Hello Kai,
I certainly don't mind you entering the discussion, quite the opposite actually! :)
I would like to see the top-level project address both what you call a "solution space", by providing basic enablers (if I were to make an analogy with Eclipse RT, that would be the enabling component model brought by Equinox, and the implementation of core and compendium OSGi services), as well as a set of specialized/vertical components targeting specific domains such as home automation (in Eclipse RT, that would be higher-level frameworks such as Smila, Riena … that expose tailored services for building search engines, client/server applications, …)
FWIW, tooling for a general purpose language (you are referring to the Lua tools) would probably not be in the scope of the TLP in my opinion – although plug-ins enabling "M2M" development on top of those tools would be (e.g a set of wizards, builders, etc. for making use of the Mihini framework, or any other embedded runtime that would be considered an "M2M" runtime). Note that when I use "M2M" here it's really as a temporary nickname for how we want to call this set of "connecting" technologies :-)
I'd like to hear others' opinion on this, but to me the scope of the top-level project should really encompass more than just what you put behind your definition of M2M (i.e. Connecting devices using unreliable low-bandwidth networks), by also welcoming tools and runtimes that re-use the basic building blocks that I mentioned earlier, but for examples focus on adding more value on the edge, without necessarily relying on "Cloudy Things".
All,
I hope you don't mind if I enter this interesting discussion! Werner's comments seem to bring some general problem to the surface: What is actually the categorization criteria for projects? Looking at what is there, it seems to be technical/developer view of the world: "Here's stuff for the runtime, there is modelling
and over there the tooling." I am wondering whether an m2m top-level project would fit into this scheme - as Werner points out, the current m2m projects are quite diverse (runtimes for servers, runtimes for embedded, protocols, tooling for application development, tooling for a general
purpose language (Lua) etc.). So m2m is for me rather a "solution space" than a technology-defined category. A second issue that I see in Ben's comments is that m2m might be actually a too narrow space for a top-level project. m2m is just one aspect of the IoT (or IoE or however you want to call it), so in a more general way I agree with Ben that it is about
"connecting things". To give you a very concrete example: I, coming from the Smart Home domain, would never feel at home under m2m - although at a first glance things seem to be technologically similar (and I already had many discussions around that with Ben
and Andy), the audience, the solutions and the business models are very different (If you are interested in some more details, have a look at http://kaikreuzer.blogspot.de/2013/05/eclipse-technologies-for-internet-of.html).
I am actually aware of big telcos where the m2m and the smart home departments are both huge, but very much separated from each other - obviously for the reasons mentioned above.
To summarize, I think if there is the need for a new top-level project, it should be "more" than m2m: "IoT", "Connected World", whatever. But then again, it would be probably difficult to find a suitable charter, because the projects can be too diverse
and their goals too different… So maybe just fitting the projects into the current categories is still the best choice?
Just my two cents, Kai
Ben,
While it's still in Proposal phase, the fact, that Concierge is proposed under RT while clearly aiming at M2M raises questions I mentioned yesterday. Would projects like Concierge
belong to a hypothetical M2M/IoT umbrella, RT as currently intended or something else?
Given a majority of projects like Konekki, Minihi, etc. are Lua-based, has anybody thought about an LDT level, Lua Development Tools, like JDT?
Werner On Tue, Jun 18, 2013 at 2:40 AM, <m2m-iwg-request@xxxxxxxxxxx> wrote:
------------------------------
Message: 2
Date: Tue, 18 Jun 2013 02:40:26 +0200
From: UOMo <uomo@xxxxxxxxxxx>
To: m2m-iwg@xxxxxxxxxxx
Subject: Re: [m2m-iwg] M2M Top-Level Project
Message-ID:
<CAAGawe1NzTgnri6nH-zBKSfq=mz5qGSz1HA-4CUoxg7cvYWJvQ@xxxxxxxxxxxxxx>
Content-Type: text/plain; charset="iso-8859-1"
Ben/all,
One more thing if you look at examples like LocationTech (where you're at
least involved as Mentor to some extent) is that similar to PolarSys (which
is even more of a cross-cut trying to cover everything from M2M/Embedded to
Automotive, Modeling, Location/GIS and other verticals like Healthcare)
there are overlaps. STEM once under a top level umbrella that dissolved
later (OHF) became a separate Technology project. If you look at both
LocationTech and PolarSys their broad scope definition and taglines would
theoretically qualify STEM there, too, probably more LocationTech. The only
project under the LocationTech umbrella is an RCP tool not so different
from STEM but for more general purpose location-based modelling (well,
there are several distros, one for Tsunami Prediction and Warning goes in a
similar direction as STEM)
Protocols or standards are not found there right now and similar to e.g.
