Live data from Hacker News

The Glacial Pace of Scientific Publishing

fasebj.org

1–10 of 35 posts

Re: The Glacial Pace of Scientific Publishing

#5
Huge irony in Leslie Vosshall's suggestions. She is a HHMI investigator and has "made it". Telling others not to try for the top 3 journals which is the EXACT path she used to get to her spot is insulting. Basically don't do all this stuff she did to get ahead because you'll make the system worse.

It's like Mark Zuckerberg telling young programmers, don't start an internet company because society is suffering from all the evils of internet.

Re: The Glacial Pace of Scientific Publishing

#7
I'm glad I'm in programming languages, where we usually publish in conferences instead of journals and you can get a paper out in 7 months if you want top-tier, 4 months if you want mid-tier, or as little as 3 months if you just want it published somewhere respectable.

And really you can shorten those timelines by a couple of months because people make their papers available as soon as they're publicly accepted.

Re: The Glacial Pace of Scientific Publishing

#8

Perhaps this problem is worse among biomedical journals? I've routinely had publications in ACS journals that go from submission to acceptance within two months, usually closer to one---I hear the same from colleagues.

yeah, is worst for biomedical journals.

Re: The Glacial Pace of Scientific Publishing

#10

Perhaps this problem is worse among biomedical journals? I've routinely had publications in ACS journals that go from submission to acceptance within two months, usually closer to one---I hear the same from colleagues.

Yes, I'm not sure what other fields may be affected, but I can confirm that this is a huge issue in biomedical journals. It's especially a problem for bioinformatics, where the combination of multi-year publication latency and the typical fast pace of software development means that tools are often long obsolete by the time they are published, and this results in researchers relying on publications and choosing algorithms that are years out of date.

For example, if a researcher chooses an algorithm with a "fairly recent" publication date of 2 years ago, that paper may have been first written 3 years prior to that, after a 1-year process of developing the algorithm and another 2-year process of generating and analyzing a dataset that showcases the algorithm so they can get it published. So that "2 year old" algorithm is actually 2 + 3 + 1 + 2 = 8 years old. If the algorithm is in a fast-moving field like next-generation sequencing analysis, it is probably obsolete several times over (i.e. the tool that obsoleted it is already itself obsolete), if it isn't completely useless now due to increases in dataset size during the past 8 years (e.g. it can only handle 1 million sequences, but today's datasets are 100 million sequences).

Any bioinformatics programmer knows this, of course, and only relies on the publication record as a last resort for information about bioinformatics algorithms. But a biologist who has a dataset they want to analyze and is looking for an algorithm to use probably won't know this, and will end up selecting a hopelessly outdated algorithm as a result.

Post reply on HN