[Miwi-middleware] KIARA/Java on github

Philipp Slusallek philipp.slusallek at dfki.de
Wed Sep 24 16:28:36 CEST 2014


Hi Thomas,

I have raised concern regarding the support of current users (within the
FI-PPP!) of existing GEis. This is a real question that we face and that
that people will certainly approach me with. Why do you have a problem
with that very relevant issue?

Your write: "Agreement is to put ALL existing implementations on hold".
I cannot agree with the statement in this strict form for the above
reasons. All I am arguing is that some (smaller) efforts be reserved for
support of existing and actively used technology.

I also made it clear that I fully support the focus of our work on the
Java API as suggested, Just not exclusively. I hope this is fine with all.


Best,

	Philipp




Am 24.09.2014 um 11:04 schrieb Thomas Michael Bohnert:
> Philipp,
> 
> I will stop this discussion here. It is pointless and a waste of time.
> 
> We are a team and we have to consider the exact context and
> organisationl structures and responsibilities.
> 
> As I have said so many times in the past, we all have good reasons why
> this is good and this is bad and this is better than this or that.
> 
> But the point now is to stop this waste of time with arguing but to
> start something that is first and foremost driven by a joint effort and
> team.
> 
> I really hope you can accept this and look forward to your contribution
> as a team member.
> 
> Best - Thomas
> 
> 
> 
> 
> On 09/24/2014 06:39 AM, Philipp Slusallek wrote:
>> Hi all,
>>
>> Sorry Thomas, for some reason I had not seen your email before I sent
>> mine last night. Thanks for the update. Has the text you sent been
>> agreed upon by Chapter 1.8 at the meeting in Madrid or is this more a
>> summary of the discussions that took place at the meeting?
>>
>>
>> Regarding the external expert: I believe Juanjo himself is probably one
>> of the best experts we could find. If he suggested others, this is
>> probably fine as well. Who is it that you are contacting now?
>>
>>
>> While I very much agree (as I already told you) to focus on the Java
>> version, I disagree strongly that we should stop all activities on the
>> existing implementations. We must at least maintain the current
>> technology and support it where necessary.
>>
>> Let's not forget that there are implementations that make active use of
>> KIARA. They are within FI-WARE (Sync GE, CH. 1.5) as well as running
>> demos within the FI-Content and FITMAN projects that make use of the
>> existing KIARA GEi. Even more uses are planned within the next few
>> months.
>>
>> Abandoning these fully working GEs without any technical reasons for
>> doing so, would be damaging to FIWARE and the GE process. As you state,
>> the Java version (especially if we start from an empty page) will take
>> significant time to become available (you have even talked about one
>> year). This is not acceptable for the existing users of the technology.
>> Additionally, in those projects the C/C++ interface is needed, which
>> would likely take even longer to become available.
>>
>> Also I see no impact whatsoever that the "old" implementationcould have
>> on the new Java API. Once we have agreed on a new API for Java, it
>> should not be too difficult to backport that to the current C/C++
>> version (at least if we keep that goal in mind during our discussion,
>> which we should do anyway). Given the flexible design of current KIARA
>> implementation, we should even be able to do so without having to
>> abandon the current API and have them coexist for compatibility until
>> existing users had a chance to switch over.
>>
>> Also there should not be any reason that the new Java version and the
>> "old" implementations cannot interoperate, even while they are using
>> different API approaches. So, I see also no reason that the "old" design
>> would limit how we design the new Java API and implementation (which
>> seems something that Jaime is very worried about).
>>
>> So unless someone comes up with solid technical reasons for abandoning
>> the current work (which I cannot see and which have never been presented
