Talk:Feature model

Suggestions for what else should be covered

* Visualization of Feature Models
* From propositional formula to a feature model (Czarnecki, Wasowski 07; Janota, Kuzina, Wasowski 08)
* Probabilistic feature models (Sample spaces and feature models: There and back again, K Czarnecki, S She, A Wasowski)

• Can tools be classified somehow? I am looking for a Web-based feature modeler and I would like to avoid clicking on 20 links to find the right one. — Preceding unsigned comment added by 217.162.99.24 (talk) 18:44, 6 March 2019 (UTC)

About the authors of the page

The work on this page was started by Don Batory, a renown researcher in the field of Software Product Lines. Mikolas Janota and Sergio Segura has provided most of the text up till now; both are doing a PhD in computer science. I believe that the citations of Don's and Mikolas's (mine) papers are justified.

The section "Configuration" is a focus of mine and it might be the case that some topics of similar relevance has not been covered. Hence suggestions and contributions are welcome. Mikolasj (talk) 16:06, 23 April 2009 (UTC)

Where's the discussion?

Looking at some of the edit comments ("Don comments added", "Thomas suggestions added") it seems as if the discussion about this article is led elsewhere - likely off-wiki, maybe via email or company communication tool. May I suggest you use this page to discuss the article?--Pgallert (talk) 11:03, 22 April 2009 (UTC)

I also noticed that all the editors to this page so far seem to be single purpose accounts (see WP:SPA) and that may mean that there may be some degree of sockpuppetry/meatpuppetry going on. I can't see any bias atm but just thought that this should be pointed out. Sorry forgot to sign this before Smartse (talk) 17:14, 23 April 2009 (UTC)

Naming

Wikipedia recommends using lowercase in titles and section names. `Feature Model' should be `Feature model'; `Software Product Line' should be `Software product line', and so on. The following is especially relevant: "Outside of Wikipedia, and within certain specific fields (such as medicine), the usage of all-capital terms may be a proper way to feature new or important items. However these cases are typically examples of buzzwords, which by capitalization are (improperly) given featured status." Note the word `outside'. Rgrig (talk) 22:00, 22 April 2009 (UTC)

Conflict of interest

googling "feature model segura" reveals a conflict of interest. Smartse (talk) 17:15, 23 April 2009 (UTC)

I'm little puzzled, Segura has worked and published articles about feature models, so did Janota (me), and Batory. I understand that this may be seen as a conflict of interests but on the other hand who else should be writing about it than people that do research on the topic? Mikolasj (talk) 17:41, 23 April 2009 (UTC)

WP:COI clearly states that you are allowed to cite yourself, provided that the paper appears in a peer-reviewed venue. I think COI is not an issue here. What can be an issue is neutrality (are you giving proper coverage or is the article biased towards your 'slice'?) and notability (is this a topic of interest to many more people than you, the authors?). Not being an expert I can't judge, but the fact that half of the citations are self-citations is a definite warning sign. If you include more sources this will help establish both neutrality and notability. (Disclaimer: MikolasJ is a friend of mine. I'm commenting here because I'm more used to Wikipedia policies and guidelines than him. I am not an expert in this domain; in fact I couldn't force myself to read the whole article :p) Rgrig (talk) 23:19, 23 April 2009 (UTC)

Your comment is absolutely to the point, Mikolasj, and to clarify for this article: None of the editors of this article is currently accused of any policy violation whatsoever. However, there is a fine line between "the experts write on the subject" (which is perfectly fine) and "the topic would not exist without the three expert authors" (which is not ok in WP as content should appear here only after the knowledge it represents has been tested by the scientific community). I know it is not nice to hurl policy links at new editors, but if you want to read about that regulation, it is here.--Pgallert (talk) 08:01, 24 April 2009 (UTC)

This is my main concern too. You may be experts in the field but is this topic really ready for inclusion in an encyclopedia if only the experts are creating a page? I've noticed that the articles that this article is related to, e.g. software product lines and Product Family Engineering are tagged and need improving. Perhaps it would be better if the editors of this page concentrated on these articles instead of their own field of research. This would remove any possible COI problems. Smartse (talk) 13:43, 24 April 2009 (UTC)

