Hi everyone,

Does anyone know of any good references around model transformation

formalisms and in particular the formalisms of model transformed using

languages ``like'' ATL?

What I mean is, given a Model a and Model b (with MMa and MMb) and a

transformation T (from MMa to MMb), is there anything we can say about

MMb with respect to MMa. (Is it some sort of subset, etc...). I am

trying to understand if there is any way of determining formally if a

transformation is possible or impossible.

i.e. If MMb is a subset of MMa is it always possible? etc...

any tips or pointers would be great.

Thanks,

Ian]]>

Hi Ian,

AFAIK, we can consider a transformation T as a partial function (total

or not) defined by the mapping written in ATL for instance. MMa and MMb

only are domain and codomain of this function. I don't think we can

consider more properties between MMb and MMa in the general case.

Hope this helps.

Best regards,

Mikael

Ian Bull wrote:

> Hi everyone,

>

> Does anyone know of any good references around model transformation

> formalisms and in particular the formalisms of model transformed using

> languages ``like'' ATL?

>

> What I mean is, given a Model a and Model b (with MMa and MMb) and a

> transformation T (from MMa to MMb), is there anything we can say about

> MMb with respect to MMa. (Is it some sort of subset, etc...). I am

> trying to understand if there is any way of determining formally if a

> transformation is possible or impossible.

>

> i.e. If MMb is a subset of MMa is it always possible? etc...

>

> any tips or pointers would be great.

>

> Thanks,

> Ian

>

--

Mikaël Barbero - PhD Candidate

ATLAS Group (INRIA & LINA) - University of Nantes

2, rue de la Houssinière

44322 Nantes Cedex 3 - France

tel. +33 2 51 12 58 08 /\ cell.+33 6 07 63 19 00

email: Mikael.Barbero@{gmail.com, univ-nantes.fr}

http://www.sciences.univ-nantes.fr/lina/atl/]]>

Perhaps the following workshop (around verification and validation for MDSD)

is a one-stop-show on current work in the field. Proceedings at

http://www.modeva.org/2007/modevva07.pdf

The charter of that workshop states:

"... Major questions that cross-cut V&V and MDE include: Is the result of a

transformation really what the user intended? Is the model correct with

respect to the expected security, time, and structural constraints? What

models can be used for validation or for verification? Does the

implementation, generated after several model transformations, conform to

the initial requirements? ..."

If I have to choose one paper, my pick is:

Analysis of Model Transformations via Alloy

Kyriakos Anastasakis, Behzad Bordbarand and Jochen M. K]]>

Thanks Mikaël,

Does it necessarily have to be a partial function (i.e. does each

element map to *at most* one element in the codomain). If we use the

impartive part of ATL, can this be broken (because we can always add a

body to a rule that calls other rules to create more than 1 new element)?

Sorry if this doesn't make sense, I have been doing these

transformations for a few years now, but haven't really thought about

them formally until recently.

thanks,

ian

Mikaël Barbero wrote:

> Hi Ian,

>

> AFAIK, we can consider a transformation T as a partial function (total

> or not) defined by the mapping written in ATL for instance. MMa and MMb

> only are domain and codomain of this function. I don't think we can

> consider more properties between MMb and MMa in the general case.

>

> Hope this helps.

>

> Best regards,

> Mikael

>

> Ian Bull wrote:

>> Hi everyone,

>>

>> Does anyone know of any good references around model transformation

>> formalisms and in particular the formalisms of model transformed using

>> languages ``like'' ATL?

>>

>> What I mean is, given a Model a and Model b (with MMa and MMb) and a

>> transformation T (from MMa to MMb), is there anything we can say about

>> MMb with respect to MMa. (Is it some sort of subset, etc...). I am

>> trying to understand if there is any way of determining formally if a

>> transformation is possible or impossible.

>>

>> i.e. If MMb is a subset of MMa is it always possible? etc...

>>

>> any tips or pointers would be great.

>>

>> Thanks,

>> Ian

>>

>

>

>]]>

Thanks Miguel,

These links are great!

cheers,

ian

Miguel Garcia wrote:

> Ian,

>

> Perhaps the following workshop (around verification and validation for MDSD)

> is a one-stop-show on current work in the field. Proceedings at

> http://www.modeva.org/2007/modevva07.pdf

>

> The charter of that workshop states:

>

> "... Major questions that cross-cut V&V and MDE include: Is the result of a

> transformation really what the user intended? Is the model correct with

> respect to the expected security, time, and structural constraints? What

> models can be used for validation or for verification? Does the

> implementation, generated after several model transformations, conform to

> the initial requirements? ..."

