Eclipse Community Forums: ATL » Invitation to Engage in a Study

Home » Modeling » ATL » Invitation to Engage in a Study

Invitation to Engage in a Study [message #1186174]

Thu, 14 November 2013 09:29

Eclipse User

Hi,

I'm doctorate student and now I need to conduct a study with ATL users.

I would like to invite you to engage in this study. For this, you only need to have knowledge on ATL transformations. Please, let me know if you are interested and I will give you information about the study.

Your participation is very important to the conclusion of my work.
Thank you for your attention.

Best regards,
Andreza

Re: Invitation to Engage in a Study [message #1194231 is a reply to message #1186174]

Mon, 18 November 2013 06:59

Eclipse User

Sounds interesting. What about just posting (a link to) further information right here?

Re: Invitation to Engage in a Study [message #1201018 is a reply to message #1194231]

Thu, 21 November 2013 09:35

Eclipse User

Roger that, put up a link Smile

Re: Invitation to Engage in a Study [message #1202735 is a reply to message #1201018]

Fri, 22 November 2013 04:38

Eclipse User

Any details ? Razz

Re: Invitation to Engage in a Study [message #1202993 is a reply to message #1202735]

Fri, 22 November 2013 07:24

Eclipse User

Hi,

I proposed a technique for measuring the change impact in ATL transformations. This technique is comprised by a set of metrics that together calculate the impact of changes in ATL transformations. These metrics use some characteristics of the transformations and a relevance (weight) must be assigned for each one. Therefore, this study is to define these relevances.

I will ask the participants to define a relevance in the range between 0 and 1 for a set of characteristics. That's basically it.

I will apreciate your participation Smile

Thank you.

Re: Invitation to Engage in a Study [message #1203206 is a reply to message #1202993]

Fri, 22 November 2013 09:38

Eclipse User

Hi,

Here is the form of the study:
docs.google.com/forms/d/1Y4wzYEAVxhKNuD5JMiO3CtDfTgbi5KWWNH7XBsgXjKA/viewform

It will take about 30 minutes.

I will apreciate your participation.

Thank you.

Re: Invitation to Engage in a Study [message #1209031 is a reply to message #1203206]

Mon, 25 November 2013 04:55

Eclipse User

Hi Andreza,
It was a slow Monday morning so I started answering your questionaire Wink

I don't think I get it to be honest, so I stopped half way. What are you after here?

You want to know the relevence of certain ATL constructs, but they are all relevent. Every choice I make in using an ATL construct has an impact. Every single one of the constructs you discuss I have used because I have needed them to build a given solution. I of course have a preference as to which solution I will use in a given situation but the survey does not consider this. Somethimes I must use the less elegent solution in ATL because the problem dictates this through some complexity. For example I'd always used matched rules if I could but this is not always possible. Should I score a certain construct lower than another because it is less elegent? How is this fair if this construct is in fact the only way of achieving what I desire.

I can't see how you will distill any useful information form the survey. Or maybe I am missing something?

Regards,
Ronan

Re: Invitation to Engage in a Study [message #1209283 is a reply to message #1209031]

Mon, 25 November 2013 07:33

Eclipse User

Hi Ronam B,

First, thank you for helping me with my study Smile

All the elements in an ATL transformation are relevant, but in this study I need to know how relevant they are in respect to the impact they can cause in a transformation (because of the concepts they have). For instance, if I change a rule that "receives parameters" it would cause high impact in the transformation because this rule can be invoked by others (dependencies) or because other reasons you think.

You must think this way, for instance: if a rule has parameters, how relevant it is within the transformation? If it is changed, the impact would be low or high due to this characteristic (have parameter)?

You will not score a certain construct lower than another because it is less elegant.

If it is not clear, please let me know.

Thank you.

Re: Invitation to Engage in a Study [message #1209545 is a reply to message #1209283]

Mon, 25 November 2013 10:22

Eclipse User

Hi Andreza,
Nope I still don't get it. Are you trying to assess the maintainability/evolvability of an ATL Transformation. For example, how easy it is to write a transformation that will be easy to update without breaking existing users of the transformation (for example a library transformation that helps with UML profile handling)? I don't see how ATL is of interest here. You could ask the same about any language.

Perhaps you are trying to reveal the benefits of a declarative langauge compared to an imperative one? If so there is no survey required as the benefits are obvious and stated elsewhere.

I think if your survey is in fact needed it needs to define exactly what your asking otherwise you will get useless results. You of course don't need to bias the survey by giving away too much but the assumptions must be clear.

Regards,
Ronan

Re: Invitation to Engage in a Study [message #1209592 is a reply to message #1209545]

Mon, 25 November 2013 10:51

Eclipse User

Hi Ronam B,

Quote:

Are you trying to assess the maintainability/evolvability of an ATL Transformation. For example, how easy it is to write a transformation that will be easy to update without breaking existing users of the transformation (for example a library transformation that helps with UML profile handling)? I don't see how ATL is of interest here. You could ask the same about any language.

Yes, I'm trying to assess the maintainability/evolvability of ATL Transformations. It is exactly as you said.

I agree with you, I could ask the same about any language. But each language has its own constructs and some characteristics are different from one language to other. My study is in the scope of ATL transformations.

Quote:

Perhaps you are trying to reveal the benefits of a declarative langauge compared to an imperative one? If so there is no survey required as the benefits are obvious and stated elsewhere.

That is not the point.

I tried to answers yours questions. Please let me know any doubt.

Thank you.
Best regards,
Andreza

Re: Invitation to Engage in a Study [message #1219907 is a reply to message #1186174]

Fri, 06 December 2013 10:49

Eclipse User

Andreza, I have just taken part in your survey after solving an ATL (actually an Ecore) issue that has worried me for more than a year.
I have to agree with Ronan that the relevance of most features is high - it would not work without any of them. So, it is difficult to answer your questions, but as you sometimes specify them further (attribute of model type or of a primitive type), it was possible for me to decide differently sometimes.

Can you already tell us when the results and their examination will be available? Do you think that a sufficient number of people is goint to take part?

Re: Invitation to Engage in a Study [message #1219908 is a reply to message #1219907]

Fri, 06 December 2013 11:05

Eclipse User

Hi Gunnar Arndt,

First, thank you for helping me with my study.

Maybe next week the results will be available. I will notify here.

Only 9 persons engaged in this study until now. I will accept answers until next monday (09 December).

Previous Topic:	[EMFTVM] Accessin output elements' attributes
Next Topic:	[ATL] Load the genmodel instead of the Ecore file?

Goto Forum:

-=] Back to Top [=-

Current Time: Wed Jul 23 07:21:27 EDT 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter