possible new implementation?

[safareli]

I have implemented something like daggy here

how is it different?:

• Instead of constructor Functions it uses Object.create
• when sum type is created it returns type representative and you can call it as it's just an object
• values used to create object is stored as array and there is no need to allocate new array in cata
• type definition is stored in type representative and it allows to have only one cata and toStrings

I believe it has a bit better performance too. would like to hear what others think.

[safareli]

ping @puffnfresh @SimonRichardson

[SimonRichardson]

I'll check on Monday if that's ok 😀

…

On Sat, 21 Jan 2017, 19:08 Irakli Safareli, ***@***.***> wrote: ping @puffnfresh <https://github.com/puffnfresh> @SimonRichardson <https://github.com/SimonRichardson> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#7 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACcaGAgedt2wdEL0S5C54srstpsAiy1Pks5rUlebgaJpZM4LoSqh> .

[safareli]

np :P

[SimonRichardson]

I'm fine with this, it would be good if we can move sanctuary-type-classes to fantasy-sorcery, otherwise this seems like it's a good start.

@puffnfresh you mind if we do a radical thing like this, I'm more than happy for it to move forward!

[puffnfresh]

Don't mind.

Irrelevant; but can it be typed with TypeScript?

[safareli]

I'm fine with this, it would be good if we can move sanctuary-type-classes to fantasy-sorcery, otherwise this seems like it's a good start.

/cc @davidchambers

Irrelevant; but can it be typed with TypeScript?

If you are asking about:

• sanctuary-type-classes, i think it's somehow possible.
• about this implementations, I think no.

What would be the best way to move forward on this?

1. create PR here (it will be a breaking change)
2. create new repo in FL and create PR there (then create a note here about new implementation)
3. move daggy to FL and do 1.

---

btw there is also a related work sanctuary-union-type and @JAForbes/sum-type which started as using sanctiary-def for type ckecking on @paldepind/union-type

[SimonRichardson]

Question how does a typed union work with types i.e. M A -> M B do you have to define new union types when this happens? I have no idea, some link to code would be amazing :) Really interested!

[safareli]

@SimonRichardson I think I didn't understand your question, if you want some example of union
here is Par, Seq, Concurrent from some from my free.

[SimonRichardson]

I was on about type checking versions and how it knows what to type check when you've provided a chain for example.

[safareli]

ah, I think @davidchambers is the one who can answer that question :P

[SimonRichardson]

@safareli I vote we go for it, we can clean up as we go a long, nothings perfect, we just need forward momentum!

[safareli]

What would be the best way to move forward on this?

1. create PR here (it will be a breaking change)
2. create new repo in FL and create PR there (then create a note here about new implementation)
3. move daggy to FL and do 1.

@SimonRichardson

[SimonRichardson]

Move daggy to FL and do 1 imo

…

On Fri, 27 Jan 2017, 08:08 Irakli Safareli, ***@***.***> wrote: What would be the best way to move forward on this? 1. create PR here (it will be a breaking change) 1. create new repo in FL and create PR there (then create a note here about new implementation) 1. move daggy to FL and do 1. @SimonRichardson <https://github.com/SimonRichardson> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#7 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACcaGFF-0anVFfNI-Rk0STT_uE7vP5pvks5rWaYIgaJpZM4LoSqh> .

[safareli]

ok will open PR here then, we can move it later too.