>

>

> If I have to choose one paper, my pick is:

>

> Analysis of Model Transformations via Alloy

> Kyriakos Anastasakis, Behzad Bordbarand and Jochen M. Küster

> (pp. 47--56)

>

>

> Miguel

>

>

> "Ian Bull" <irbull@cs.uvic.ca> wrote in message

> news:fjimsk$8mm$1@build.eclipse.org...

>> Hi everyone,

>>

>> Does anyone know of any good references around model transformation

>> formalisms and in particular the formalisms of model transformed using

>> languages ``like'' ATL?

>>

>> What I mean is, given a Model a and Model b (with MMa and MMb) and a

>> transformation T (from MMa to MMb), is there anything we can say about MMb

>> with respect to MMa. (Is it some sort of subset, etc...). I am trying to

>> understand if there is any way of determining formally if a transformation

>> is possible or impossible.

>>

>> i.e. If MMb is a subset of MMa is it always possible? etc...

>>

>> any tips or pointers would be great.

>>

>> Thanks,

>> Ian

>>

>

>]]>

Ian,

In fact, I was a bit simplifying the problem. First of all, I only

consider *declarative* part of ATL (imperative part only is for some

very specific issues that can not be deal with declarative part and is

not very interesting for reasoning on transformations).

I was also simplifying the case where a rule match only 1 model element

(despite ATL can match more than one) and create one element. I agree

this is over simplification.

If we consider declarative ATL in the large (n,n), we are facing a n-ary

relation. This property may not be very interesting by itself and may

not allow to extract derived properties from the transformation.

I hope this will be helpful to you.

Best regards,

Mikael

Ps: It makes sense ;) but i'm not very formally thinking on model

transformations neither.

Ian Bull wrote:

> Thanks Mikaël,

>

> Does it necessarily have to be a partial function (i.e. does each

> element map to *at most* one element in the codomain). If we use the

> impartive part of ATL, can this be broken (because we can always add a

> body to a rule that calls other rules to create more than 1 new element)?

>

> Sorry if this doesn't make sense, I have been doing these

> transformations for a few years now, but haven't really thought about

> them formally until recently.

>

> thanks,

> ian

>

> Mikaël Barbero wrote:

>> Hi Ian,

>>

>> AFAIK, we can consider a transformation T as a partial function (total

>> or not) defined by the mapping written in ATL for instance. MMa and

>> MMb only are domain and codomain of this function. I don't think we

>> can consider more properties between MMb and MMa in the general case.

>>

>> Hope this helps.

>>

>> Best regards,

>> Mikael

>>

>> Ian Bull wrote:

>>> Hi everyone,

>>>

>>> Does anyone know of any good references around model transformation

>>> formalisms and in particular the formalisms of model transformed

>>> using languages ``like'' ATL?

>>>

>>> What I mean is, given a Model a and Model b (with MMa and MMb) and a

>>> transformation T (from MMa to MMb), is there anything we can say

>>> about MMb with respect to MMa. (Is it some sort of subset, etc...).

>>> I am trying to understand if there is any way of determining formally

>>> if a transformation is possible or impossible.

>>>

>>> i.e. If MMb is a subset of MMa is it always possible? etc...

>>>

>>> any tips or pointers would be great.

>>>

>>> Thanks,

>>> Ian

>>>

>>

>>

>>

--

Mikaël Barbero - PhD Candidate

ATLAS Group (INRIA & LINA) - University of Nantes

2, rue de la Houssinière

44322 Nantes Cedex 3 - France

tel. +33 2 51 12 58 08 /\ cell.+33 6 07 63 19 00

email: Mikael.Barbero@{gmail.com, univ-nantes.fr}

http://www.sciences.univ-nantes.fr/lina/atl/]]>

Thanks again.

I assumed you were thinking the declarative part, but I just wanted to

make sure.

The reason I was asking is because I am just finishing my dissertation

on using MDE to create interactive visualizations of information. I

have several MM for for viewers (graphs, charts, heatmaps, etc...) and

using model transformations I can specify how the visualization should

work. I know this works in practice, but I wanted to give some advice

on how well it works in theory (can this always be done, etc...). There

are some cases when I have to use operational (imperative)

transformations, and it usually is a result of needing multiple elements

in the visualizations for a single source.

