These are chat archives for jdubray/sam

20th
Jun 2016
Edward Mulraney
@edmulraney
Jun 20 2016 08:04
good slides. is the presentation recorded?
Kelyak
@kelyak
Jun 20 2016 08:49
Is it just me or on page 13 the content and the footer (with urls) are overlapping each other ?
Apart from that, the part that show how each framework are wrong seems too big for me. What I mean is : for someone who followed this conversation since the beginning and already tried something with SAM, it can be understandable. But from an external point of view, it could really be seen as "everything else is crap, me better". I know it isn't like that but you should be careful about the possibility that some people could feel that way. Saying that SAM could be a great improvement is a good thing, but relying too much on SAM is better than x and y could have the opposite effect.
But that's only my opinion ^^
(and it's only by reading the slides, your speech might not give that feeling, I don't know :D)
Riccardo
@riccardoferretti
Jun 20 2016 09:39
I really like that you do a review of what’s out there. I agree with @kelyak , that just by looking at the slides, the review comes across as negative. Even just removing the “messy” images to the various frameworks could remove the negative association, then the conversation can focus on what you believe are benefits/drawbacks of those approaches
also, I didn’t realize mutation was a cornerstone of the approach - what do you mean by that?
Jean-Jacques Dubray
@jdubray
Jun 20 2016 10:10
@edmulraney the talk is this afternoon. Yes, I believe it will be recorded.
@kelyak I would not present it as everything else is crap, SAM is better, since you are comparing frameworks with a pattern, that being said, I'd be happy if I get people to think about their programming model. I had this discussion last night with someone who just fell in love with Elm and we talked about tasks/commands and the notion of "step" in Elm, and the reality is that there is no notion of Step in Elm and he could immediately relate to the boilerplate code he had to write because of that.
Jean-Jacques Dubray
@jdubray
Jun 20 2016 10:17
@riccardoferretti I paired down the figures, originally, I was trying to get a spatial view of where the concepts fit, but I agree it makes the figures very hard to digest. I have updated the slides with just a thumbnail.
Kelyak
@kelyak
Jun 20 2016 10:32
@jdubray I understand that, it's just the feeling that I had when looking at the sliders ;)
Jean-Jacques Dubray
@jdubray
Jun 20 2016 10:35
@kelyak @edmulraney @riccardoferretti thank you for the feedback!
Edward Mulraney
@edmulraney
Jun 20 2016 13:35
good luck with it. looks like it will be a useful talk for all of us
btw for the future, if youre interested in getting fancy looking slides out of the box in a simple editor: http://lab.hakim.se/reveal-js/
creates slides like Andre Staltz's "Don't react": http://staltz.com/dont-react/#/
David Turissini
@davidturissini
Jun 20 2016 22:07
Hello!
Jean-Jacques Dubray
@jdubray
Jun 20 2016 22:29
Hi David, great to meet you.
Riccardo
@riccardoferretti
Jun 20 2016 22:45
@jdubray I find this Elm code here interesting from that POV: https://gist.github.com/pdamoc/c96714479d9f531fbc7468d5670ef576
fundamentally the state is encoded in the model, and by organizing the functions “by state” instead of by funcionality (that is have the state as "top level”, instead of switching on it inside the various functions), and renaming subscriptions to nap, you get pretty much the SAM pattern
for the propose/accept it shouldn’t be very challenging to implement in Elm, it’s just a matter of organizing code/functions
although I am not sure what you mean by “no step” - isn’t the loop Msg -> Update -> View/Subscription a step?