CS2125 Paper Review Form - Winter 2018

Reviewer: Mikhail Berezovskiy

Paper Title: When and How to Use Multi-Level Modelling

Author(s): JUAN DE LARA, ESTHER GUERRA, JESUS S ´ ANCHEZ CUADRADO

1) Is the paper technically correct?
 [v] Yes
 [ ] Mostly (minor flaws, but mostly solid)
 [ ] No

2) Originality
 [ ] Very good (very novel, trailblazing work)
 [] Good 
 [v ] Marginal (very incremental)
 [ ] Poor (little or nothing that is new)

3) Technical Depth
 [ ] Very good (comparable to best conference papers)
 [v] Good (comparable to typical conference papers)
 [ ] Marginal depth
 [ ] Little or no depth

4) Impact/Significance
 [ ] Very significant
 [ ] Significant
 [v] Marginal significance.
 [ ] Little or no significance.

5) Presentation
 [ ] Very well written
 [] Generally well written
 [v ] Readable
 [ ] Needs considerable work
 [ ] Unacceptably bad

6) Overall Rating
 [ ] Strong accept (award quality)
 [] Accept (high quality - would argue for acceptance)
 [v ] Weak Accept (borderline, but lean towards acceptance)
 [ ] Weak Reject (not sure why this paper was published)


7) Summary of the paper's main contribution and rationale
   for your recommendation. (1-2 paragraphs)

  According to authors, this paper aims at filling "how and when to use multi-
level modeling" gap by identifying a set of patterns and idioms where the use of 
multiple levels makes sense. They made a vast overview of different modeling 
patterns with their pros and cons. Such patterns as type-object, dynamic features, 
dynamic auxiliary domain concepts, relation configurator and element classification.

  Authors did research on a corpus of 400 models, with a conclusion that 
most of the models are using explicit modeling approach and can be improved 
with multi-level modeling approach. 


8) List 1-3 strengths of the paper.  (1-2 sentences each,
identified as S1, S2, S3.)
 - Paper is well structured, content is sequential and coherent
 - Quite a full description of Two-level and Multi-level architectures. 
    A detailed review of type-object, dynamic features, dynamic auxiliary 
    domain concepts, relation configurator and element classification patterns.
    "When" and "how" to apply.


9) List 1-3 weaknesses of the paper (1-2 sentences each,
identified as W1, W2, W3.)
 - Benefits of multi-level meta-modeling not clearly shown in section of rearchitecting 
    from two to multi-level models, reducing models size are not obvious and require higher 
    abstraction and complexity of modeling.
 - Research of corpus of 400 models doesn't cover problems of two-level modeling, 
    experience, reasons, and consequences. There no metrics of evaluating "better" model, 
    except model size.
 - Despite detailed and structured content, the idea of paper looks as a small feature rather 
    than the new concept of modeling. Since models are focused on writers and readers of models, 
    research should cover at least one of them, and follow their needs.