Joined: 1969.12.31 00:00:00
Location: Watsonville, CA
Here's an email thread between Nicolai Svendsen and myself exchanging initial observations re the new TriadCity talking client, in alpha this week.
I believe I can work with anything. I haven’t had a look at the forums, but I’m sure it’ll work. I’m not overly worried about it. I am a fighter! Of course, some are more difficult to navigate than others, but it depends entirely on the skill of the user as well. Although, I’m lying a little as I continue this paragraph. I did look at it very briefly, and it looks pretty fine. The more people we can get to discuss this the better.
Well, if the popup windows could be automatically spoken, it would solve the issue. Have you tried WAI-ARIA? You probably have already, but I got curious though I haven’t always found it so consistent when it reads. Turning VoiceOver on for that is as you say kind of awkward, but if it can’t be spoken automatically somehow, then I don’t see many solutions unfortunately.
Twitter Cavern sounds fantastic, and no popups gains instant approval of course!
TriadCity with voice commands would be awesome. Being able to verbally instruct it on what to do would be fantastic, and very efficient. Gone will the annoying typos be especially when doing it on your smartphone.
You’re very welcome! I try to be as helpful as I can, but I think the forums are a brilliant idea.
On Oct 22, 2014, at 7:41 PM, Mark Phillips <email@example.com> wrote:
- If we were to add a section to the SmartMonsters Web forums for public discussion of these talking client issues, how convenient or inconvenient would that be for you? I see an advantage to sharing these dialogs with others who are participating. But I’m not sure how easy or difficult the forums are for you to use.
- TriadCity was designed a long time ago with an explicit goal of integrating Web content with the game world. This is where the concept of popup windows originates. We use it in simple ways for signs, like road signs; in more interesting ways to integrate visual content such as images of artworks; and as a way of representing books in libraries - we can link to the Project Gutenberg texts for example. IMO this is one area where we want to go back through the entire game world and really be ruthless about deciding where do popups add real value to the user experience, and where can they be replaced by ordinary text. Obviously if we replace them with text, the speech engine can simply read them and there’s no need for VO. Right now the voice interaction between popups and the main game output feels really awkward for exactly this reason: you need VO for the popups.
- (Aside: I’m writing a “Twitter Cavern” right now where Tweets fly around like bats, and Tweets we disapprove of will attack, and you blast them like monsters. It’s pretty funny! No popups here, just an example of integrating the TriadCity game universe with the external one.)
- The longer run solution to these problems is to write new clients which are *entirely* spoken. For example, nobody’s gonna want to type TriadCity commands into their smart phone, and, lack of mobile clients is holding back TriadCity’s wider adoption in a world which is increasingly moving away from desktops and laptops. I think the right paradigm is to speak your TC commands into the phone and have the entire experience be verbal, for sighted and blind players alike. We’ll get to that next once we feel the talking client you’re interacting with now is ready for release.
Again many many thanks! The feedback you’re providing is already very helpful!
On Oct 22, 2014, at 10:18 AM, Nicolai Svendsen <firstname.lastname@example.org> wrote:
That sounds fantastic! I’ll be looking forward to what you folks have in store for this.
About the spoken text you are absolutely correct. I’d like to have it displayed in the browser window as well so that VoiceOver users can review it if necessary. Whether they should leave VoiceOver on or off is a tough choice. I personally have a Keyboard Command set to mute speech so that VoiceOver doesn’t interrupt or speak over the talking client, although the speech manager handles this a little oddly when speech is not silent. This means that you may hear VoiceOver speak one thing and the client says something else, or VoiceOver is cut off until the client is done talking. Those results vary. Another problem with this approach is that when VoiceOver’s speech is muted, it doesn’t always speak when you turn speech back on. You could also turn VoiceOver off, but that means that when you look at a sign, for example, you have to turn VoiceOver back on to read it. Turning on VoiceOver might also take a few seconds on some Macs.