Thanks for all your help. I did find some good links on this stuff as

well, I will get them organized and post them here in case others have

questions too.

Cheers,

ian

Mikaël Barbero wrote:

> Ian,

>

> In fact, I was a bit simplifying the problem. First of all, I only

> consider *declarative* part of ATL (imperative part only is for some

> very specific issues that can not be deal with declarative part and is

> not very interesting for reasoning on transformations).

>

> I was also simplifying the case where a rule match only 1 model element

> (despite ATL can match more than one) and create one element. I agree

> this is over simplification.

>

> If we consider declarative ATL in the large (n,n), we are facing a n-ary

> relation. This property may not be very interesting by itself and may

> not allow to extract derived properties from the transformation.

>

> I hope this will be helpful to you.

>

> Best regards,

> Mikael

>

> Ps: It makes sense ;) but i'm not very formally thinking on model

> transformations neither.

>

>

> Ian Bull wrote:

>> Thanks Mikaël,

>>

>> Does it necessarily have to be a partial function (i.e. does each

>> element map to *at most* one element in the codomain). If we use the

>> impartive part of ATL, can this be broken (because we can always add a

>> body to a rule that calls other rules to create more than 1 new element)?

>>

>> Sorry if this doesn't make sense, I have been doing these

>> transformations for a few years now, but haven't really thought about

>> them formally until recently.

>>

>> thanks,

>> ian

>>

>> Mikaël Barbero wrote:

>>> Hi Ian,

>>>

>>> AFAIK, we can consider a transformation T as a partial function

>>> (total or not) defined by the mapping written in ATL for instance.

>>> MMa and MMb only are domain and codomain of this function. I don't

>>> think we can consider more properties between MMb and MMa in the

>>> general case.

>>>

>>> Hope this helps.

>>>

>>> Best regards,

>>> Mikael

>>>

>>> Ian Bull wrote:

>>>> Hi everyone,

>>>>

>>>> Does anyone know of any good references around model transformation

>>>> formalisms and in particular the formalisms of model transformed

>>>> using languages ``like'' ATL?

>>>>

>>>> What I mean is, given a Model a and Model b (with MMa and MMb) and a

>>>> transformation T (from MMa to MMb), is there anything we can say

>>>> about MMb with respect to MMa. (Is it some sort of subset,

>>>> etc...). I am trying to understand if there is any way of

>>>> determining formally if a transformation is possible or impossible.

>>>>

>>>> i.e. If MMb is a subset of MMa is it always possible? etc...

>>>>

>>>> any tips or pointers would be great.

>>>>

>>>> Thanks,

>>>> Ian

>>>>

>>>

>>>

>>>

>

>

>]]>

That topic has to do with another ongoing PhD blogged at

http://visual-languages.blogspot.com/

And there's a very short (2-page) paper also touchign that topic,

"Mapping visual notations to MOF compliant models with QVT Relations"

The URL is http://portal.acm.org/citation.cfm?id=1244228

So I guess we all want to read your dissertation once it's ready :-)

Miguel

"Ian Bull" <irbull@cs.uvic.ca> wrote in message

news:475EB085.8030502@cs.uvic.ca...

> Thanks again.

>

> I assumed you were thinking the declarative part, but I just wanted to

> make sure.

>

> The reason I was asking is because I am just finishing my dissertation on

> using MDE to create interactive visualizations of information. I have

> several MM for for viewers (graphs, charts, heatmaps, etc...) and using

> model transformations I can specify how the visualization should work. I

> know this works in practice, but I wanted to give some advice on how well

> it works in theory (can this always be done, etc...). There are some

> cases when I have to use operational (imperative) transformations, and it

> usually is a result of needing multiple elements in the visualizations for

> a single source.

>

> Thanks for all your help. I did find some good links on this stuff as

> well, I will get them organized and post them here in case others have

> questions too.

>

> Cheers,

> ian

>

> Mika]]>

Thanks Miguel,

This is a great reference.

I am doing my research in a group focused more on Information

Visualization than MD(E/A/D). It has been a good battle learning all

the modeling technologies, but a lot of fun. My primary research

objective is to determine how MD(E/A/D) can be used in the creation and

customization of interactive information visualization.

