Removing Etoys (was Re: A process proposal for 3.10)

Bert Freudenberg bert at freudenbergs.de
Wed Oct 18 12:37:53 UTC 2006


As Juan wrote, removing Etoys from Morphic while keeping it both  
loadable and functioning properly is futile.

So either you leave it in, or you consciously give up compatibility  
with anyone using Etoys now, like the squeakland distribution, OLPC  
distribution, Smalland, the Spanish LinEx version, the Japanese  
Nihongo version etc. Already synchronizing Squeakland and 3.8 was  
hard, nobody has tried yet for 3.9, but this would make it outright  
impossible.

I'm *not* saying you should not do this, but please be aware of the  
possible consequences.

- Bert -

Am 18.10.2006 um 14:11 schrieb jvuletich at dc.uba.ar:

> Hi Giovanni,
>
> I don't think what you say is doable. We don't have the means to  
> break all
> dependencies on eToys, and at the same time keep eToys working. It  
> would
> need the same kind of work as to make it unloadable and loadable back.
>
> When I realized how much refactoring is needed to do that, I  
> stopped the
> MorphicSplitters effort and quited as the Morphic Steward.
>
> What I propose can be done. I did it for 3.7. You can download it from
> http://www.jvuletich.org/Squeak/EToysFreeMorphic/ 
> EtoysFreeMorphic.html . I
> believe Pavel did something similar (although I haven't looked at it.
>
> Anyway, I'd like to be proven wrong. Are you volunteering?
>
> Cheers,
> Juan Vuletich
>
>> When resource is scare, the smarter come out ;)
>>
>> How is big etoys?
>> And How is deeply integrated in Squeak?
>>
>> If Etoy is not so big (as I remembered), we can simply start to  
>> "cut its
>> wires" from Morphic and let it aging around.
>> We can start creating a mechanism to deprecate some methods.
>> By the way the deprecation engine seems to me very important to do.
>>
>> We can do this thing with less then 8 hours of work in my own  
>> opinion,
>> I suggest to rethink our apprach and to adopt a "Panta  
>> Rei" (verything
>> changes - the philosophy of Heraclitus).
>> We should start to plan the new Squeak version WITHOUT throwing  
>> away the
>> bad
>> things.
>>
>> We can start taking bad thing alone in a room, putting heavy doors  
>> in it,
>> and then
>> prohibing them to exit :)
>>
>> The other developers will start stopping using EToy in a smooth way.
>>
>> After a while we can think how to dismiss them... or not.
>>
>> In my own opinion frequently Squeak home change, so the problem  
>> will solve
>> smootly without so much pain.
>> Let's discuss on this approach.
>>
>>
>> On 10/18/06, Juan Vuletich <jvuletich at dc.uba.ar> wrote:
>>>
>>> So, this seems a good time to remove eToys from the official  
>>> release. I
>>> can team with Pavel and Stef (and any other volunteer) to do this.
>>>
>>> However, it will take some 20 or 30 hours of work, and I think we  
>>> need
>>> to know if this will be adopted, otherwise I won't spend time on it.
>>>
>>> I guess the Board could lead, and make a decision, or enlight me  
>>> about
>>> the decision process for issues like this one.
>>>
>>> Cheers,
>>> Juan Vuletich
>>>
>>> Marcus Denker wrote:
>>>> The SqueakLand people don't use 3.9, and I am quite sure they never
>>> will.
>>>>
>>>> Etoys 1 is past live-cycle. There is 3.8/OLPC which is a cool Etoys
>>>> image for eToys1.
>>>>
>>>> For the future, there needs to be a new eToys2 that is  
>>>> maintainable.
>>>> There is a very cool demo of a next-gen eToys based on Tweak.
>>>> That seems, to me, much more the thing to take a look at for the
>>> future
>>>> eToy system.
>>>>
>>>>        Marcus
>>>
>>>
>>
>>
>> --
>> "Just Design It" -- GG
>> Software Architect
>> http://www.objectsroot.com/




More information about the Squeak-dev mailing list