This is probably very easy to fix, and I think I agree.

we now have nvda+f twice though.

While this is certainly a lot of info to rattle off in one go, inserting commas would make it a lot slower to get through all the info. Note that NVDA+f twice will display each piece on a separate line, so if you need to review it closely, that's probably the better option.

@Qchristensen, @michaelDCurran, thoughts?

I think having NVDA+f twice now makes this less urgent and does mean that those who want to skim through the information to quickly find one attribute can do so. While those who want to go through each piece of information can do that as well now.

There aren't too many other places where we read numerous pieces of information in one block like this.
Otherwise perhaps a setting in maybe Voice Settings to "separate grouped information with" and have the options of a comma, full stop or space.

@feerrenrut, @Qchristensen, coming back to this, I would tend to agree with @jcsteh in #3283 (comment). Thoughts?

Why not have it seperated with a PauseCommand of given length? If there's not a pauseCommand, we could make one.

We do have a BreakCommand which inserts a pause of specified length (though it's probably not supported by third party synths). Pauses of specific length are problematic with regard to different rates, though.

When I press NVDA +Delete, the three pieces of information are separated by pauses of just the right length. Can we copy the same approach for NVDA+F? I agree that this isn't high priority though given NVDA+F twice.

I agree the specific original case of separating attributes with comma (or other pause) when using NVDA+f is now effectively negated by being able to press that twice to open a window with the information you can move through as you like.

I do think, however, that it would be worth being consistent. As @bhavyashah notes, NVDA+delete puts a comma between pieces of information.

NVDA+TAB (report focus) and NVDA+end (Status bar) both add an extra space in between elements (that doesn't adjust the length of time between reading, at least with eSpeak-NG but could that vary between synthesizers?) Pausing between reading elements specified in NVDA's document formatting when moving through text uses the same (two spaces between elements).

I know there are other examples I haven't thought of
I think having the option to choose how much to pause between reading multiple pieces of information would be useful. I expect there would be those who would be happy of the setting in order to remove pauses as much as possible, and those who would be happy because they could insert a longer pause.

Actually I could see the case for having a longer pause between document formatting changes when proof reading in say Word (which could change to anything), compared to element types on a web page. In such cases, having the one option would enable this to be saved in a configuration profile so it could be varied between programs.

I think a human similar pause would be enough here to meet every user expectation, the ones who don't like longer pauses could set the speech rate to a higher speed and the pause will be shorter anyway then. I guess a komma between the expressions should suffice here, otherwise the work on this would be really huge if you want to make every pause configurable. I don't know if this is worthy to do. Note that you can also stop the speech with shift and resume it again also with shift, so that you already can adjust the pause between expressions manually without changing any setting.

I would also agree to insert commas between each field. Jaws has such pauses for its equivalent script and I remember having difficulty with the absence of pause in NVDA+F when transitioning from Jaws to NVDA.

IMO, with a high rate, it does not lengthen too much the time to get all the information.
@jcsteh do you have a precise example where the length of text spoken with and without comas differ a lot? (specific synth, specific rate) Also, it should be noted that punctuation in OneCore were making too long pauses in speech not so much time ago. But this is now fixed.

Regarding the double space question, I guess it is generally used in NVDA to separate things that should be handled independently by the synthesizer. In some case, a single space is not enough, e.g. when space is used as thousand separator when reading number in some languages such as French or German.