Smalltalk › Pharo › Pharo Smalltalk Users

Ring Vs Ficus

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

4 messages Options

abdelghani ALIDRA

Jun 06, 2015; 7:10am

Ring Vs Ficus

43 posts

Hi Guys,

I believe the Ring metamodel has been proposed 4 or 5 years earlier.

Is it acctually in use in the current pharo tools? which ones?

I have been told that Martin has develloped a different model : Ficus

What are the limitations that Martin tried to fixed?

What can be further unhanced?

Thanks.

Abdou.

Uko2

Jun 06, 2015; 7:13am

Re: Ring Vs Ficus

1537 posts

As far as I remember Ficus relies on Ring and is a model for handling history. It’s used by Epicea which is a new changes framework…

Uko

On 06 Jun 2015, at 09:26, abdelghani ALIDRA <[hidden email]> wrote:

Hi Guys,

I believe the Ring metamodel has been proposed 4 or 5 years earlier.
Is it acctually in use in the current pharo tools? which ones?

I have been told that Martin has develloped a different model : Ficus
What are the limitations that Martin tried to fixed?
What can be further unhanced?

Thanks.

Abdou.

stepharo

Jun 06, 2015; 1:15pm

Re: Ring Vs Ficus

5159 posts

In reply to this post by abdelghani ALIDRA

Hi abdelghani

Welcome to this list :)

Hi Guys,

I believe the Ring metamodel has been proposed 4 or 5 years earlier.

Is it acctually in use in the current pharo tools? which ones?

Ring should be improved (probably with stubs) to support that we can create a method and that it does not create
automatically the class of the method.

Ring proposed strategy to query the code of a method (the one currently being in the image), or the method represented in the past or another one.
and we should have a look at it.

I have been told that Martin has develloped a different model : Ficus

Martin explained to me that ficus is built on two parts: one that describe the model and they other the "body " of an entity (data)
and such data can be shared between multiple versions.

What are the limitations that Martin tried to fixed?

What can be further unhanced?

Thanks.

Abdou.

tinchodias

Jun 17, 2015; 2:40pm

Re: Ring Vs Ficus

624 posts

Upppssss, I just see this email. Since Abdelghani visits RMoD, I've discussed with him in person about this topic.

Ring-Ficus is an alternative implementation to Ring (as it is in the Pharo image). Like Ring, Ficus-Ring provides models that represent system's code entities, and they are also polymorphic with them. For example, RFiClass shares the API of Class (partially) . The main differences are:

- closed-universe approach: a Ring-Ficus code entity will always answers a Ring-Ficus code entity (in Ring it is not always like that)

- efficient for importing and querying whole Pharo environments: for example, you can import all Pharo 40xxx code entities and query them, e.g. find all the implementors or senders of a selector at each version.

Martin

On Sat, Jun 6, 2015 at 3:34 PM, stepharo <[hidden email]> wrote:

Hi abdelghani

Welcome to this list :)

Hi Guys,

I believe the Ring metamodel has been proposed 4 or 5 years earlier.

Is it acctually in use in the current pharo tools? which ones?

Ring should be improved (probably with stubs) to support that we can create a method and that it does not create
automatically the class of the method.

Ring proposed strategy to query the code of a method (the one currently being in the image), or the method represented in the past or another one.
and we should have a look at it.

I have been told that Martin has develloped a different model : Ficus

Martin explained to me that ficus is built on two parts: one that describe the model and they other the "body " of an entity (data)
and such data can be shared between multiple versions.

What are the limitations that Martin tried to fixed?

What can be further unhanced?

Thanks.

Abdou.

... [show rest of quote]