hi. i'd just like to ask, if one thing could be done for the games on here, that use money. could it be made, so that in stead of having a space in the numbers, we either have a , [comma], or nothing at all? so for example. in stead of it reading, you have 1 500 dollars, it should read something like, 1,500 dollars, or even, 1500 dollars. of course, if you go with a propper numbering system, you could drop the longer wording for dollars, and have it presented like this. you have $1,500, or, you have $1500.
you could do the same for the pounds sign, and so on. this would make the synths read it faster usually, and also, makes more sense.
this would go for points as well, so if it said that i have 2 300 points, it should really read, i have 2300 points, or 2,300 points.
hope this can be done, to make reading all the numbers a bit faster.

Hello,

The notation currently used is already a compromise, unfortunately.

Dots and commas aren't properly seen in all languages as being thousend separators. There exists confusions with decimal point. For example, the number 1234 written 1,234 or 1.234 can be viewed as one comma three four five as well as one thousend two hundreds thirty four, depending on the language. Same for the dot which might also confuse people.

IN the speech synthesis area, it's also true that not all synthesisers properly recognize comma-separated or dot-separated numbers :

• Jaws US with eloquence recognize dot and space notation, but not commas; the later are seen as decimal point. Idem for NVDA with eloquence.
• We also tried NVDA with a number of french and english SAPI voices and most of them recognize one, perhaps two, but never the three notations. so we are anyway screwed.
• Don't use any separator at all is bad. Sighted users have already hard time coming on the playroom, and many asked to add a space between thousends when there wasn't.

We have chosen space, and actually unbreakable space, because :

• There is no possible decimal point confusion in any language
• It works well with many of the screen readers tested, without any noticible slow down in speech rate
• and unbreakable space so that there couldn't be any new line in the middle of a number, very important for partially sighted users

IF you are using NVDA with a voice that doesn't support space-separated numbers, I suggest doing the following: add a dictionnary entry like "[0-9]{1,3} [0-9]{3}" and map it to "$1.$2" or "$1,$2" depending on the separator properly recognized by your synthesizer. This is a regular expression entry, so also check the correspondig box. You can also add a similar entry for millions and billions if you want. I haven't tested, it's just a though. In any case, there is no jaws equivalen except that you might try to set jaws so that it controls number pronunciation unstead of being controlled by the voice or unselect that option if it was set (it's somewhere in the settings center).

ok, thanks for the explination as to why it is, as it is. i seem to be having a problem with the expression thing though. i can get it to seemingly match the numbers that are there, but how do i make it speak the result, without replacing numbers of my own? is there a variable or variables i can use, to pull the numbers from the actual match?
don't know much about this kind of stuff, and from looking on google, i got even more confused, if anything. lol.
perhaps, if you could, give me more of an idea of the kinds of number matches i could use, and how to make them speak correctly in the result box.
thanks.

Sorry, but I'm not expert in NVDA so I don't even know if you are allowed to reuse matched subpatterns in the result. you may try $1 and \1 in the result, and \d interchangeably with [0-9] for the match.