MQTT there are standard-defining bodies like OASIS or OGC doing just that.
They may interact with top level umbrella projects like LocationTech, while
some members of these organizations probably see less value to join by
themselves.
Papyrus is a tool that seems torn between EMF, the (not much more than
Science it feels) Automotive IWG and PolarSys. That for some tools or
projects is also a risk to keep in mind. Mylyn started as Technology if I
remember correctly (still called Mylar[?]) then moved to Tools and
eventually became an umbrella.
There are various ways to categorize them, but even using the existing
Eclipse umbrella top-level groups more efficiently instead of keeping it
all in Technology could be an option. You referred to Ponte as "fully
fledged server", but Quest so far mostly builds on underlying servers like
Node.js, so I don't think I'd call it RT. while the description of Minihi "The
Mihini project delivers an embedded runtime..." speaks for itself and could
place Minihi under the Eclipse RT umbrella, too.
Ponte has some characteristics of stuff that's in WebTools now as well as
some projects under the SOA Umbrella. Others like Paho may feel like SOA,
too as the focus is on Messaging. While Koneki and other M2M Tooling
projects are certainly something that may feel at home under the Tools
umbrella and its various branches. No significant activity there for 2-3
years now, but Seqoyah and Mobile Tools Project underneath could be a shelf
for M2M related, tools, too. Kepler does not seem to have a "Mobile" distro
any more either. There is a fuzzy "Automotive" one that doesn't seem to
know, whether it's the "C/C++" distro or "Modeling", nor does it show any
clear results of the Automotive IWG as there seems to be little output
there yet. Having an Eclipse distro for M2M development partly to take over
from the Mobile one could be a good step. Maybe for Juno as it looks too
late this year,
Werner
> On Mon, Jun 17, 2013 at 5:47 PM, <m2m-iwg-request@xxxxxxxxxxx> wrote:
>
>> Send m2m-iwg mailing list submissions to
>> m2m-iwg@xxxxxxxxxxx
>>
>> ----------------------------------------------------------------------
>>
>> Message: 1
>> Date: Mon, 17 Jun 2013 06:25:37 -0700
>> From: Benjamin Cab? <bcabe@xxxxxxxxxxxxxxxxxx>
>> To: m2m Industry Working Group <m2m-iwg@xxxxxxxxxxx>
>> Subject: Re: [m2m-iwg] M2M Top-Level Project
>> Message-ID: <CDE4A829.441D9%BCabe@xxxxxxxxxxxxxxxxxx>
>> Content-Type: text/plain; charset="windows-1252"
>>
>> Hi Mike, all,
>>
>> +1, definitely. I think a top-level project would really help to federate
>> M2M projects under a common umbrella.
>> It would be good to discuss what domains we would like the top-level
>> project to cover. At the moment the M2M IWG charter states that it
>> coordinates the work around tools, frameworks and protocols for M2M ; which
>> is probably a good basis for a top-level project charter. However, I think
>> we should be more clear regarding the fact that some projects are targeting
>> constrained embedded devices (Mihini, Paho clients), and some other target
>> full-fledged servers (e.g. Ponte). Also, the M2M IWG as of today is very
>> much focused on enabling M2M by providing the basic building blocks (device
>> management, communication protocols etc.), while I think there should also
>> be room for technologies that are further down the development chain,
>> especially the ones allowing the development of end-user applications
>> making use of M2M data.
>>
>> >From a very high-level perspective, I'd like to have a top-level project
>> with a tagline that could be "Connecting". Connecting machines to machines
>> (embedded agents, protocols, brokers, protocol "adapters"?), connecting
>> developers to machines (tools, debuggers, simulators, code generators ?),
>> connecting end-users to machines (SCADA, visual dashboards, ?). What are
>> your thoughts?
>>
>> Benjamin--
>
>
|
