Thanks to @Hillelogram, @ztellman, @garybernhardt, @dribnet, and @colinfleming for early feedback on the tutorials and overall scope.
-
-
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Very cool work!
-
Thanks! I'm just following the playbook I learned from the Clojure community: Read papers from 30 years ago and implement 'em = )
-
Can you share some of the papers that went into this? It looks like a great tool for thinking through application design
-
Harel's original: http://www.inf.ed.ac.uk/teaching/courses/seoc/2004_2005/resources/statecharts.pdf … history: http://www.wisdom.weizmann.ac.il/~harel/papers/Statecharts.History.pdf … Horrocks UI/discussion book: https://amzn.to/2J68rFw Samek code/embedded book:https://amzn.to/2J23Xzz
- 1 more reply
New conversation -
-
-
For those of us who already know a bunch about formal methods, care to compare this to other tools? To Alloy, to model checkers, to …? What's the underlying state exploration logic?
-
I enjoyed the traffic light demo, but my "hello world" is a PAIR of [perpendicular] traffic lights, so you can also write basic safety (no two greens) and liveness (eventually each gets green) properties. I'm curious what that looks like.
-
To my understanding this is intentionally just the specification part, not any verification part. It's to make it as easy as possible to communicate with designers and implementers about the spec.
-
But having put all that work into developing a spec, wouldn't you want to do more with it? Put differently, there's a fine line between a "property" (to verify) and a "fact" [as Alloy calls them] (to assert). Would be nice to fluidly move between the two.
-
http://Sketch.systems is trying to make the very idea of abstract states tangible and compelling to folks who've never been exposed to such thinking. It's trivial compared to Alloy, TLA+, etc. --- no state space enumeration, invariant checking, SAT solver, variable expansion.
-
People must experience the usefulness of counting with their fingers before they'll be motivated to learn logistic regression. http://Sketch.systems is not for people who already love formal methods; it's for people who've never heard of them.
-
That seems like an unnecessarily aggressive response, and a bit of a red herring to someone who's thought about these issues a LOT, but thanks for replying and I certainly wish you well.
-
My apologies --- I was trying to explain our design choices, not make a value judgement or dismiss your question. Our tool doesn't model data or check invariants because *we* don't know how to implement and explain those (useful!) features without compromising accessibility.
- 1 more reply
New conversation -
-
-
What’s the best way to chat about new features? :) I’d love to be able to model concurrent states.
-
I'd love to discuss, you can email me: kevin@generalreactives.com Concurrent states are on the roadmap (it's why prototype code gets array of `active_states` rather than a single object). Not sure when, though, and how useful they'll be without, e.g., transition guards.
-
Yes I noticed the array. Good thinking. I’ll send you an email!
End of conversation
New conversation -
-
-
Neat! I think it will be very interesting when these specs are attached more directly to code so both developers and non-developers are on the same page. E.g. understanding the slack notification algorithm from reading code is probably a nightmare: https://twitter.com/mathowie/status/837735473745289218?s=21 …
This Tweet is unavailable.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Yep,
@lynaghk showed me it a while ago! Looks really good.
End of conversation
New conversation
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.