For someone learning modeling on their own, newsgroups like this have

been a great resource.

cheers,

ian

Miguel Garcia wrote:

> Ian,

>

> That topic has to do with another ongoing PhD blogged at

> http://visual-languages.blogspot.com/

>

> And there's a very short (2-page) paper also touchign that topic,

> "Mapping visual notations to MOF compliant models with QVT Relations"

> The URL is http://portal.acm.org/citation.cfm?id=1244228

>

> So I guess we all want to read your dissertation once it's ready :-)

>

>

> Miguel

>

>

> "Ian Bull" <irbull@cs.uvic.ca> wrote in message

> news:475EB085.8030502@cs.uvic.ca...

>> Thanks again.

>>

>> I assumed you were thinking the declarative part, but I just wanted to

>> make sure.

>>

>> The reason I was asking is because I am just finishing my dissertation on

>> using MDE to create interactive visualizations of information. I have

>> several MM for for viewers (graphs, charts, heatmaps, etc...) and using

>> model transformations I can specify how the visualization should work. I

>> know this works in practice, but I wanted to give some advice on how well

>> it works in theory (can this always be done, etc...). There are some

>> cases when I have to use operational (imperative) transformations, and it

>> usually is a result of needing multiple elements in the visualizations for

>> a single source.

>>

>> Thanks for all your help. I did find some good links on this stuff as

>> well, I will get them organized and post them here in case others have

>> questions too.

>>

>> Cheers,

>> ian

>>

>> Mikaël Barbero wrote:

>>> Ian,

>>>

>>> In fact, I was a bit simplifying the problem. First of all, I only

>>> consider *declarative* part of ATL (imperative part only is for some very

>>> specific issues that can not be deal with declarative part and is not

>>> very interesting for reasoning on transformations).

>>>

>>> I was also simplifying the case where a rule match only 1 model element

>>> (despite ATL can match more than one) and create one element. I agree

>>> this is over simplification.

>>>

>>> If we consider declarative ATL in the large (n,n), we are facing a n-ary

>>> relation. This property may not be very interesting by itself and may not

>>> allow to extract derived properties from the transformation.

>>>

>>> I hope this will be helpful to you.

>>>

>>> Best regards,

>>> Mikael

>>>

>>> Ps: It makes sense ;) but i'm not very formally thinking on model

>>> transformations neither.

>>>

>>>

>>> Ian Bull wrote:

>>>> Thanks Mikaël,

>>>>

>>>> Does it necessarily have to be a partial function (i.e. does each

>>>> element map to *at most* one element in the codomain). If we use the

>>>> impartive part of ATL, can this be broken (because we can always add a

>>>> body to a rule that calls other rules to create more than 1 new

>>>> element)?

>>>>

>>>> Sorry if this doesn't make sense, I have been doing these

>>>> transformations for a few years now, but haven't really thought about

>>>> them formally until recently.

>>>>

>>>> thanks,

>>>> ian

>>>>

>>>> Mikaël Barbero wrote:

>>>>> Hi Ian,

>>>>>

>>>>> AFAIK, we can consider a transformation T as a partial function (total

>>>>> or not) defined by the mapping written in ATL for instance. MMa and MMb

>>>>> only are domain and codomain of this function. I don't think we can

>>>>> consider more properties between MMb and MMa in the general case.

>>>>>

>>>>> Hope this helps.

>>>>>

>>>>> Best regards,

>>>>> Mikael

>>>>>

>>>>> Ian Bull wrote:

>>>>>> Hi everyone,

>>>>>>

>>>>>> Does anyone know of any good references around model transformation

>>>>>> formalisms and in particular the formalisms of model transformed using

>>>>>> languages ``like'' ATL?

>>>>>>

>>>>>> What I mean is, given a Model a and Model b (with MMa and MMb) and a

>>>>>> transformation T (from MMa to MMb), is there anything we can say about

>>>>>> MMb with respect to MMa. (Is it some sort of subset, etc...). I am

>>>>>> trying to understand if there is any way of determining formally if a

>>>>>> transformation is possible or impossible.

>>>>>>

>>>>>> i.e. If MMb is a subset of MMa is it always possible? etc...

>>>>>>

>>>>>> any tips or pointers would be great.

>>>>>>

>>>>>> Thanks,

>>>>>> Ian

>>>>>>

>>>>>

>>>>>

>>>

>]]>