Publish a Survey Paper Based on Literature Survey - The Good, The Bad, & The Ugly

Yesterday, I reviewed a survey paper which was perhaps done by a student of a graduate program just after his/her literature survey. How do I know? Because it was just a summary of a set of papers, and lacked a critique, comparison, identification of burning questions/challenges, and any insights for future research directions. My decision as reviewer was easy, "Strong Reject", as several such papers I rejected in the past. Today, one of the students told me "I hope to publish 2 papers based on my research as mentioned in one of the seminars. First, as a survey paper after literature survey and second after finishing research".

The idea of trying to publish a survey paper just after a literature survey looks so wrong to me. While I have had a couple of debates regarding this with my colleagues (both who agree & disagree), I guess it's time to document this so that I don't need to repeat myself again.

Let's start with the most important person when it comes to writing, the Reader. While not many read today, ones that read are overloaded with enough low-quality contents. Alternatively, enough sources are ready to publish low-quality contents (even IEEE Xplore is not an exception). Following are Reader's objectives of a good Survey Paper:

• To be a good introductory material on a given topic
• Should have new interpretations of results, points to make, or new conclusions to draw
• Cover broad range of key and connected material
• Cover each material to a sufficient depth, rather than just summarizing
• Compare and critically evaluate existing work based on related material and writer's expertise
• Identify burning questions/challenges and insights for future research directions
• Confidence that the material is reliable, as it is prepared by experts in the field who have a thorough understanding of the topic and who bother to go and find the details

While it is hard to filter out good contents, at least tools like Google Scholar help us by ranking publications based on citation count (where other people have already filtered out good vs. bad for you). Another option is to search for reputed journals, magazines, or researchers by names. Unfortunately, not many peer reviewers seem to be helping out in filtering these low quality survey papers, though that's in their job description.

Then comes the Writer. As most graduate programs want students to have a couple of publications, writers are pressed to publish something, somewhere. Following is the Good, Bad, and  Ugly side of Writer:

The Good	The Bad	The Ugly
Get to practice academic/research writing	Paper likely to get rejected	Talking of things that you know but don't understand yet (you only read, but never played with ideas to really understand them)
Get feedback from supervisor and reviewers	Harsh review could demotivate researcher	Misrepresentation as an expert when you are an novice
Learn conference/journal process	Yet another low quality survey paper that doesn't address Reader's objectives	Bad reputation for supervisor
If lucky, get a publication	Hardly gets a citation	Repeated attempts could get blacklisted by journal/conference

If a student wants to ever appear as a researcher, the Bad and the Ugly outweigh the Good. Even if the student is not, supervisor who is already having a career as a researcher don't want the Ugly.

While it is up to the supervisor and student to decided, I guess it's very clear when a student just start he/she does not have the maturity, expertise, and perspective to write a good survey paper. They may write a good survey paper by the time they graduate, if they have done substantial work. Based on my own experience, survey paper I wrote after 4 years of PhD work to date has 45+ citations, which I consider as a better metric than paper count...