>> since we started our working on KIARA), we should refrain from
>> suggesting to waste EU money that has been spend on the current
>> implementations.
>>
>> To be clear, I am perfectly fine is focusing the current development on
>> Java (as we have started some time ago already), I am just opposing to
>> abandon the existing work as your text seems to suggest.
>>
>>
>> Best,
>>
>>     Philipp
>>
>>
>> Am 23.09.2014 um 12:19 schrieb Thomas Michael Bohnert:
>>> Dear all,
>>>
>>> Nice to see that there is so much energy behind our proposal to start
>>> over with a Java version. If we continue like that there will be soon a
>>> very nice product that serves the needs of our primary customers - that
>>> is the FIWARE community.
>>>
>>> About the next steps.
>>>
>>> First, we need to complete the hand-over from MIWI to I2ND. This
>>> includes all the stuff agreed upon during the WPL/WPA meetings, like new
>>> mailing lists, etc. But this is mostly admin stuff.
>>>
>>> One item is the weeky KIARA meeting. This is now linked to Task 1.8.3
>>> and to be taken over by eProsima (Jaime?). That means to prepare agendas
>>> in advance, take minutes, record action items, report to the WPA/WPL,
>>> and any other task related to the task leadership.
>>>
>>> Regarding the technical directions to be taken. Agreement is to put ALL
>>> existing implementations on hold and to start over with the JAVA
>>> version. Starting point is a blank page but we will carefully analyze
>>> existing code once we have agreed on the design, i.e. issues first and
>>> foremost the KIARA API.
>>>
>>> About the KIARA API. Here the most important issue is to define an API
>>> that meets the expectations of professional software developers. Since
>>> we have very good expertises in the team, with a broad spectrum of
>>> know-how and experiences we need to find a way to consolidate this into
>>> one approach and ultimately specification. One element in achieving this
>>> is to consult an external expert that will take a fresh and independent
>>> view on things and thus secure a design that meets this requirement. The
>>> exact candidate is not yet decided (do you have proposals?) but there is
>>> already one person we know about and we currently are contacting.
>>>
>>> Action items:
>>> Jaime - please prepare an agenda for coming Friday meeting, share it at
>>> least a day before with the team such that we can add items to the
>>> agenda in advance.
>>>
>>> We will discuss any aspect in next Friday's meeting, that is technical
>>> aspects as well as implementation and procedures on task level.
>>>
>>> Thanks for your contributions and commitment in advance!
>>>
>>> Christof and Thomas
>>>
>>> On 09/22/2014 04:15 PM, Jaime Martin Losa wrote:
>>>>      Hi Philipp,
>>>>
>>>>      This time we are going to start with the design, and not with the
>>>> code, and this design should be externally approved. Yes we should
>>>> start with a blank page.
>>>>
>>>>      We didn't developed any document. Basically the Java Idea and the
>>>> external audit has come from Juanjo.
>>>>
>>>>      We are working now in a proposal (design, spec) of API. Next
>>>> Friday meeting, we should start talking about the first design tasks
>>>> of JAVA KIARA.
>>>>
>>>>      Our main goal should be the usability, because we need something
>>>> stable and easy within one year.
>>>>
>>>> Regards,
>>>> Jaime.
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: Philipp Slusallek [mailto:philipp.slusallek at dfki.de]
>>>> Sent: lunes, 22 de septiembre de 2014 16:07
>>>> To: Jaime Martin Losa; Dmitri Rubinstein; KIARA Mailing List
>>>> Subject: Re: [Miwi-middleware] KIARA/Java on github
>>>>
>>>> Hi Jaime,
>>>>
>>>> Java was important for us before and Dmitri has worked on this
>>>> implementation for quite some time now. we are ualready sing this now
>>>> for another project. But changing it for better versions is always a
>>>> good thing and we are very open for it.
>>>>
>>>> Please do not blame us for having already done some of the work that
>>>> is now becoming relevant also for the project as a whole.
>>>>
>>>> You/we do not have to use this directly, but I see this as an
