<div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Date: Thu, 22 Jan 2009 12:10:28 -0600<br>
From: David Bruce <<a href="mailto:davidstuartbruce@gmail.com">davidstuartbruce@gmail.com</a>><br>
Subject: Re: [Tux4kids-tuxtype-dev] Gsoc ideas<br><br>
Hi Sreerenj,<br>
<br>
(list members - we had been discussing adding sound files for each<br>
word to be typed)<br>
<br>
><br>
> Hi David<br>
> But courresponding to each wav file ,we need to identify the typed word.And<br>
> if the word is incorrect ,the code should identify that.So for that we need<br>
> to do something with library.Is it possible without adding any additional<br>
> libs to our program?<br>
><br>
<br>
It depends on what sort of activity we make this into. If it is<br>
something with only one "target" at a time, like the phrase typing<br>
activity, the program would simply play the .wav instead of displaying<br>
the text to be typed, and everything else would be the same.</blockquote><div><br>I think better than having .wav files for each letter, we should link some kind of text to speech package with tuxtype. Otherwise having .wav files for each and every language can make the software too big. Moreover, it would be easier for adding new languages. <br>
</div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><br>
<br>
We could adapt the cascade or comet games to play a .wav of each word<br>
when it appears on the screen, either in addition to displaying it<br>
(easier) or instead of displaying it (harder). Either way, the<br>
program would already know the answer word without any speech-to-text<br>
conversion being required. The checking of the player's answer would<br>
be just like it is now. If we did this (and it might be a very good<br>
enhancement), we would need to take care to keep the words spaced far<br>
enough apart so that the .wavs don't overlap, and limit how many words<br>
are in play at once because the player might not remember what was<br>
said. Perhaps we could have the target blink when it gets halfway<br>
down and repeat the .wav, or something.<br>
<br>
If we add a lot of sound files, we will likely go over the 10MB file<br>
size limit for Alioth downloads, but I'm sure we could work something<br>
out.</blockquote><div><br>Don't you think that we should have speech output for each key the user presses also? This would be a great enhancement for the visually challenged people. I don't know how we will do this. This is just a suggestion. I can imagine something like - a male voice for the key pressed by the user, and a female one for each character that comes up on the screen, or something like that. I think this can be easily done if we use a text to speech package.<br>
<br>I think we should start by adding this feature to the practice game because there is only one character at a time and then from there go on and add the feature to the other games that display more than one character at a time.<br>
</div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><br>
<br>
Anyway, something of this nature sounds like a great GSoC idea. We<br>
still need to figure out who can mentor this year.<br>
<br>
David</blockquote><div><br><br>Regards,<br>Sreyas<br></div></div>