OpenXTalk Forums

A person called Nick Antonaccio has provided a series of learning stacks:

https://live-code.net/livecode-tutorial.html

The link, however, to the example stacks is dead.

The lessons were 'live' as on Sunday 21 April 2024.

You mean 2021, but hadn't been changed since 2016. <-- the link here works

The lessons: however old they might be still have some value for those just starting with xTalk.

Does revSpeak still work, and on which operating systems?

Aaaaah:
- -
Here are ALL the example stacks:
- -

99BottlesOfBeer.livecode
Calculator.livecode
CashRegister.livecode
CoinFlip.livecode
FreehandDraw.livecode
HTMLCalendar.livecode
HappyBirthday.livecode
Madlib.livecode
Match Button (drag_drop).livecode
PigLatin.livecode
TileGame.livecode
TimeBetween.livecode
TimeList.livecode
TipCalculator.livecode
ToDoList.livecode
server1.livecode
server2.livecode

I couldn't find a license declaration anywhere at the site or in the stack file archive. Is it somewhere on his site that I'd missed?

Nor could I: and as they are freely downloadable from the sequestrated version on the Wayback Machine I don't quite see what your point is.

In the absence of any licence:

All example apps can be downloaded

seems unfetteredly permissive.

richmond62 wrote: ↑Sun Apr 21, 2024 8:02 pm Nor could I: and as they are freely downloadable from the sequestrated version on the Wayback Machine I don't quite see what your point is.

In the absence of any licence:

All example apps can be downloaded

seems unfetteredly permissive.

For personal use, I suppose. But redistribution and modification are unaddressed without a licence.

https://troz.net/post/2015/livecode-stacks/

Feel free to use or adapt as you see fit. They have not been tested using recent versions of LiveCode, so may not work as expected. As stacks are provided as is. Use at your own risk.

tperry2x wrote: ↑Sun Apr 21, 2024 10:28 pm https://troz.net/post/2015/livecode-stacks/

Feel free to use or adapt as you see fit. They have not been tested using recent versions of LiveCode, so may not work as expected. As stacks are provided as is. Use at your own risk.

Are those stacks from Sarah or Anthony?

If they're Sarah's that makes it much easier. She's often thorough about such things.

Does revSpeak still work, and on which operating systems?

It works on macOS, Win, iOS, Android, ad it can easily be implemented on a Web/Emscripten (HTML5 API) Engine.

For completeness, TTS also works in HyperSimulator and OpenXION (depending on the JavaVM it's running on having the needed component)

For regular Linux(es) not so much... I only have a simple Extension wrapper for that Linux 'speak-N-spell' sounding eSpeak(-ng) TTS (I actually built and tested this on macOS too). You can also get TTS by using a command line tool too (using shell() or 'open process' for non-blocking).

I'd like to use a better sounding Text-To-Speech API, but wasn't even sure if there is one that is as prevalent as eSpeak, and preferably a library that isn't too large that could be packaged with the IDE.

OpenXTalkPaul wrote: ↑Mon Apr 22, 2024 8:57 pm For regular Linux(es) not so much...
You can also get TTS by using a command line tool too (using shell() or 'open process' for non-blocking)

It is doable on Linux, even though it doesn't sound anywhere as human as MacOS (demo video), and I was just thinking, to make it non-blocking you could do: (replacing the path to my desktop with yours). The "speakplugin" is a little custom c++ program I created tonight that is based on eSpeak (so I'd credit them if we used it in OXT).

What I'd really like though is also the other way around: speech to text.

I know this is built in in MacOs, but if we had some way of implementing this in Linux and Windows too, that would be really cool.

(I could then program something for my dad who can't see much anymore, to hopefully automate his UI on his computer. ).

tperry2x wrote: ↑Tue Apr 23, 2024 6:39 pm

OpenXTalkPaul wrote: ↑Mon Apr 22, 2024 8:57 pm For regular Linux(es) not so much...
You can also get TTS by using a command line tool too (using shell() or 'open process' for non-blocking)

It is doable on Linux, even though it doesn't sound anywhere as human as MacOS (demo video), and I was just thinking, to make it non-blocking you could do:
cmd.png
(replacing the path to my desktop with yours). The "speakplugin" is a little custom c++ program I created tonight that is based on eSpeak (so I'd credit them if we used it in OXT).
size.png

Right, the '&' on the end of the shell command makes that run as a sub-process, the only problem is you then have to repeatedly-check the process ID to see if the processes is still alive or it finished running.
The nicest thing about eSpeak is its small file sizes (compared to other TTS)...unless you actually WANT that 1970s TTS sound (in which case I know there's an actual Speak-N-Spell emulator out there ). Not sure that there's any real benefit to having it in Extension module form, I can't remember if I had it running as non-blocking or not.
For macOS the Extensions module for NSSpeech (and probably AVSpeech, which someone else wrote) can do things that revSpeech can't do, for example my NSSpeech module can render speech to a sound file.

Speech To Text would be great, not just for visually impaired, but for people who don't like typing too.
I wonder are there even any FOSS Voice dictation libraries in the wild?

OpenXTalkPaul wrote: ↑Fri Apr 26, 2024 1:36 am Speech To Text would be great, not just for visually impaired, but for people who don't like typing too.
I wonder are there even any FOSS Voice dictation libraries in the wild?

