Hi,

I am thinking of an interface that goes in this direction for more than a year now, but I am not there yet.

A classic REPL works fine for sequential programs, but the problem with this is that as soon as you do anything meaningful and want to iterate, you get a bit lost. For example, imagine you demo on a system complexity like visualization using a Mondrian builder:

You first do:
view nodes: classes.
view edgeFrom: #superclass.
view treeLayout

This can work just great. But, now you need to add at the beginning:
view shape rectangle height: #numberOfLinesOfCode

You have to start again. The REPL interface was bound by the abilities of a terminal. But, we do not have those constraints anymore, and that interface has to be rethought.

One thing I am thinking of is having groups of text that get a preview. For example, imagine that you see a preview only after you say Cmd+Enter.

Now, about your points. You are right in that the implementation of Cmp+p was a nice solution 40 years ago, but now we can do better. There is no need for Cmd+p to affect the current text. This can be solved easily, and it is on the soon to show roadmap. This alone should alleviate the current playground usability.



Doru


On Wed, Apr 16, 2014 at 11:08 PM, Alexandre Bergel <alexandre.bergel@me.com> wrote:
> On 16/4/14 12:44, Alexandre Bergel wrote:
>> Yes, as the one of scheme, drRacket...
>
> how is it different than a workspace?
> I do not see the difference.

There are several differences:
        - In an evaluator, you just need to press enter to evaluate an expression. In a workspace you need to press Cmd-p
        - In an evaluator, you have the nice highlighting for _each_ entry. In a workspace, everything is messed up after evaluating the first expression
        - In an evaluator, you have usually pseudo variables to access previous results. In a workspace you cannot do that, else by putting parenthesis everywhere
        - an evaluator is well suited for demonstration (e.g., the Wolfram language, Light table…)

Alexandre


>
>>
>> Alexandre
>>
>>> Le 16-04-2014 à 2:23, Serge Stinckwich <serge.stinckwich@gmail.com> a écrit :
>>>
>>> What do you mean by evaluator ? A kind of REPL with history management ?
>>>
>>> Regards,
>>>
>>> On Mon, Mar 31, 2014 at 8:31 AM, Alexandre Bergel
>>> <alexandre.bergel@me.com> wrote:
>>>> It would be great to have an evaluator. A bit like a terminal or what they have with Wolfram or Aurora.
>>>>
>>>> I guess Doru has been thinking about that already. Is there something cooking ?
>>>> I plan to do one based on Roassal...
>>>>
>>>> Alexandre
>>>>
>>>> --
>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>> Alexandre Bergel  http://www.bergel.eu
>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Moose-dev mailing list
>>>> Moose-dev@iam.unibe.ch
>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>>
>>> --
>>> Serge Stinckwich
>>> UCBN & UMI UMMISCO 209 (IRD/UPMC)
>>> Every DSL ends up being Smalltalk
>>> http://www.doesnotunderstand.org/
>>> _______________________________________________
>>> Moose-dev mailing list
>>> Moose-dev@iam.unibe.ch
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>> _______________________________________________
>> Moose-dev mailing list
>> Moose-dev@iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> _______________________________________________
> Moose-dev mailing list
> Moose-dev@iam.unibe.ch
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev


_______________________________________________
Moose-dev mailing list
Moose-dev@iam.unibe.ch
https://www.iam.unibe.ch/mailman/listinfo/moose-dev



--
www.tudorgirba.com

"Every thing has its own flow"