4 ms·

Is goto really the best way to link together flows? Personally, and this could be my complete ignorance of the chatbot topic, I can't even imagine the possible

Is goto really the best way to link together flows? Personally, and this could be my complete ignorance of the chatbot topic, I can't even imagine the possible hellscape that would be large applications. Like... how do you ensure you don't just end up in some huge loop? How do you even reliably test it? Does it have some sort of built in state-machine to keep you from constantly shitting on your feet? Also... where is the source code? What's it implemented in? Is there a timeframe for its release? Overall, I think this is a pretty damn cool idea but without knowing more about how it works under the hood it's really hard to want to dig in more. FWIW, the name makes me think it's an ML (which it's not), or something to do with machine learning and chat bots (which it is also not).

frsechet 7y ago
Hi rubyn00bie, Thanks for the feedback! Let me answer your comments one by one, but first of all, keep in mind that we announced the beta this week, so very early days and feedbacks like yours are what will steer us on the right path :-) goto: I know this looks a bit old-fashioned, but the main goal of the language (and framework) is to provide an easy way for developers (and our goal is that even non-developers should get it) to understand how the conversation will work. "Do this then go there" is an easy enough construct. The point is that it is easy to explain. It might not be the "best™" way to do it, but it is a very easy way, one that we have never had to spend more than 5s explaining to absolute beginners. hellscape: in our short-term roadmap we have a "issue warnings at compile time" so that people know when they do something that is probably wrong. But we have tested it in large-scale apps: for instance, one of the chatbots we developed is a dialogue-based interactive fiction, which has a game lifetime of about 45-60 minutes. Pretty large, but fits in about 1000 lines of code. Basically, the fact that we enforce a flat file structure makes it easy to find your way. But once more, happy to improve our system in the future. Anything we can do to make things simpler is welcome. source code: It's written in Rust. We plan to open source it in a few weeks, can't say a date for sure but "soon". We just need some time to clean up some stuff. If you absolutely need to have a look at the code before testing, I'm happy to oblige, send me an email to francois@clevy.io! name: naming is hard. Our target is to make it easy to build a rich chatbot even for non-developers. We chose the name CSML (Conversation Standard Meta Language) because we aim to make it a standard programming language for building conversational experiences. And the ML at the end kind of looks like HTML which kind of looks friendly enough for beginners. It's definitely not a markup language, of course there is absolutely 0 machine-learning in it :-) Thanks again for the feedback! Let me know if you have any more questions!

frsechet 7y ago

Hi rubyn00bie, Thanks for the feedback! Let me answer your comments one by one, but first of all, keep in mind that we announced the beta this week, so very early days and feedbacks like yours are what will steer us on the right path :-) goto: I know this looks a bit old-fashioned, but the main goal of the language (and framework) is to provide an easy way for developers (and our goal is that even non-developers should get it) to understand how the conversation will work. "Do this then go there" is an easy enough construct. The point is that it is easy to explain. It might not be the "best™" way to do it, but it is a very easy way, one that we have never had to spend more than 5s explaining to absolute beginners. hellscape: in our short-term roadmap we have a "issue warnings at compile time" so that people know when they do something that is probably wrong. But we have tested it in large-scale apps: for instance, one of the chatbots we developed is a dialogue-based interactive fiction, which has a game lifetime of about 45-60 minutes. Pretty large, but fits in about 1000 lines of code. Basically, the fact that we enforce a flat file structure makes it easy to find your way. But once more, happy to improve our system in the future. Anything we can do to make things simpler is welcome. source code: It's written in Rust. We plan to open source it in a few weeks, can't say a date for sure but "soon". We just need some time to clean up some stuff. If you absolutely need to have a look at the code before testing, I'm happy to oblige, send me an email to francois@clevy.io! name: naming is hard. Our target is to make it easy to build a rich chatbot even for non-developers. We chose the name CSML (Conversation Standard Meta Language) because we aim to make it a standard programming language for building conversational experiences. And the ML at the end kind of looks like HTML which kind of looks friendly enough for beginners. It's definitely not a markup language, of course there is absolutely 0 machine-learning in it :-) Thanks again for the feedback! Let me know if you have any more questions!