Convincing a harvester (was on SqF list)

Stephane Ducasse ducasse at iam.unibe.ch
Wed May 7 07:01:02 UTC 2003


Ok this is perfect for me. I think that this way we could be sure that 
the best
happens because people are concerned with the changes. For example all 
the people that want
TimZones, should sit, merged and done.

This is what we are doing for the kernel. We need it.

By the way I do not make distinction about the future direction of 
Squeak. My only concerned is that we can continue to develop stuff in 
Squeak without the need to be a superocder like nathanael to fix all 
kinds of bugs and errors to arrive to the point.

I also want to use Squeak for teaching good programming and not just 
say to the student: this is fun but it stinks. Look at SystemDictionary 
to convince you or the mess with UI access from deep core places. 
Anyway everybody knows that. So the cleaning and improvements will 
benefit to ***EVERYBODY***. If Squeak would have been clean, nathanael 
would not have needed to fight with it and we would be far ahead, I 
imagine that this is the same for lot of people.

We should not confuse creation and basic principles of good OO and 
software engineering. As a creator I need to tools that let me 
experiment FAST.

This is why the network fix are important too. and certainly all kind 
of other improvements.

stef

On Wednesday, May 7, 2003, at 08:50 AM, Cees de Groot wrote:

> On Wed, 2003-05-07 at 08:41, Stephane Ducasse wrote:
>> Now I would not like that everything gets automatically accepted 
>> before
>> we have serious tests
>> to check that people are not breaking what other are doing.
>
> Well, I wasn't exactly arguing for *everything* from *everyone*. I was
> arguing more along the lines of:
> - A group of people with good standing in the community proposes a
> largish refactoring project;
> - They outline some sort of vision;
> - The community declares it is happy with that vision;
> - The results are automatically harvested.
> Yes, some sort of barrier to enter into the auto-harvesting process 
> must
> be in place. It would be ludicrous to accept patches from me without
> prior review. But if there's a project, and especially if there's a
> project done by a *group* of people who I assume:
> - Work together;
> - Don't want to look like a bunch of fools;
> I'd assume that this group would make sure that they put only the
> quality bits into the update stream, and would *voluntarily* ask the
> community for feedback if they are unsure about patches.
>
> Communities scale by means of trust, and nothing else.
>
>
> <signature.asc>



More information about the Squeak-dev mailing list