>>>> interesting basis that we can use as a starting point for our
>>>> discussions. This is why we are making it available for all to look at
>>>> (and please do so).
>>>>
>>>> Having something that is already working and is already compatible
>>>> with our previous work is (at least from my point of view) a better
>>>> place to start than a blank page. I hope you and others agree.
>>>>
>>>> As in the C/C++ version we are very open to suggestions and
>>>> alternative interfaces and implementations. So please feel free to
>>>> suggest those, ideally directly with your implementation. Fortunately,
>>>> we will not have the C/C++ issues, so we should be in a much better
>>>> place to collaborate based on comparing each of our concepts and
>>>> actual implementations of them.
>>>>
>>>> I am looking forward to a closer collaboration within FI-Core.
>>>>
>>>> BTW, at the Kickoff you have apparently developed a document
>>>> describing how you (as in KIARA group) would want o work together. We
>>>> have not been in the loop and so it would be good to get that
>>>> document, so we are able to comment (and hopefully just agree to it).
>>>>
>>>>
>>>> Best,
>>>>
>>>>      Philipp
>>>>
>>>>
>>>>
>>>>
>>>> Am 22.09.2014 um 15:37 schrieb Jaime Martin Losa:
>>>>>      Dmitri, we are not going to make the same mistake twice.
>>>>>
>>>>>      First we have to define what we want to do, an API that should be
>>>>> externally approved, the different modules, and then start to assign
>>>>> tasks.
>>>>>
>>>>>      Don't start just writing code without any design consensus and
>>>>> later try the other parties to follow you.
>>>>>
>>>>> Regards,
>>>>> Jaime.
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: miwi-middleware-bounces at lists.fi-ware.org
>>>>> [mailto:miwi-middleware-bounces at lists.fi-ware.org] On Behalf Of Dmitri
>>>>> Rubinstein
>>>>> Sent: lunes, 22 de septiembre de 2014 13:10
>>>>> To: KIARA Mailing List
>>>>> Subject: [Miwi-middleware] KIARA/Java on github
>>>>>
>>>>> I just made KIARA/Java v0.1 available on the github:
>>>>>
>>>>> https://github.com/dmrub/kiara-java
>>>>>
>>>>> License: LGPL version 3
>>>>>
>>>>> Current features:
>>>>>
>>>>> Supported transports: HTTP 1.1, TCP Block Transport
>>>>>
>>>>> Supported protocols: JSON RPC 2.0, JavaObjectStream (Java
>>>>> serialization)
>>>>>
>>>>> JSON RPC 2.0, HTTP 1.1 and TCP Block Transport are compatible to the
>>>>> KIARA C/C++ version.
>>>>>
>>>>> Example applications are in:
>>>>> ./KIARA/src/test/java/de/dfki/kiara/test/AosTest.java
>>>>>
>>>>> Best,
>>>>>
>>>>> Dmitri
>>>>> _______________________________________________
>>>>> Miwi-middleware mailing list
>>>>> Miwi-middleware at lists.fi-ware.org
>>>>> https://lists.fi-ware.org/listinfo/miwi-middleware
>>>>>
>>>>> -----
>>>>> No virus found in this message.
>>>>> Checked by AVG - www.avg.com
>>>>> Version: 2014.0.4765 / Virus Database: 4015/8172 - Release Date:
>>>>> 09/08/14 Internal Virus Database is out of date.
>>>>> _______________________________________________
>>>>> Miwi-middleware mailing list
>>>>> Miwi-middleware at lists.fi-ware.org
>>>>> https://lists.fi-ware.org/listinfo/miwi-middleware
>>>>>
>>>>
>>>>
>>
>>
> 


-- 

-------------------------------------------------------------------------
Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI) GmbH
Trippstadter Strasse 122, D-67663 Kaiserslautern

Geschäftsführung:
  Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender)
  Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats:
  Prof. Dr. h.c. Hans A. Aukes

Sitz der Gesellschaft: Kaiserslautern (HRB 2313)
USt-Id.Nr.: DE 148646973, Steuernummer:  19/673/0060/3
---------------------------------------------------------------------------
-------------- next part --------------
A non-text attachment was scrubbed...
Name: philipp_slusallek.vcf
Type: text/x-vcard
Size: 441 bytes
Desc: not available
URL: <https://lists.fiware.org/private/miwi-middleware/attachments/20140924/a0a7653e/attachment.vcf>


More information about the Miwi-middleware mailing list

You can get more information about our cookies and privacy policies clicking on the following links: Privacy policy   Cookies policy