I don't understand your concern Smartse. You are concerned that the page is created by experts. I would be concerned if it was not created by experts. I dare to suggest that most WP readers are not interested in articles written by dilettantes. I wholeheartedly agree that the other articles need work, unfortunately they need a lot of work and I think they need a complete rehaul and at the moment I don't have time for such.Mikolasj (talk) 17:15, 27 April 2009 (UTC)

It seems to me that "the topic would not exist without the three expert authors" is ruled out simply by the fact that there's at least another 10ppl involved in the cited articles. Note also that the seminal article is from 1990, i.e., very old given the computer science perspective.Mikolasj (talk) 17:15, 27 April 2009 (UTC)

Let me speak as a researcher not involved in the writing of this article, but involved in researching feature models. The topic of feature models has not been created by Mikolas and the two other people. There is at least in the order of 100 research papers published on this topic, and the original proposal goes back to reference no 1 (published in 1990s). Feature models prominently feature (pun intended) in the Generative Programming undergraduate textbook by Czarnecki and Eisenecker published around year 2000 (BTW. I do think that it is a serious omission that this work is not cited - this is probably the single most popular work read about feature modeling, also very approachable). Feature models are well presented in several series of research conferences, even the most prominent ones (like ICSE). Another proof for this is an extensive list of feature modeling tools, some commercial, found in the article. Need for the article is clear - I had my students complaining about lack of it multiple times. Wazow (talk)

Now let me speak as an (inactive, mostly Polish) wikipedia editor, I do have full understanding for the worries about COI, and I think the article could benefit from being edited by another group of people. But I do not think any more drastic steps other than waiting until the other researchers join in are necessary (we just need to find time for this!) Wazow (talk) —Preceding undated comment added 19:58, 27 April 2009 (UTC).

Hmm I'm not sure about this so I'm going to post some info on the conflict of interest noticeboard so that some other editors can taka a look. Smartse (talk) 20:31, 27 April 2009 (UTC)

Sorry I didn't join to the discussion before. Firstly, I would like to thanks Smartse for his helpful comments and recommendations for improvement. Regarding the conflict of interest, I would like to remark that the reason of citing one of my own papers is because it is a survey, the only one in the field of automated analysis as far as I know. It provides a detailed overview of the analysis of feature models and quotes the most important works in the field so far. I firmly believe it is a helpful starting point for those interested in the analysis of feature models. Having said this, I guess I could provide some other references to meet neutrality requirements but I will have to find time for that. Segura09

About the relevance of the topic for the WP community I totally agree with Mikolasj and Wazow comments. Few will dispute the fact that feature models are nearly a 'de facto' standard for variability modeling in product lines. You can also use Google scholar to obtain some evidences of this: "feature models" product lines returns more than 2000 entries. Segura09

Finally, I would like to remark that our contributions to the page are just a starting point. We are aware of the needs for improvements and refinements. Undoubtedly, adding more references and subtopics (visualization, refactoring, applications, etc.) would improve the work notably. However, this cannot be done only by three people (we simply don't have time for this). We just hope that this initial effort may encourage other people to contribute and improve the work little by little. I would say this is how this work ;) Segura09 —Preceding undated comment added 19:05, 28 April 2009 (UTC).

The discussion can be found here Wikipedia:Conflict_of_interest/Noticeboard#Feature_model Smartse (talk) 23:17, 27 April 2009 (UTC)

About the Reference for "Feature Tree"

The page uses the term "Feature Tree" (refering to the Feature Diagrams) and includes a reference to a type of diagram that does not correspond.

• The linked page (http://www.bawiki.com/wiki/techniques/feature-tree/) presents a graph instead a tree
• The page mention the steps to create the Feature Trees proposed by Seilevel (http://www.seilevel.com/business-analyst-resources/business-requirements-models-templates/feature-tree/).
• These feature trees, probably based on Feature Models, do not correspond to the typically used in the Feature Models.

For instance, the Seilevel Feature Trees do not include mandatory or alternative features. These models are used to represent hierarchies of features and cannot be analyzed using the mentioned techniques or libraries.

Probably, we can include an additional section with "related models":

• We can include the "Feature Trees" proposed by Seilevel (with "three levels" and similar to the Ishikawa/fishbone diagrams)
• and the model used in the Classification_Tree_Method used in Testona (http://www.testona.net/en/, http://www.systematic-testing.com/documents/qualityweek1995_1.pdf) for test design.

Jaime.chavarriaga (talk) 15:47, 6 February 2016 (UTC)