This question would be very easy to resolve if VoiceOver would remain quiet when you press enter (at least this isn’t the case on my iMac), or if public APIs existed to communicate with VoiceOver. I’m going to mess around with this a little bit more to see what I think. I’m having a difficult time coming up with a straight answer as you can probably tell, though it probably comes down to user preference in the end. If information that opens in popup windows could read automatically, the issue is almost completely negated, aside from the reality that some users might want to manually review information with VoiceOver . There are a lot of variables here. If VoiceOver did not sometimes fail to turn speech back on when muted, it would be much more convenient to just tell users to set a keyboard command, but much less so to turn VoiceOver back on.
I’m going to think some more on this. Having said all this, though…or written it anyway…I think the bottom line is to tell users to turn VoiceOver off. If you are not going to review the incoming text, you won’t need VoiceOver and especially not if my wish regarding popup windows could be implemented. Once the information is done reading, the user can simply hit CMD-W and have focus stay in the text field.
The speech manager generally does need improvements to multiple areas including latency issues, but that isn’t relevant here. I live in a dream world where applications can communicate with VoiceOver without private APIs, but I wake up every morning and realize this isn’t happening.
This is maybe a bit of a mouthful, and perhaps a confusing one, but I hope it makes sense anyway. This is my thought process. To be honest, I actually just wrote this as the thoughts came to me before they could slip away. They have a habit of doing that!
I’m glad to be a part of this, and I’m very enthusiastic about it.
On Oct 22, 2014, at 6:36 PM, Mark Phillips <email@example.com> wrote:
Excellent, I’m glad you were able to get logged in.
Yes, saving voice and other preferences is on the way. Perhaps as early as tomorrow. Also yes, we did choose not to clear commands entered, specifically to make it easier to repeat things. Along with voice settings, you’ll sometimes find that to get somewhere in The City rapidly you’ll want to just enter “north” six times in a row - things like that. And one more yes - we haven’t added the “help preferences” menu yet. Will do that as soon as settings are savable.
So your suggestion re spoken text. Making sure I understand correctly. You’d like the spoken text to be written to the browser, so that you could for instance select the text to hear it again? You make a very good point about uncommon names, as players often like to choose names which are pretty exotic. It’s certainly possible to write the spoken output to the browser. I think the more fundamental question to resolve first is: do we want to suggest that users leave VoiceOver turned on, or turn it off while interacting with the client? What do you think?
I think the disconnects you experienced were coincidence - not related to looking in any direction. I can’t duplicate that experience. This client, and the GUI HTML5 client it shares code with, is sensitive to network latency. In time this’ll get sorted out by the browser vendors but in the meantime it’s pretty bleeding edge.
Thanks Nicolai, really appreciate the excellent feedback!
On Oct 22, 2014, at 9:19 AM, Nicolai Svendsen <firstname.lastname@example.org> wrote:
Yes, it eventually sorted itself out. I gave it a little while, then just decided to try again.
I’ve had pretty good experiences with the client. it was a little confusing in the beginning when I realized that the commands entered did not clear themselves out, but then I found I could of course just hit enter to repeat the command. Of course, I’m so sure that this wasn’t the case earlier this morning, but it probably was. I assume this is intended behavior, but regardless I like it that way. It’s especially handy when adjusting speech rate, though I would make the suggestion that perhaps the voice preferences are stored somehow as well as other settings, presumably those that will eventually be located in help preferences. It just handed me a blank help topic when doing “help preferences,” so I’m going to assume that this is probably being worked on.
I’d also like to suggest the ability to review the spoken text in the browser window. As far as I can tell, this wasn’t possible to do with VoiceOver, but I can see this becoming a bigger issue when you start to interact with everyone else around you. Since it appears that you have to type a complete match when it comes to objects and names, I see the problem with names in particular. Certain synthesizers on various platforms may pronounce names quite differently than Alex might, especially if they are not common or spelled a certain way.
I ran into a weird problem. If I tried to look in a certain direction, the game would disconnect me and suggest that I have a bad internet connection. For instance, in the in-game tutorial, I tried to look north and l n. Both caused the same issue. All other commands have not so far.
One other oddity I ran into is that when you interrupt the speech with the CTRL key, it is sometimes necessary to do this multiple times. This isn’t always the case, so you might have to move around a little bit before reproducing this.
This is all I have so far before I got interrupted and had to quit.