[IAEP] GPA Notes from visit on Wed July 8

Gary C Martin gary at garycmartin.com
Wed Jul 8 19:51:47 EDT 2009


Hi Greg,

First, thanks for the great write up!

On 8 Jul 2009, at 20:40, Greg Smith wrote:

> Seemed  like a lot to me but most kids did it with a little help at
> some steps. The other hard work flow was creating two images, one
> right after the other. Here are the steps for that:
>
> - Open paint
> - Create painting
> - Change name
> - Click keep button (there's no feedback that it was saved, definitely
> need something there)
> - Change the name again
> - Click Edit tab and erase the whole picture
> - Paint new picture
> - Save new picture

These steps show a very common mistake/misunderstanding based around a  
central design idea in Sugar – and likely show that idea to be a  
dangerous and/or poorly implemented one. It ties into the whole idea  
behind Journal, Activities, Keep, and Stop.

Activities do not save, in the traditional sense, when you hit that  
keep button. The object you generate with 'keep' is of the the same  
'Activity' but with a different state. You end up with a trial of what  
looks like saved objects in your Journal with different states (sounds  
fine so far...). But try resuming any 2 of them at once... The first  
resume will work as expected, the 2nd will just switch you to the  
already opened Activity instance... Sugar believes them to be the same  
Activity, just different revisions, like a hidden chain of versions.  
I'd certainly suggest you not try collaborating with objects created  
like this, especially Write. You WILL be in a world of pain and WILL  
very likely lose work. But in the above case, the importing of such  
objects into Memorise luckily avoided the 'same activity id' problem.

The correct procedure with current (and past) Sugar is:

- Open new Paint Activity
- Create painting
- Change name
- Click stop
- Repeat

This way each Journal entry is a new, separate Activity – no hidden  
barbed-wire code hooking it to other seemingly separate entries.

Regards,
--Gary



More information about the IAEP mailing list