Re: [ucf-tac] Urban Computing Track at Open Source Summit EU

Hi Lorenzo

I gave this idea some thoughts and it seems it comes with a couple of
difficulties.

> **As first step** we should ask around if somebody in the CG can contact any researcher/student/expert in the field of urban planning or logistics or related fields

I hope anyone from the community, especially those related to academic
environments could provide some help. Anyone's got any idea?

> During our tools-making effort, we have collected some scenarios about how "clients-servers agents" could be imagined as clusters ("Hydra networks") to manage data-exchange among ground sensors or networked machines [1],

I don't know real limitations of such devices (I'm not an expert in
IoT), but I expect a mixture of EXI and MQTT (EXI as an binary XML
format, MQTT as a protocol - HTTP may be a bit to heavy for poor
computing power equipped devices).
The former, while somehow different to what we commonly use (JSON-LD,
Turtle/N3/etc, maybe XML-RDF) is still an RDF serialization of which
support is doable without a bigger effort, including low-level
programming languages.
The latter presents a more complicated puzzle as it is completely
different to HTTP. While hydra was designed to be protocol agnostic,
we never reached a phase where hydra is in use on non-HTTP approach.

> Main challenge in my opinion, with these applications we may step into the world of embedded applications so hardware limitations and software safety requirements may require a campaign of optimisation

My first idea was to provide a "fixed" JSON-LD like context - a
receipt on how an API and hypermedia should be provided so a simple
client can find in the payload what's needed. Something that will get
rid of any RDF specific weirdness (type entailing, value cardinality),
something that will eliminate all those heavy from computing
perspective tasks as iterating through sets, etc.

> and/or porting to lower-level languages

I believe we could setup proper repos on HydraCG's GH to give place
for such development. Unfortunately, while it sounds interesting (and
unless C# and i.e. .net core is considered a lower-level language) I
personally won't be able to help in this process other way than code
reviews and general support. HydraCG, heracles.ts and other topics
occupy me well enough.

I hope those thoughts makes any sense. Feel free to deliberate more so
we can discuss it with other community members (including conf-calls,
which BTW is tomorrow!)

Regards

Karol Szczepanski

pt., 7 cze 2019 o 13:14 Lorenzo Moriondo <tunedconsulting@gmail.com> napisał(a):
>
> Ok, let's try to imagine a two-years road-map that will lead to elaborate a proposal to present to Linux Conference and/or be submitted to the Urban Planning Foundation to demonstrate Hydra capabilities in the field.
>
> **As first step** we should ask around if somebody in the CG can contact any researcher/student/expert in the field of urban planning or logistics or related fields, we need domain experts that are willing to work with us to see if there is viability to such use-cases for Hydra. I am asking around among my contacts so probably you can do the same. **Then** we can start writing a proposal to join the Foundation and/or write a paper/presentation to present in the Conference tracks.
>
> During our tools-making effort, we have collected some scenarios about how "clients-servers agents" could be imagined as clusters ("Hydra networks") to manage data-exchange among ground sensors or networked machines [1], and we have a demo about how a flock of drones with IR sensors can communicate using a simple protocol based on a Hydra API to cover a vast area and find out heat spots that may suggest presence of fire: each drone is a client-server agent that can issue and receive requests, and all these agents are connected to a Hydra server to form a Hydra network [2].
>
> Main challenge in my opinion, with these applications we may step into the world of embedded applications so hardware limitations and software safety requirements may require a campaign of optimisation and/or porting to lower-level languages, and the necessity of finding a standard on our side for storing data with low memory footprint. In the case of Javascript tools, for example, we should consider streamlining compilation via WebAssembly (or using something like NectarLang [3]) for execution on embedded micro-controllers; another example, the Python Heracles-inspired agent (that may be in alpha for September) supports an in-memory graph datastore that can work down to 16Mb of RAM to map a local network of sensors. These considerations may be pretty relevant once trying to port concepts from the Internet to the IoT domain.
>
> **Summing up**, we may want to start collecting all the bibliography and the developing experience of the last years (from all the different teams that have been experimenting with Hydra) in a single document that targets urban planning applications/data with the support of domain experts.
>
> Thanks for your interest and your attention. Let me know your points about.
>
> PS. as Hydra Ecosystem we could have a specs-complete server (hydrus beta) for the end of this Summer of Code.
>
> [1] https://www.hydraecosystem.org/Design
> [2] https://github.com/HTTP-APIs/hydra-flock-demo
> [3] https://nectar-lang.com/
>
>
> On Thu, 6 Jun 2019, 22:02 Karol Szczepański, <karol.szczepanski@gmail.com> wrote:
>>
>> Hi Lorenzo
>>
>> Thanks for the news. Sounds interesting!
>>
>> As for the POC - heracles.ts reference client is on NPM [1] ready for
>> some action. It has some capabilities on board (I'm working on a new
>> hydra console prototype) and I think it could be used for that POC of
>> yours, at least for the client part.
>> I'm not sure about the server side. Do you have any ideas for that?
>>
>> We're working also on more examples and stabilizing the spec to be
>> less testing/more stable so we can think about wider adoption.
>>
>> Feel free to elaborate more on how we, as a community could get involved.
>>
>> Regards
>>
>> Karol Szczepanski
>>
>> [1] https://www.npmjs.com/package/@hydra-cg/heracles.ts
>>
>> czw., 6 cze 2019 o 19:29 Lorenzo Moriondo <tunedconsulting@gmail.com>
>> napisał(a):
>> >
>> > Hi,
>> >
>> > As you may know Linux Foundation, Google and other big players created the Urban Computing Foundation and have added a track to this year's Linux Conference EU. Hydra tools could be players in this field, at least in the future. I suggest you to subscribe in the UC mailing list if you didn't yet, to follow as things are evolving fast, if we can have a complete client-server POC for the end of the year we could propose Hydra for these issues:
>> >
>> > * Mining data from the Internet of Things in urban areas
>> > * Managing urban big data on the cloud
>> > * Knowledge discovery from sensor data for saving energy and resources
>> >
>> > This year Call For proposal ends 11:59 PM PST on Monday, July 1 but at the moment seems too close as we (Hydra Ecosystem) are still engaged in developing at least until September, but if you have something ready it may be a good path to growth. What do you think about? It is a viable path to real-world applications based on Hydra? Involving
>> > Cheers,
>> >
>> > [1] https://uc.foundation/
>> >
>> > Lorenzo Moriondo, from mobile
>> > https://it.linkedin.com/in/lorenzomoriondo
>> >
>> > ---------- Forwarded message ---------
>> > From: Travis Gorkin via Lists.Uc.Foundation <tgorkin=uber.com@lists.uc.foundation>
>> > Date: Wed, 5 Jun 2019, 23:15
>> > Subject: [ucf-tac] Urban Computing Track at Open Source Summit EU
>> > To: <ucf-tac@lists.uc.foundation>
>> >
>> >
>> > Hi TAC Members,
>> >
>> > I wanted to share that the Linux Foundation events team has offered to support a dedicated track for Urban Computing as part of the Open Source Summit EU in Lyon, France!  I've gone ahead and provided them with an abstract and suggested topics for the track (see below).  They have added this to the event's Call for Proposals page here.  As members of the UCF TAC, we will need to review all the proposals and select the presentations to form the track for the summit.
>> >
>> > We should all promote the event and encourage people in our networks to present their work in the Urban Computing space!
>> >
>> > Looking forward to seeing what topics get submitted,
>> > Travis
>> >
>> > Urban Computing Track
>> >
>> > Urban computing combines acquisition, integration, and analysis of big data generated in urban spaces to provide insight into the major issues that cities face around environment, human life quality, and city operation systems.  Sessions in this track are aimed at Engineers, Data Scientists, and Urban Planners interested in applying open source tools and techniques to urban computing and smart city applications.
>> >
>> > * Urban sensing and city dynamics sensing
>> > * Mining data from the Internet of Things in urban areas
>> > * Managing urban big data on the cloud
>> > * Interactive visual data analytics for urban computing
>> > * City-wide traffic modeling, visualization, analysis, simulation, and prediction
>> > * Knowledge discovery from sensor data for saving energy and resources
>> > * Intelligent delivery services and logistics industries in cities
>> > * Machine learning algorithms applied to spatial and spatio-temporal data
>> >
>> > _._,_._,_
>> > _._,_._,_
>>

Received on Monday, 10 June 2019 19:37:29 UTC