The questions posed are somewhat vague, Marcel, but we shall try to understand and answer them anyway. :-) Organizing systems research from top (general) to bottom (detailed, specific) appears to me to be not very fruitful. I have indeed seen systems research built on big "boxes" of dynamic (or other) systems interacting in various ways. BUT: if you don't understand exactly what is going on in each box first, you will have NO clue what happens at the even bigger system level, because the interactions among the more detailed systems are not fully understood. So: the lesson is to make sure sufficient information is generated at the bottom level first. 

- Is it wise to switch from one model system to another?

- Is it wise to cite research concerning model systems you don't know, accepting that different organisms have different rules of functioning you don't really know?

A great deal of collaborative international research on global environmental change is coordinated through four Global Environmental Change Research Programmes and the Earth System Science Partnership. In light of the need for an overarching set of solution-focused and integrated research priorities for these institutions, the International Council for Science (ICSU) and the International Social Science Council (ISSC) carried out a consultative process to rethink the focus and framework of Earth system research.

http://www.icsu.org/future-earth/media-centre/relevant_publications/Science_article_12_11_2010.pdf

Dear Marcel,

It's not wise to switch from one model system to another. Because time, efforts, and cost invested by the researcher will go waste.

Before hand, better to choose team of researchers, clear your research objectives, extensive literature review, make yourself clear about your research direction --- then choose the appropriate research model system or customise the model according to specific requirements of research project. 

The questions posed are somewhat vague, Marcel, but we shall try to understand and answer them anyway. :-) Organizing systems research from top (general) to bottom (detailed, specific) appears to me to be not very fruitful. I have indeed seen systems research built on big "boxes" of dynamic (or other) systems interacting in various ways. BUT: if you don't understand exactly what is going on in each box first, you will have NO clue what happens at the even bigger system level, because the interactions among the more detailed systems are not fully understood. So: the lesson is to make sure sufficient information is generated at the bottom level first. 

Yes, Dear Marcel. Scientific-based research is a very slow process that requires long time consuming and a great tenacity and perseverance. Sometimes we think that we are losing our time when we test wrong paths or we do not see any visible advance through our research, but we are just establishing the bases that will be indispensable for the new real developments. With our long time consumed and efforts, sometimes we achieve results that we initially thought that would not be possible of being reached. Of course that this does not apply when the objective is “fast food” type research and only metrics.

@António Manuel Abreu Freire Diogo: Scientific-based research is a very slow process that requires long time consuming and a great tenacity and perseverance. 

A key component of scientific research is time.    Often, quite a bit of time is needed to obtain a satisfactory result in solving a research problem.    For those of us who have been fortunate to be given the opportunity to do research, the time component is built into the research process.

Remarkably, the seemingly long time spans between research satisfactory results is in keeping with growth processes in nature.    A good example is a lily bulb, which is planted in the fall, letting a lily bulb build up energy during the winter months and then in the springtime a lily bulb keeps pace with rising temperatures and increased sunlight to produce a blossom.    This growth process is analogous to what a research experiences from the start of the work on a research problem to the discovery of a solution.    Initial satisfactory results on a new research problem can sometimes take several or more months to achieve a good result.

More to the point, a research that works within a well-known framework (process model) has a better chance of solving a research problem in a reasonable amount of time.  

The second key component in the research process mentioned by @António  is patience and perseverance.   One must have sufficient patience and perseverance (steady work) to achieve a good result.

Dear @Marcel, never change the model, that is my suggestion. As my predecessors have mentioned, (@Subhash, @Anthonio, @James...) time is the key issue, so researchers should be patient,persistent, and optimistic. Faith in the research result must exist.

@Ljubomir:Faith in the research result must exist.

I agree.    In terms of faith in a research result, we can carry your very good idea a step further.    Faith in a result result is a byproduct of demonstrating (proving) that a result satisfies the constraints on a research problem.    Such a result satisfies the axioms in the space in which a result is found.    By space, I mean, for example, a metric topological space endowed with one or more relations such as a proximity relation or a tolerance relation.    

I have understood dear @James. Constraints do always exist and optimization process in research is needed in order to satisfy these constraints.

To understand the functioning and evolution of a local model system with unique spatiotemporal characteristics, observation and experience will ultimately be more important than the knowledge from libraries dealing with model systems that express other characteristics?

I did not down-vote, but I think people need field experience, not network experience, to get familiar with study system, right?

In an overambitious publication in a so-called top journal, key words are so general that they can cover many, often irrelevant, research field/publications complicating literature searches to find the relevant information? 

If you use a longer list of key words, you should master the topic better, right?

Each book is unique in expression, so if you use as key words all the different words  found in a book 'X' (e.g. key word 'the' can be used only once), you will only find that book 'X', but then this would imply you know the content of book 'X'?

Thus, how to find an appropriate list of key words between a single word (e.g. 'the') and all the different key words reflecting the content of a whole book (e.g. 1000 different words in a book 'X' used as key words)? 

Journals ask to indicate let say '6'

You mean, the English part of the data base, also accepting that other languages can use 'the' within words (e.g. menthe)....... (?)

What is the sense of familiarity with a model system accepting that people do not even truly know their own body they experience every day?