DBPL venue: quick responses needed

[James Cheney]

Hi,

After discussions with and VLDB and ICFP workshop chairs, unfortunately, the preferred option of a "virtual" DBPL at both venues joined by a video link turns out not to be  viable (it would cost at least $9000 up-front, which we don't have).  More information about this below, if you are interested.

The two most viable remaining options for DBPL 2015 are:

A) co-located with ICFP 2015, in Vancouver, on September 5,
B) co-located with SPLASH 2015, in Pittsburgh, sometime between October 25-30

ICFP has more or less given us the "green light", if we can decide soon, while we are less certain about SPLASH, so we are leaning towards option A (despite the known disadvantages) unless there is a strong preference for SPLASH.

If you are interested in attending/participating in a DBPL symposium in 2015, please respond to me by *this Friday*, with one of the following responses:

- Neither option A nor B works for me

- Only option A would work

- Only option B would work

- I could attend either option A or B

If you know others (not on this list) who might want to attend, I'm happy to count their votes too, so please forward!

--James

== More information ==

At the Dagstuhl seminar in December, we discussed DBPL 2015 and some complications in organizing it raised by the fact that VLDB 2015 (Hawaii) and ICFP 2015 (Vancouver) overlap, and these are two conferences that are important for potential participants from DB an dPL communities respectively.

We discussed several options:

1. organize "jointly" using a video link: this option seemed to be preferred.
2. organize as part of VLDB: not ideal because difficult for anyone attending ICFP to attend.

3. organize as part of ICFP: not ideal because difficult for anyone attending VLDB to attend.

4. organize somewhere else, such as SPLASH or ECOOP: I didn't feel much enthusiasm for this idea, and there wasn't a general sense that these conferences were ones that significant numbers of potential DBPL participants would normally attend.

Torsten and Fritz volunteered to investigate the options available with the VLDB and ICFP organizers respectively.  Both events' workshop chairs were receptive, but ICFP estimated that the A/V cost for their end alone would be $4500 and we assumed that VLDB would be similar.  This would have added significantly to the cost of the workshop, and in the absence of a sponsor able to provide this at short notice, we had to let the VLDB workshop deadline pass without making a proposal for DBPL to be held there. 

Holding DBPL in Hawaii is problematic because of distance and because VLDBs workshop days are a subset of ICFP's.

Holding DBPL in Vancouver seems viable because it is possible to travel from Hawaii to VLDB overnight September 3-4 or 4-5 (though is a pain) and for US or European participants, this travel will be on the way back home anyway. 

Holding DBPL with ECOOP is no longer possible (their workshop deadline was in January).  SPLASH/OOPSLA is still possible, but we would have to submit a proposal for evaluation first whereas with ICFP the workshop organizers have already given us the "green light", provided we are confident that the event will be successful there.

[Dan Suciu]

Only option A works for me.

A quick thought, before I forget. I encourage you to involve Alvin Cheung in DBPL; the best would be to ask him to give a keynote talk (he can access Vancouver easily from Seattle).

Dan

> --
> You received this message because you are subscribed to the Google Groups "Data-Centric Programming Languages" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to data-centric-progr...@googlegroups.com.
> To post to this group, send email to data-centric...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/data-centric-programming/CAAdciC2P43WKZ6E_sGHW%2BvMWwu%2B5%3DhXzrBBWghSrEnP0oo7UiQ%40mail.gmail.com.
> For more options, visit https://groups.google.com/d/optout.

[James Cheney]

Hi again,

Here's a summary of the 7 responses so far:

5 OK with option A

4 OK with option B, with 2 "maybes"

3 of the above are OK with either option.

So, at the moment option B is potentially slightly better but it's a small sample.

We are aware that an event shortly after VLDB is not ideal; we're trying to establish whether it's better or worse than the SPLASH option. 

So, if you would normally attend VLDB, it would be very useful to hear from you to see whether a DBPL event on September 5th is a dealbreaker (and whether the SPLASH option is any better).

If you would normally attend ICFP, it would be very useful to know whether attending DBPL as a 1-day SPLASH workshop would be realistic, or a dealbreaker.

If you already responded, thanks and sorry for the extra traffic.

--James

[James Cheney]

Hi again,

Here is a summary of responses:

Of 5 responses from people who I'd characterize as "from the DB community", option A worked for 1 person and option B for 2.5 of them.  (I counted "maybes" as 0.5, i.e. probability of attendance is 0.5.)

Of 6 responses from people who I'd characterize as "from the PL community", option A worked for all 6 and option B worked for 4.5.

Extrapolating irresponsibly from this small sample, we can estimate the following probabilities:

P(attend option A | DB background) = 0.2

P(attend option A | PL background) = 1.0

P(attend option B | DB background) = 0.5

P(attend option B | PL background) = 0.75

P(attend option A) = 0.6363...

P(attend option B) = 0.6363...

Overall, my impression from these (way statistically underpowered) estimates is that option A would lead to a "successful" event, with lots of participants who would normally attend ICFP, but not a "successful DBPL event" due to lack of participants who need or prefer to attend VLDB.  So, going with ICFP seems likely to lead to a symmetric situation to past DBPLs, where it is very attractive to one community but not for another.

On the other hand, not everyone in either community is able to attend an event co-located with SPLASH, but it appears to be plausible for a majority of respondents, and not disproportionately terrible for either community.  So, on the basis of this quick and unscientific survey, it sounds like it would be a mistake to proceed with the ICFP option and would be less bad (but still not ideal) to explore the SPLASH option this time around. 

Please let us know if you have any additional information that might affect these conclusions.

--James

On Wed, Feb 11, 2015 at 10:25 AM, James Cheney <james....@gmail.com> wrote: