Why do we need a new journal in systems? The current landscape has many problems: - Top venues like SOSP/OSDI and Eurosys don't do revisions: good work that is not perfect gets rejected and bounced around (wastes both author and reviewer cycles).
-
-
Prikaži ovu nit
-
- Artificial constraints like being single track limit the number of papers accepted and presented - Coupling publication and conference attendance is unfortunate, shutting out authors from countries like Iran and author who simply dont have the money to travel
Prikaži ovu nit -
- Conference papers also have an unfortunate limit on number of pages both at submission and in final camera-ready time. Some conferences try to squeeze authors by asking for money for extra pages in camera-ready version.
Prikaži ovu nit -
I think all of these problems would be solved simply by moving to a journal format. Think SOSP/OSDI, but with as many papers accepted as are deemed high quality with revision! We would still have conferences, but they would be explicitly for networking, with more focus on that.
Prikaži ovu nit -
Why do we need a new journal? Can't we use the ACM journals like Transactions on Computing Systems? No, I don't trust ACM to get this right, and I do think it is important the journal is not tied to an org like ACM. It should be managed by the Systems Research Foundation.
Prikaži ovu nit -
How would this work? 48 senior grad students from all over the world would be selected to be editors, for the period of one or two years. Students would do the reviews for submitted papers. The review quality of a fresh-graduate and senior grad student is not very different.
Prikaži ovu nit -
Students are more enthusiastic, and are more open to suggestions about changing their review style or tone. I truly believe with some training the review quality would be the same as that of SOSP/OSDI, which is the best I've seen.
Prikaži ovu nit -
A sort of revolutionary aspect of this journal review process: - continuous back and forth between authors and reviewers during a one month period! - No rebuttals; instead, using something like HotCRP, authors can ask questions throughput the one month period
Prikaži ovu nit -
The journal would return an accept, revise, or reject decision to the authors in a month Important: the review process will be non-confrontational. The goal of everybody involved is to get the paper to the quality needed to publish: authors can choose if they want to put in work
Prikaži ovu nit -
If the decision is revise, the authors get to submit a revised version in a 1 month to 3 month timeframe, to be evaluated by the reviewers and continue the conversation.
Prikaži ovu nit -
Running this journal will not be expensive. See JMLR: the biggest cost is filing taxes! Apart from the volunteer student editors, there won't be: - copy-editing - type-setting We expect the authors to do this themselves, as is normal in CS.
Prikaži ovu nit -
Obviously, it will take time for this journal to become "prestigious". But there is a clear void for a venue like this. There are many authors frustrated with current system. We would tap into all that. I estimate journal will be considered "top" in 5-7 years.
Prikaži ovu nit -
Finally, the culture in systems seems to be "starting a new thing" rather than "fixing the old thing". OSDI was started because the folks running SOSP were convinced accepting 10 papers every two years was the right thing to do.
Prikaži ovu nit -
If folks are interested in doing this, please ping/email/DM me! Putting together a team and figuring out the details of the journal will be the first step. Doing this will take time, we better get started if we want a good journal in 5 years time :)
Prikaži ovu nit - Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.