[LLVMdev] LLVM Conference 2007 ?

[Domagoj Babic]

Hi all,

On 11/28/06, Reid Spencer <rspencer at reidspencer.com> wrote:
> * Venue: West Coast, USA. Probably either San Francisco Bay Area or
> Seattle
> * Time Frame: Post 2.0 release, summer 2007
> * Topics: Anything related to use or development of LLVM

I'd vote for Seattle. The costs of organization should be lower than in
the Bay Area.

> * An indication of how likely you would be to attend (0=no, 5=maybe,
> 10=have my ticket)

Right now I'm not sure. I might be away during (some part of) summer.

> * What you could offer (present a paper or topic, run a workshop, teach
> a tutorial, etc.)

I'd suggest a workshop with proceedings. Everyone submits a short paper
(6 pages), and gives a talk. There could be several tracks: development,
optimizations, static analysis, ... Depending on the number of
accepted/submitted papers, we should organize paper reviews.

Reviews are a double edge sword: If all the papers get in, the
conference becomes uninteresting for publishing. So, it is very
likely that you'll see a sharp decline in the number of submitted
scientific papers later (if the conference becomes a regular annual
event).

On the other hand, since this is the first LLVM conference, I'd suggest
to keep the acceptance rate moderately high (50%).

To assure a high number of quality papers, I'd strongly encourage you
to make the conference more general than just LLVM-based developments.
I'd suggest that optimization and static analysis tracks also accept
papers that are not strictly based on LLVM, but could provide valuable
ideas for further LLVM development, or are interesting from software
analysis perspective.


As I suggested in the channel, a poll on llvm.org could actually be a
simpler and faster way to gauge the interest and expectations.

Regards,

    Domagoj Babic