Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:111029
MIME-Version: 1.0
References: <21C5073D-3F39-49DA-8686-E027AE780793@joshbruce.dev>
 <CAEsg9X1sCm=G5Z5Np__GmO2FbeNY+p-ue1rwUQ6cjFQyaf9rSg@mail.gmail.com>
 <dc315306-3e3a-41f0-9e20-6d66de3bb6d9@www.fastmail.com> <CADyq6s+_GoMcWWQqoCi9w=v336cVLYiCjojXNrnf_PbGFFkTUw@mail.gmail.com>
 <2d0fea79-0729-4cce-8b3a-be834614f242@www.fastmail.com> <CADyq6sLTLG5bZtv3TC0-ra-M6wMswLpMVCT4L=MCnOjtKfb5fg@mail.gmail.com>
 <1a34b907-4766-4c13-9da3-34a3df527679@www.fastmail.com>
In-Reply-To: <1a34b907-4766-4c13-9da3-34a3df527679@www.fastmail.com>
Date: Wed, 15 Jul 2020 19:22:01 +0200
Message-ID: <CADyq6sLc0EWfHLjU+qNYD2H1oS1V8KwzchatB+W=5Rj+XtH8Hg@mail.gmail.com>
To: Larry Garfield <larry@garfieldtech.com>
Cc: php internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="0000000000001e817f05aa7e2ac8"
Subject: Re: [PHP-DEV] [RFC][Discussion] Objects can be declared falsifiable
From: ocramius@gmail.com (Marco Pivetta)

--0000000000001e817f05aa7e2ac8
Content-Type: text/plain; charset="UTF-8"

 <http://ocramius.github.com/>Hey Larry,


On Wed, Jul 15, 2020 at 7:15 PM Larry Garfield <larry@garfieldtech.com>
wrote:

> On Wed, Jul 15, 2020, at 11:59 AM, Marco Pivetta wrote:
> > Hey Larry,
> > <http://ocramius.github.com/>
> >
> >
> > On Wed, Jul 15, 2020 at 6:55 PM Larry Garfield <larry@garfieldtech.com>
> > wrote:
> >
> > > I disagree entirely.  The value of a Maybe over just null is
> > >
> > > 1) You can bind to it even if the result is empty; the "if is null"
> check
> > > gets abstracted away from the main code.
> > >
> >
> > That's correct: that's how the `Maybe` monad works too.
>
> That... is what I'm describing?  A Maybe Monad?
>

Sorry, my wording wasn't right.
`>>=` implemented for `?T` does the exact same thing as `>>=` implemented
via `Maybe T`: same abstraction, different implementation (`instance Monad
?T` vs `instance Monad Maybe`).

Therefore, if we had monadic syntax, you could write the following code:

```hs
do
    value <- someCall someInput
    someOperation <- someOtherCall value
    return value
```

The above would work with both `?T` and `Maybe T`, luckily.


> > > 2) You are forced to unwrap/extract the value if you're not just
> binding,
> > > which makes it less likely you'll forget to check if it's null
> >
> >
> > No need to unwrap: that's what `>>=` does for you.
> >
> > Assuming you use psalm or phpstan, a program that doesn't use `>>=`, or
> > forgets to check for `null` on `?T` does not type-check anyway, and can
> be
> > rejected before even writing any runtime tests: that's sound.
>
> My Haskell experience is zilch, but since PHP lacks a dedicated bind
> operator that also magically unwraps, that doesn't seem relevant.  (Mind
> you, I'd be all for a dedicated bind operator that also auto-unwraps!  But
> that's not on the radar at the moment.)
>

 No need for magic operators: just because `>>=` is written with an infix
operator in Haskell doesn't make it magic.

In fact, Marco Perone released a nice PHP type-safe FP library around
functional PHP abstractions today, and `>>=` is implemented as `bind()`
there:

 *
https://github.com/marcosh/lamphpda/blob/ebde827e78f0d5889da041fecfea353817890e35/src/Maybe.php#L175
 *
https://github.com/marcosh/lamphpda/blob/ebde827e78f0d5889da041fecfea353817890e35/src/Typeclass/Monad.php#L22

You can use that today: it already brings some value to the table, without
any need for language-level additions (besides generics) :)

Greets,

Marco Pivetta

http://twitter.com/Ocramius

http://ocramius.github.com/

--0000000000001e817f05aa7e2ac8--