davidg666 wrote:I ask because there's an issue that affects TTS here in Ireland - and presumably also in South Africa, where they use similarly-named roads - that could easily be resolved by making the TTS behave slightly differently.
David, I have been following this thread with some new interest, as we indeed have the same problem. I drive on the 'North One South' and 'North One North' daily, so it is a bit of a thorn in my side, especially since I have seen quite a few URs on this and know I can do nothing to fix it.
We can likely live with the 'N'1 N / 'N'1 (N) or 'N'1 S naming convention if Waze 'tweaked' TTS slightly, but I was not happy with resorting to the 'n'1 N or 'n'1 S convention to work around this.
I'm just curious as to how the name actually looks and displays in the client, and how visible the single quotes are.
Instead of just 'tweaking' this every time, Waze devs should possibly just look at using or defining some opening character and closing character that will likely NEVER appear in any actual names, and use that as a 'literal' indicator, where anything that apears inside the literal is not subjected to testing for abbreviations.
Something simple like the square brackets or even the curly braces. Naming it [N]1 N, or even [N1] N, and they simply drop the brackets in the displayed name, and do not pronounce [N] as 'north'. Even something like ~N~1 N, where the single 'tilde' characters are ignored. I presume fixed 'opening and closing' delimiters would be simpler to search. They could even be used as an indicator that this character sequence should be pronounced individually instead of treating it as a word, like [QE]2 or even [NY] being explicitly overridden as 'Queue Eee Two' or 'Enn Why'.
Something like this would be far simpler to implement in code, instead of having to populate an additional DB field to use as a TTS override using phonetics, where 99% of all existing named segments would likely never make use of the phonetic override.
It just requires some simple string manipulation during Back-End parsing which they likely do already.