I'd love to add something like that in. It would open a whole extra set of use cases and justifications for using OXT / OXT Lite, and make it so much more versatile. You could then in theory link it to ChatGPT (AI et all) and have an assistant where you simply speak to it for help.

tperry2x wrote: ↑Fri Apr 26, 2024 6:54 am

OpenXTalkPaul wrote: ↑Fri Apr 26, 2024 1:36 am Speech To Text would be great, not just for visually impaired, but for people who don't like typing too.
I wonder are there even any FOSS Voice dictation libraries in the wild?

I'd love to add something like that in. It would open a whole extra set of use cases and justifications for using OXT / OXT Lite, and make it so much more versatile. You could then in theory link it to ChatGPT (AI et all) and have an assistant where you simply speak to it for help.

I've played around with Dictation on macOS now and then (going way back to Mac Classic's "Speakable Items"), it's gotten better but still not insanely great. I always forget the key combo to invoke it, generally forget that it's available to use, or wouldn't want to use it with other people being in the room...

ChatGPT /AI aren't perfect either, certainly useful for certain tasks, but it rarely produces at first prompt ready to use code without the need for some correcting. It is great for tedious writing tasks though, like creating a lot of those Extension Builder binding strings, and for typing out syntax formatting for APIs with lots of constants.

I think it would be a better use of time if we added the choice to use an external code editor that integrates with the IDE instead of using the built-in SE. Apparently this was a feature that OXT's MetaCard ancestor had.

I know there's add-ons for Pulsar, VSCode, etc. that integrate ChatGPT and similar AI into the workflow. And those editors all have bunches of other useful code editor add-ons available (not sure about voice dictation though). We already have language support for xTalk LCS/LCB support with Pulsar, VSCode, BBEdit/TextWrangler, and a few others. (BTW, I've made a modified version of LCS/LCB language support package for Pulsar, so that it recognizes filenames ending with .oxtstack and .oxtscript )

I imagine the external code editing mechanism could work similarly to the way that the IDE (and some other apps I've used) integrate using external image editors. That's in the contextual menu for 'Image selected' context, and there's a revPref to set which editor. What it actually does is it creates a temporary image file in the temp directory and opens that in your image editor, meanwhile the IDE displays a dialog asking if you want to update the image and if so it then replaces the image's previous image data with the data from that temp file. There may be a more automatic-updating way this could work (using 'on resumeStack'), I'm not sure. I know the Script Editor already has a dialog that supposed to display when it thinks a stack may have been modified by an external editor, but I believe that's only meant for checking script-only stacks currently.

I was thinking about my previous suggestion of doing this today, and you are right - it might not be worthwhile.
I mean, perhaps it's just a gimmick to be able to talk to the machine and have it quote back AI-generated stuff at you.

I keep coming back to how siri isn't much good and cortana pretty much failed too.

However, for user interface control I think it would be brilliant.

There's a lot of talk on here, and use of the phrase "a good use of time".... surely we decide what a good use of our time is? It's not like anyone is paying us, and nobody is giving me a deadline (apart from the inexplicable 'rush to get to v1.0' of oxt lite, which happened - which I still don't understand to this day). Perhaps there's some underlying motive or rush people have to implement a feature 'yesterday', but then why haven't they come forward with example code if that's the case?

Lurking on one of my many bookshelves is the printout of my 2004/5 Master's thesis:
- -
And central to that thesis was an agent-led programming interface so that teachers and 'academics' could sit and tell a computer what to do to produce a computer program that would present their specialist knowledge without them having to go to lengths they could ill afford to learn how to program a computer.

While the prototype (developed with Runtime Revolution) was a mouse-directed thing that ran through a decision tree, the proposal was to end up with a speech-to-text interface so, say, your specialist in Scots history at the time of Alexander III could sit and talk into a microphone and the 'thing' would put together a program for the "Prof".
- -
Unfortunately I cannot locate either the software files or the Clarisworks document at present.

tperry2x wrote: ↑Fri Apr 26, 2024 10:09 pm There's a lot of talk on here, and use of the phrase "a good use of time".... surely we decide what a good use of our time is? It's not like anyone is paying us

Absolutely we do! I was basically thinking out loud there. I'm trying to make a push right now. I think I would like to make some sort of milestone release, for my own satisfaction.

OpenXTalkPaul wrote: ↑Tue Apr 30, 2024 3:04 pm Absolutely we do! I was basically thinking out loud there.

That was not a dig at you Paul. In various posts, there's a few instances where people say "a better use of your time would be..." (and similar phrases). I just find it a little 'off' when people dictate how we should be spending our time. Especially when *everyone* is welcome to contribute, even if it's finding bugs - that's a start. They can all go into the to-do-list for the next release.
I'm not under any pressure or any schedule except my own.

Indeed: I am just relaxing on the balcony in the evenibf sun having done a heavy 8 hours in the garden with a spade, a drill, and a metal cutter. . .

Tomorrow (holiday for me) is forecast to have heavy, all-day rain: so there will be touch of devotion to the cause, but not because SixthWorld or Uncle Tom Cobbley is telling me what to do, but just because I want to.

Cheers! Virtual drinks all round.