links to this page:    
View this PageEdit this PageUploads to this PageHistory of this PageTop of the SwikiRecent ChangesSearch the SwikiHelp Guide
Last updated at 1:08 pm UTC on 20 October 2006
Imagine these ideas used to teach Smalltalk or Literate Squeaking

We just finished a system called StoryHarp for creating interactive stories operated by voice with text-to-speeech, sound, and music output. We prototyped parts of this system in Squeak, but delivered it in Delphi for Win95/NT for various reasons. I though I'd list these reasons, as they are more or less future directions I'd like to see Squeak go, and I'd appreciate any help you can offer in getting Squeak moving in those directions.

Squeak does not yet have built in text-to-speech. A Squeak based TTS engine would be great. There are implemetations I know, available for non-commercial use written in C++. Anyone know of a Smalltalk based text-to-speech system? Could we get Apple to release Macintalk source for use in Squeak?
Look at Squeak speaks!Hans-Martin Mosner

Squeak does not have speech recognition built in (although it does have sound input). Again, anyone know of Smalltalk base SR software? We'd be content with something fairly simple. We don't need dictation capabilities. Ideally it would use the TTS system to generate waveforms for phrases and choose a nearest match from what the person said: I believe this is how the earliest working SR systems worked anyway, as I heard it way back from my college advisor, on the first SR-TTS project the TTS group got done first and then scooped the SR group.

I know we could link Squeak to SR and TTS engines via calls to native code, as we did with the Delphi app which uses Microsoft Agent and a SR and TTS engine. But we'd really like an all-Squeak solution. The advantage to using Squeak for this is cross platform portability. Since we now have a big chunk of the market covered with a Win95/NT product, we'd like a Squeak implementation to cover almost *all* the other platforms for one investment. Also, it creates installation problems and issues when products need other products to function. As an alternate to having Squeak do SR and TTS, maybe there could be a standard API interface to TTS and SR engines from Squeak that was supported by all platforms?

Our prototype GUI in Squeak looks really lame because we used standard ST-80 components. Morphic was too slow and also somewhat incomplete. MVC is too awkward; this from someone who started using MVC over eight years ago. Squeak really needs a crisp set of widgets analogous to TK. Ideally, they would be prototype based, like Newton widgets.

We wanted to embed Squeak in the product, but the current Squeak development interface is too confusing for novices. StoryHarp supports macros we want to link to specific bits of Smalltalk code. We need a simpler front end and debugger for programming in Squeak, which hides the power and complexity of Squeak until the end user needs it; maybe like LearningWorks? We also want to have a minimum footprint Squeak with only the components we absolutely need, with the rest demand loadable.

Please don't take these as gripes. StoryHarp improved greatly during its time as a Squeak app: you'll see the browser influence immediately. It was touch and go a few months back as we almost went straight ahead with a Squeak version instead of a Delphi version. We just got cold feet and decided to take a safe route first. Now that something is working, we can think about being more ambitious in version two. – PaulFernhout