Group Actions
Engagement
  • Author Author: mouch
  • Date Created: Date Created
  • Views 455 views
  • Likes 1 like
  • Comments 1 comment
Related
Recommended

Standing on the Shoulders of Giants!

mouch

Researchers publish for a variety of reason, suchs as claim to a discovery, sharing their findings, cordon of their research area so that everyone knows what they are working on so that they will look in a different direction, allowing them to travel and speak on conferences and basically their work is based on what their pre-decessors have done, hence: Standing on the shoulders of giants.

 

In various papers I read, there is information missing about what the researcher has done to achieve their findings, for example: I am currently reading papers (I am not going to name or shame anyone because I do not think that is in anybodies interest in this case) about dc-dc converter topologies and I find in a lot of papers that often control parameters have been left out as well as input settings etc. The reason for this - I assume: limited space in the paper or . Although, I do not see why a little table providing all the parameters cannot fit in the paper anymore. I fact, I think a table providing all parameters should be a must for every publication. If this type of information can be freely left out we would still believe that neutrinos come in before knocking.

 

On the flip side, in a recent article in the IEEE Control Systems Magazine (the rise and fall of popular control systems) it is argued that people do not read far enough back in to the future and claim to have made discoveries that actually have already appeared in earlier papers.

 

The supposedly advantage of published papers is - in my opion as follows - someone has done research so you don't have to start at the bottom, instead you can stand on his shoulders. This would work so much better if published papers have a dedicated parameter section, which will allow you to -for example- build the simulation in matlab/simulink and enter the parameters and done. And then you can play with the simulation, research it and build on it.

 

The main benefit of this would be for starting researchers who might not be interested in the nitty gritty of the internal workings but on designing further control strategies.

<edit>

Not only would this benefit the starting researcher in this way but also in the case where someone has made a big error (as in one of the articles I read lately) which confused me because in theory it would make sense but in practise more was needed to a) understand the paper and b) the results were wrong. And because of the wrong results I had to learn what went wrong and as such it was great learning but not really standing on the shoulders of giants.

</edit>

 

my2p

 

 

As I mentioned, I do not believe in naming and shaming instead I want to glorify some papers I was generally impressed with:

 

GIAOURIS, D., BANERJEE, S., ZAHAWI, B. & PICKERT, V. (2008) 'Stability Analysis of the Continuous-Conduction-Mode Buck Converter Via Filippov's Method.' Circuits and Systems I: Regular Papers, IEEE Transactions on, 55 (4), pp. 1084-1096.

 

GIAOURIS, D., MAITY, S., BANERJEE, S., PICKERT, V. & ZAHAWI, B. (2009) 'Application of Filippov method for the analysis of subharmonic instability in dc–dc converters.' International Journal of Circuit Theory and Applications, 37 (8), pp. 899-919.

 

The above does not only go for papers and proceedings but also for books, if you want to know about dc-dc converters then read the following:

 

ERICKSON, R. W. & MAKSIMOVIC, D. (2001) Fundamentals of power electronics. 2nd ed. Boston, Mass. ; [Great Britain], Kluwer Academic Publishers.

 

TSE, C. K. (2004) Complex behavior of switching power converters. Boca Raton, CRC Press.

Parents

  • I agree, many young engineers do not look at old research as much as they should.

    Personnally, I find a much better level of detail in older papers than a lot of the new papers.  In part, because the researcher writing the paper assumes that the reader is already deeply familiar with the background.  I think this approach is wrong.  To be fully useful, you should always provide at least a quick overview of the subject so that you can set up a context in which the reader and assess your claim of improvement.

     

    I have read too many papers that present their approach and then leap to the "Quit Everything is Done" (QED) without really providing the data needed to reach that point.

    I also dislike papers that include hundreds of references.  Yes it may show that you at least looked at a lot of past research, but it is more useful to cite those that contributed to your idea and led you to your new development.  The term "Keep It Simple" KIS should always be invoked.

     

    So if you are writing research papers, keep it simple, set up a context for your idea and then include data that demonstrates your achievement.  Only then can you postulate other potential applications for you idea.

    Papers writen just to keep your publication count up are generally useless to working engineers who really could benefit from your detailed understanding of the technical issues.  You just need to simplify the information so they can quickly grasp what you did and can agree that you have a good idea.

     

    Just a thought,

    DAB

Comment

  • I agree, many young engineers do not look at old research as much as they should.

    Personnally, I find a much better level of detail in older papers than a lot of the new papers.  In part, because the researcher writing the paper assumes that the reader is already deeply familiar with the background.  I think this approach is wrong.  To be fully useful, you should always provide at least a quick overview of the subject so that you can set up a context in which the reader and assess your claim of improvement.

     

    I have read too many papers that present their approach and then leap to the "Quit Everything is Done" (QED) without really providing the data needed to reach that point.

    I also dislike papers that include hundreds of references.  Yes it may show that you at least looked at a lot of past research, but it is more useful to cite those that contributed to your idea and led you to your new development.  The term "Keep It Simple" KIS should always be invoked.

     

    So if you are writing research papers, keep it simple, set up a context for your idea and then include data that demonstrates your achievement.  Only then can you postulate other potential applications for you idea.

    Papers writen just to keep your publication count up are generally useless to working engineers who really could benefit from your detailed understanding of the technical issues.  You just need to simplify the information so they can quickly grasp what you did and can agree that you have a good idea.

     

    Just a thought,

    DAB

Children
No Data