DConf 2019: Shepherd's Pie Edition

[Nicholas Wilson]

On Sunday, 23 December 2018 at 10:59:32 UTC, Joakim wrote:
>> You say that like some superior technology exists to replace 
>> the conference.
>
> It does, read the first link I gave in my first post above.

You mean the one that says "I don’t know how to fix conferences"?

>> Yes, DConf may benefit from tutorials, workshops, BoFs, 
>> whatever, but the value it brings to the community is very 
>> real.
>
> It may bring some value, but that's not the question: the 
> question is whether we could get more value out of the 
> alternatives, particularly at a cheaper cost? The fact that you 
> and others keep avoiding this question suggests you know the 
> answer.

That really depends on the objective function you mean by "more 
value".
"social networks, Slack groups, podcasts, and YouTube" are all 
well and good but they cannot compare (as in apples to oranges) 
to high-bandwidth low latency personal communication with all the 
people that have an interest (business, technical, whatever)  and 
technical expertise in the subject at hand.

>> Hardly. IME there are two kinds of conferences (or maybe they 
>> form a spectrum, whatever) academic and industrial. Academic 
>> is going nowhere, research needs presenting, organisation of 
>> collaboration needs to happen.
>
> Research conferences are irrelevant. I don't pay attention to 
> them and the fact that the Haskell link Atila gave above says 
> their conferences are for presenting research is one big reason 
> why almost nobody uses that PL in industry.

I concede that I find PL theory useless, but not all academic 
conferences are PL theory, and I don't think that the potential 
scope for more academic talks of DConf is limited to PL theory. 
Novel applications of D in anything from physics to 
bioinformatics to optimisations based on immutability to DSELs 
enabled by D's meta programming are all possible in an academic 
setting.

>> Industrial, there is project coordination, employment 
>> prospectus, business opportunities, why do you think companies 
>> sponsor conferences? They get their moneys worth out of it.
>
> Clearly not in the iOS community, and according to a commenter 
> in my second link above, the Javascript community in his 
> country, as the number of tech conferences is going down a lot. 
> It is my impression that this is true across the board for 
> pretty much every tech community, but I presented that iOS link 
> because he actually tallies the evidence.

I don't doubt those numbers and perhaps the other forms of 
communication do lessen the need for multiple conferences per 
year, but there is a large difference from many to one compared 
to one to zero, in person communication cannot be easily replaced.

Industrial sponsorship is definitely real, take a look at the 
side column of http://llvm.org/devmtg/2018-10/ which I went to 
and talked to the authors of 
https://github.com/wsmoses/Tapir-LLVM for potentially targeting 
OpenMP and other parallel runtimes with dcompute, talked to the 
people developing the SPIR-V target of LLVM, the list goes on. 
I'm going to EuroLLVM (Brussels) to continue those conversations, 
followed straight away by ACCU (Bristol) to give a talk about 
meta programming with D in the context of developing and using 
DCompute. Then a few weeks later I'll be going to DConf for many 
reasons but principally to coordinate development, deal with the 
gripes that have accumulated. I'll probably return home via 
Boston for IWOCL (OpenCL).

>> Perhaps you as an individual believe that they are not cost 
>> effective for you, fine.
>
> As I keep repeating, this is not about me. I'm pointing out 
> trends for _most_ devs,

DConf has been growing in size every year it has been held, as 
have IWOCL and the LLVM conferences. I'm sure some topics for 
some conferences are declining, it may well even be an industry 
wide trend, but I'd bet good money that the new equilibrium will 
have conferences as a staple.

> my own preferences are irrelevant.

I certainly hope not.

>> But consider that the foundation reimburses speakers and I 
>> personally would be very interested to hear what you have been 
>> doing with Andoird/ARM and I'm sure many others would as well, 
>> the question becomes: is it worth your time?
>
> I don't understand what's so special about "speakers" that it 
> couldn't simply reimburse non-speakers that the foundation 
> wants at one of the decentralized locations instead. It seems 
> like the talk is a made-up excuse to pay for some members of 
> the core team to come, when the real reason is to collaborate 
> with them. Why not dispense with that subterfuge?

The talks together with the topic of the conference are what draw 
people to the conference and make it economically viable. It is a 
perfectly rational decision. If I was running a conference trying 
to turn a profit I'd probably get more applications for the 
available speaker slots => better quality speakers => more 
attendees => $$$.

DCompute would not exist were it not for that reimbursement, as a 
poor student that made the difference between this is something I 
can work towards, afford to go to and get good value out of vs 
not. Perhaps we could run general travel grants like LLVM does 
but I don't think we're large enough for that, Mike Parker would 
be the person to talk to about that. But if, like me, they are 
students and wan't to have something to talk about to aid in 
networking, then giving a talk will help with that.

> I see little value in a full talk about a port to a new 
> platform like Android, that is basically another linux distro 
> with a different libc. It's not a matter of my time, I don't 
> think it's worth the audience's time. I wish those organizing 
> DConf would focus on that more.

You can choose the length of the talk you think would fit the 
topic. You could cover the basics of using the port for 
developing Android apps, the difficulties you experienced doing 
the port and the troubles others might have in doing their own, 
... as they say, the stage is yours. It would also present an 
opportunity to convince others of the direction  you think we 
should be going in e.g. w.r.t mobile/ARM/AArch64.