On selecting a publication venue by Mark Handley

It REALLY matters where you publish you work!

There are two main reasons for this.

 - You want to influence other researchers with your ideas.  This is
one of the main ways you have impact (and if you don't aim have
impact,  there's no point in doing research).  You can only influence
smart people if they've heard of your work.  There are so many places
to publish, but there's only so much time, so the people you want to
influence only attend a handful of conferences each year, and look at
the proceedings for a handful too.  If you don't publish in the places
smart people frequent, no-one will ever hear about your work again.

 - When you look for a research job in academia or a top lab, they
will use your publication record as a proxy for how smart you are.
It's easy to publish lots of papers.  Even machine-generated papers
can get into some conferences: http://pdos.csail.mit.edu/scigen/  So
no-one much cares how many papers you have; they care where those
papers are.  This is because it's hard to get published in top venues,
so this is a first-pass approximation for how
clever/useful/informative your ideas are (yes, it's a flawed measure,
but everyone uses it anyway).  So for a lectureship here, I'd
interview someone with a SIGCOMM, an NSDI and a OSDI paper, but I
wouldn't interview someone with 50 papers in obscure venues unless
there was something else exceptional about them.

OK, so where should you publish?

Here's some good advice about conferences vs journals:
http://people.csail.mit.edu/mernst/advice/conferences-vs-journals.html

Then you want to choose a high impact venue if the work is strong.
Sometimes you just want to get an idea out there quickly, and then a
workshop is a really good idea.  Some workshops actually  have all the
right people there who you want to influence - HotNets is a good
example.   There's no reason to publish in third-tier conferences
though.  You could have better spent the writing time doing better
work.

Citeseer has a list of venues by impact:
http://citeseer.ist.psu.edu/impact.html
It's generally OK, though a little out of date (for example, NSDI is
top-tier in my book, but doesn't appear), and it doesn't list most
workshops.

Generally speaking, when I do when I'm appraising a conference or
workshop is to see who is on the programme committee.  If more than
50% of the people on the PC are people I've heard of and respect, then
it's probably a good venue.  If the fraction is very small, steer
clear.