> On 23 Feb 2016, at 13:43, Tudor Girba <[email protected]> wrote: > > Hi, > >> On Feb 23, 2016, at 1:24 PM, Esteban Lorenzano <[email protected]> wrote: >> >> >>> On 23 Feb 2016, at 13:20, Alain Plantec via Pharo-dev >>> <[email protected]> wrote: >>> >>> >>> From: Alain Plantec <[email protected]> >>> Subject: Re: [Pharo-dev] [Bloc] Do we want <return: #Point> or <return: >>> Point> >>> Date: 23 February 2016 at 13:19:34 GMT+1 >>> To: Pharo Development List <[email protected]> >>> >>> >>> Hello Doru, >>> >>>> On 23 Feb 2016, at 10:33, Tudor Girba <[email protected]> wrote: >>>> >>>> Hi, >>>> >>>> At the moment, it is used for documentation purposes in Bloc. It is part >>>> of the effort of Alex to document Bloc thoroughly. I think it is an >>>> interesting idea, in that we would have a significant case study for that >>>> can be used later as optional types information to improve static tool >>>> support. And it does not hurt at the moment. >>>> >>>> What do you think? >>> >>> Experimenting optional types is a good idea. We could have Tools to >>> dynamically check them. >>> What about a separate typing model that could be used to add/remove the >>> types annotations ? >>> Also, I wonder if Bloc is the good package to experiment optional types. >>> Maybe it is not stable enough. >>> Cheers >>> Alain >> >> please, please, please keep the scope of those experiments aside the main >> effort who is to have Block/Brick running as soon as possible. >> >> Esteban >> >> ps: … optional types… at the end, we will realise Gilad was right all this >> time :) > > I think there is a misunderstanding. > > The current annotations are there for documentation purposes. Given that Alex > did an extensive pass to document Bloc, they should remain in place as they > document the contracts. Documentation like this is something we should > embrace for such a central piece. > > That this also provides the information for a potential optional types > experiment it’s a different issue, and it would not affect at all the > implementation of Bloc.
I understand why it is done in the context of documentation, and I am all for good documentation. But it really is a dangerous road to start on. It sends a weird signal, as if it is better to start adding static typing information. The next step will be that someone suggests to start adding the same kind of annotation typing information to all arguments, next all instance variable, it can't hurt right ? Then we write tools to use that information, then we start requiring it, next we have Java with tons of boiler plate code for nothing. > Cheers, > Doru > >>> >>> >>>> >>>> Cheers, >>>> Doru >>>> >>>> >>>>> On Feb 23, 2016, at 10:25 AM, Alain Plantec via Pharo-dev >>>>> <[email protected]> wrote: >>>>> >>>>> >>>>> From: Alain Plantec <[email protected]> >>>>> Subject: Re: [Pharo-dev] [Bloc] Do we want <return: #Point> or <return: >>>>> Point> >>>>> Date: February 23, 2016 at 10:23:33 AM GMT+1 >>>>> To: Pharo Development List <[email protected]> >>>>> >>>>> >>>>> I don’t like it too. >>>>> Alain >>>>> >>>>>> Le 23 févr. 2016 à 09:50, Nicolai Hess <[email protected]> a écrit : >>>>>> >>>>>> >>>>>> >>>>>> 2016-02-23 9:47 GMT+01:00 stepharo <[email protected]>: >>>>>> Hi >>>>>> >>>>>> I saw that something <return: #Point> or <return: Point> >>>>>> I do not know why but I have the impression that <return: #Point> is >>>>>> better. >>>>>> Because we may have code not present and still want to load the code. >>>>>> >>>>>> I would like to know for what this is used. >>>>>> I don't like it. >>>>>> >>>>>> >>>>>> Stef >>>>>> >>>>>> >>>>> >>>>> >>>>> >>>> >>>> -- >>>> www.tudorgirba.com >>>> www.feenk.com >>>> >>>> "We are all great at making mistakes." >>> >>> >>> >> > > -- > www.tudorgirba.com > www.feenk.com > > "If you interrupt the barber while he is cutting your hair, > you will end up with a messy haircut."
