" intent, Alexa has problems working out if that is "one=blah, two=blah, three=blah" or "one=blah blah blah, two=, three=" or "one=blah blah, two=blah, three=", etc. If instead you say "blah wiggle blah wobble blah", for a "{one} wiggle {two} wobble {three}" intent, Alexa will more often return "one=blah, two=blah, three=blah".