Next steps for MCP?
Stephane Ducasse
ducasse at iam.unibe.ch
Sun May 25 11:04:53 UTC 2003
Hi daniel
This is really interesting. Could you come up with a set of description
of "architectural" dependencies.
For example when working on KCP, I noticed that Dictionary was
referring to MorphGraphicalMenuDictionary.
While it should be in the other directions. So I fixed it if I remember
correctly.
It would be excellent to have a bunch of Spaghetti tracer scripts
identifying different flaws in Squeak.
We could then evaluate them to see if we progress. Alex plays with
SpagTr but I did not had the time to discuss
with him.
I created a page on the KCP for all kinds of stuff we could fix such as
the deepHorror from Ctuluh: Utilities.
Please add ideas there too.
Diego for the KCP we are not a lot, I'm mainly pushing and discussing
with others encouraging them to pick
a topic and fixing it. Alex worked on tools. I know that roel is
interested to fix certain aspects,
nathanael other so at the end we will have a cool result but this is
mainly ant moves. So I try to do regularly some cleaning.
Stef
On Sunday, May 25, 2003, at 01:19 PM, Daniel Vainsencher wrote:
> Well, seemed to me for 2 people you got quite a lot done, so it seems
> to
> me you're getting along fine. Of course, there's work enough in
> cleaning
> morphic for 10 people, but I'm sure more will join in.
>
> BTW, if some parts of the task are less interesting than others, focus
> on the interesting ones, we're happy if you're happy :-)
>
> I wonder if you find this interesting - the class category
> Morphic-Experimental sounds to me unessential to Morphic. But looking
> at
> the reference graph using SpT, I see that actually, Morphic-Kernel does
> depend on it in a few ways. The *PropertiesMorph classes are referenced
> by Morph. This requires some sleuthing as to whether the
> *PropertyMorphs
> should actually be in the Squeak core (in which case, they should be
> moved out of the category *Experimental), or that they should be made
> somehow optional extensions, which would require some design solution.
>
> I'm building a test+tool suite that surfaces this kind of issues for
> Morphic specifically, and Squeak in general. For example, I did -
>
> mma_MultiModuleAnalyzer onCategoriesWithPrefix: 'Morphic'.
> mma allShortestPathsFromPackage: (mma packageDefiningClass: Morph) to:
> (mma packageDefiningClass: TextPropertiesMorph)
>
> and got a whole list of paths full of design darkness. For example,
> TheWorldMenu knows the Flaps (sigh), the Flaps know the
> ProjectNavigatorMorph (Huh? flaps should be pluggable mechanism, and
> nothing more) which knows the EToyProjectHistoryMorph, which is a bit
> weird since the latter is in Experimental.
>
> Which suggests already two (IMHO) interesting redesign and cleanup
> projects -
> 1. How do we make Flaps pluggable enough to express everything it knows
> about other classes, without hardcoding?
> 2. Same for TheWorldMenu.
>
> Daniel
>
> diegogomezdeck at consultar.com wrote:
>> Hi Daniel & list...
>>
>> The very first point to attack is to get more people on MCPTeam. We
>> was 5
>> in the beggining and now we're only 2 of us.
>>
>> We're planning to call (again) for help. To cleanup the morph part
>> is a
>> big job (and boring) and the "future" looks not so good with only 2
>> of us.
>>
>> After this point, we have a set of really-small-changes to do
>> (basically
>> some changes we rollback to avoid conflicts). And then our plan is to
>> discuss in public the next steps.
>>
>> Cheers,
>>
>> Diego
>>
>>> Hi MCP, everybody.
>>>
>>> Now that we have the first MCP stuff in the image, I'm sure I'm not
>>> the
>>> only one curious about your intended next steps. I'm very happy with
>>> the changes done so far, and also have some suggestions about where
>>> to
>>> go next.
>>>
>>> One of the big problems with understanding Morphic is that, as one of
>>> you guys said a couple of weeks ago, it's limits are unclear. This is
>>> because the framework is often codependent with it's clients (the
>>> morphic applications and extensions).
>>>
>>> This is an area which I would be very happy if you decide to attack,
>>> because every application which is extricated from this mess will
>>> make
>>> Morphic clearer for all of it's users, in addition to users of that
>>> specific application.
>>>
>>> It is also an area in which I hope SpaghettiTracer can significantly
>>> help locate and resolve problems, and I'd be glad to help you try it
>>> out.
>>>
>>> Daniel
>
More information about the Squeak-dev
mailing list
|