[e2e] double bland reviewing

RJ Atkinson rja at extremenetworks.com
Wed May 12 12:23:00 PDT 2004


On May 12, 2004, at 12:49, Mark Allman wrote:
> Note that I did not intend to be quite that crass in my assessment and
> the above is a little out of context.  The above comment is not to say
> that reviewers do not strive to understand new and interesting 
> concepts.
> The above is to say that if your paper is so horribly written and
> presented that I cannot possibly understand the new concept you're
> trying to impart then I'll recommend the paper be rejected (e.g., some
> plots in papers I review are utterly unreadable -- what am I to do to
> get the results the plots purport to present?  should I give the 
> authors
> the benefit of the doubt and let the paper slide through?  that seems 
> to
> me to not be doing my job as a reviewer).  I am not trying to "destroy
> what I cannot comprehend".  I am trying to remind authors that the 
> point
> in publishing *should* be to impart new insights on the world.  And, if
> you cannot do so in a clear way then your paper is useless to the
> world.

I'm with Mark on this.

I've reviewed an unusually high (for me) number of papers lately.
I am surprised at the percentage of unclear, poorly written papers
crossing my desk (entirely putting aside questions of technical
substance or relevance for the moment).  This includes basic stuff
like having grammar more or less correct, having at least subject and
verb in the purported sentences, and so forth.  I have to believe
that some of these papers were quickly dashed off and sent in
-- without the author getting some colleague of the author to read 
through the paper once before submission to double-check readability.
While I've not been on program committees, my own limited experience
seems consistent with Craig Partridge's assertion that the bottom 50%
of submissions are clearly awful and need a lot of work.

It is very legitimate to review a paper and suggest a full rewrite with
2nd full review cycle when the paper is so poorly written that the 
content
is unreadable or does not get explained in a clear logical manner.

Note well that this issue is different from (1) being unable to 
understand
the concept despite a clear presentation and also different from
(2) disagreeing with the concepts asserted.  Authors need to ensure that
their papers (including figures and graphs) are readable and that the 
paper's
organisation is reasonably sensible and clear.

I'm reviewing one paper right now where it is pretty clear to me that
at least one of the listed authors could not have actually read through
the submitted draft, which is curious (to my mind) and yet another 
problem.

Note well that I'm not a huge fan of any of the reviewing systems in
common use.  For many of the "blind" papers I've reviewed through the
years, I could tell who the author(s) was(were) by the end of the first
reading of the paper.  So I don't think "blind" reviewing actually works
as its advocates assert.  I do find the notion of signed reviews 
intriguing,
though I wonder whether or not it would be a net improvement.  And I
very much like Jon Crowcroft's idea of providing an optional 
non-publication
forum for reviewing papers and providing feedback to authors.

Sigh.

Ran



More information about the end2end-interest mailing list