Gender Diversity and CS Conferences
Last week in Barcelona, I had an interesting conversation with Koen De Bosschere, the editor of ACM Transactions on Architecture, Compilers, and Optimization (TACO), about the computer science publication process. His contention is that the CS community's publication process, which treats conferences as the primary and certainly the most prestigious publication venues, is a major cause of gender inequality in the CS academic and research community. Because women are primarily responsible for child care, they quickly realize that regular travel to conferences is incompatible with a family but is necessary for career advancement. Hence many women leave the research field. In addition, conferences put researchers at non-US schools at a disadvantage because of the cost in money and time of attending conference that primarily occur in the US, and because of the airplane travel, is not sustainable.
Koen argued for making fast-turnaround journal publication (as he is doing with TACO) the primary venue for research papers and creating 3 annual CS conferences (in the US, Europe, and Asia) for the community to get together and enjoy the social and technical interactions that occur at conferences.
My initial reaction was bemused horror that anyone would suggest getting rid of conferences, which have been my primary focus of publication and technical exchanges for my career. But, the more I thought about it, the more value I saw in the idea. The conference culture started in a different time, when CS was a new and rapidly evolving field; journal publication was a slow, often-multiyear process; and communication was limited to mail, telephone, and email. None of this is still true, except that CS continues to evolve rapidly.
We can now publish papers as quickly or even quicker in journals than conferences, and conferences such as VLDB and OOPSLA have blurred the distinction between conference and journal reviewing. Why not shift publication to journals and move to a few big meetings? I think there will still be room for workshops and symposia, where a focused group of participants spend their time and working together, but I could easily live without the larger "conferences" such as PLDI, POPL, ASPLOS, ISCA, HPCA, PPoPP, ...
Will this solve the gender problem in CS? Not by itself, but it clearly could help make the field more family friendly by allowing researchers of both sexes to travel less and still publish in the top venues.
This change, like open publication, is obviously a bit challenge to the budgets of professional organizations such as ACM and IEEE, which rely on conference revenue. But, the financial model is going to change anyways, so why not factor this improvement into the planning for the new business model?
Koen argued for making fast-turnaround journal publication (as he is doing with TACO) the primary venue for research papers and creating 3 annual CS conferences (in the US, Europe, and Asia) for the community to get together and enjoy the social and technical interactions that occur at conferences.
My initial reaction was bemused horror that anyone would suggest getting rid of conferences, which have been my primary focus of publication and technical exchanges for my career. But, the more I thought about it, the more value I saw in the idea. The conference culture started in a different time, when CS was a new and rapidly evolving field; journal publication was a slow, often-multiyear process; and communication was limited to mail, telephone, and email. None of this is still true, except that CS continues to evolve rapidly.
We can now publish papers as quickly or even quicker in journals than conferences, and conferences such as VLDB and OOPSLA have blurred the distinction between conference and journal reviewing. Why not shift publication to journals and move to a few big meetings? I think there will still be room for workshops and symposia, where a focused group of participants spend their time and working together, but I could easily live without the larger "conferences" such as PLDI, POPL, ASPLOS, ISCA, HPCA, PPoPP, ...
Will this solve the gender problem in CS? Not by itself, but it clearly could help make the field more family friendly by allowing researchers of both sexes to travel less and still publish in the top venues.
This change, like open publication, is obviously a bit challenge to the budgets of professional organizations such as ACM and IEEE, which rely on conference revenue. But, the financial model is going to change anyways, so why not factor this improvement into the planning for the new business model?
Comments
Post a Comment