proposal for EvoOnt/BOM ontology improvements

[Valentin Vlasceanu]

Hello everyone,

My name is Valentin Vlasceanu and I'm a PhD student at
TELECOM&Management SudParis. The subject of my thesis is “Bugtracking”
and I'm also involved in the Helios project, on which I'm working
closely with Olivier Berger.
As part of the project we thought of improving the EvoOnt ontology
(more precise the BOM ontology - http://www.ifi.uzh.ch/ddis/evo/) with
bits of information that we consider important for our work, but might
be useful to someone else too.
Thus, we consider that a “product” is made of “component”s and that
each “component” can contain one or more “software packages” (which
can also might be compiled for different distributions). Then, every
“component” and “software package” has it's own version. This leads
to two possible relationships:
a) Issue / Component : an “issue” belongs to a “component” or a
“component” has an “issue”;
b) Issue / ComponentVersion : an “issue” can be found or fixed in a
precise component version.
You can find a more detailed diagram and information on the project's
wiki: https://picoforge.int-evry.fr/cgi-bin/twiki/view/Helios_wp3/Web/HeliosBtOntology.
We would appreciate any thoughts or remarks regarding our proposal and
of course if there are other relationships or elements that you might
consider important and that we have missed.

Thank you and I await your feedback on this topic,

Valentin Vlasceanu

[Henry Story]

Btw, I will be in Paris for from next week for a couple of weeks at
least.
So perhaps we can meet up then.

Henry

[Olivier Berger]

Hi.

Let me add a few bits, reminding you that we mentioned previously other
additions that we propose, that addressed the modeling of links between
bugs.

The ultimate goal that we seek is to address a large scale application
on bug facts modeling, in the context of the global Open Source
ecosystem.

Here, we need to model the fact that bug reports/Issues are linked to
distribution packages (more than to source files), and for instance that
"similar" bugs are filed in different bugtrackers (for the various
distributions that package a same software, or for the upstream
projects), and such bugs need to be identified and monitored by
maintainers if they want to not reinvent the wheel.

We're working on implementing a first real application making use of
this with Debian and Mandriva bugs triplified, and I hope we can have an
interesting demo in the next month.

Maybe we're moving again even closer to a new revision of the "baetle
ontology", with all these additions to the original EvoOnt BOM ?

Looking forward to reading your feedback.

Best regards,

Le mercredi 16 septembre 2009 à 06:29 -0700, Valentin Vlasceanu a
écrit :

> As part of the project we thought of improving the EvoOnt ontology
> (more precise the BOM ontology - http://www.ifi.uzh.ch/ddis/evo/) with
> bits of information that we consider important for our work, but might
> be useful to someone else too.
> Thus, we consider that a “product” is made of “component”s and that
> each “component” can contain one or more “software packages” (which
> can also might be compiled for different distributions). Then, every
> “component” and “software package” has it's own version. This leads
> to two possible relationships:
> a) Issue / Component : an “issue” belongs to a “component” or a
> “component” has an “issue”;
> b) Issue / ComponentVersion : an “issue” can be found or fixed in a
> precise component version.
> You can find a more detailed diagram and information on the project's
> wiki: https://picoforge.int-evry.fr/cgi-bin/twiki/view/Helios_wp3/Web/HeliosBtOntology.

--
Olivier BERGER <olivier...@it-sudparis.eu>
http://www-public.it-sudparis.eu/~berger_o/ - OpenPGP-Id: 1024D/6B829EEC
Ingénieur Recherche - Dept INF
Institut TELECOM, SudParis (http://www.it-sudparis.eu/), Evry (France)

[Olivier Berger]

Le mercredi 16 septembre 2009 à 18:39 +0200, Henry Story a écrit :
> Btw, I will be in Paris for from next week for a couple of weeks at
> least.
> So perhaps we can meet up then.
>

Sure.

Will you attend the Open World Forum, which could be a good occasion to
discuss bugs and SemWeb things ?

Regards,

[Chaudhry Muhammad Aftab Iqbal]

Hi all,

On Wed, Sep 16, 2009 at 5:51 PM, Olivier Berger <olivier...@it-sudparis.eu> wrote:

Hi.

Let me add a few bits, reminding you that we mentioned previously other
additions that we propose, that addressed the modeling of links between
bugs.

The ultimate goal that we seek is to address a large scale application
on bug facts modeling, in the context of the global Open Source
ecosystem.
 

Here, we need to model the fact that bug reports/Issues are linked to
distribution packages (more than to source files), and for instance that
"similar" bugs are filed in different bugtrackers (for the various
distributions that package a same software, or for the upstream
projects), and such bugs need to be identified and monitored by
maintainers if they want to not reinvent the wheel.

We're working on implementing a first real application making use of
this with Debian and Mandriva bugs triplified, and I hope we can have an
interesting demo in the next month.

We submitted a paper last year in SWESE 2008(Semantic Web Enabled Software Engineering) workshop where we have proposed the idea of customizing / enhancing web applications (e.g. Bugtracker) to provide useful information (in this case : bugs) which are not explicitly mentioned or interlinked. We have developed a prototype version for JIRA bug tracker which enhance JIRA bug tracker by providing small fancy widgets which collects related information about a bug (e.g, mentioning of a bug-id in  a comment or in summary or in description of other bug) and as well as look for dependent bugs which are explicitly mentioned accross bug tracker and provide all these information as an integrated view in a bug tracker against a bug. please find an attached snapshot that will give an overview about it.

Maybe we're moving again even closer to a new revision of the "baetle
ontology", with all these additions to the original EvoOnt BOM ?


Looking forward to reading your feedback.

I would highly appreciate for furhter discussion and possible extension or usage of such interfaces or applications on big datasets.
 
regards
Aftab Iqbal

[Henry Story]

On 16 Sep 2009, at 18:58, Olivier Berger wrote:

> Le mercredi 16 septembre 2009 à 18:39 +0200, Henry Story a écrit :
>> Btw, I will be in Paris for from next week for a couple of weeks at
>> least.
>> So perhaps we can meet up then.
>>
>
> Sure.
>
> Will you attend the Open World Forum, which could be a good occasion
> to
> discuss bugs and SemWeb things ?

Ah yes, thanks for reminding me. I did not get time to contact them yet.
http://openworldforum.org/

I will be (talking perhaps) tomorrow Saturday at the Social Web Camp
in Paris

http://barcamp.org/SocialWebCampParis1

Those who are in Paris should come.

Henry

[Olivier Berger]

Le vendredi 18 septembre 2009 à 16:08 +0200, Henry Story a écrit :

> I will be (talking perhaps) tomorrow Saturday at the Social Web Camp
> in Paris
>
> http://barcamp.org/SocialWebCampParis1
>
> Those who are in Paris should come.
>

Unfortunately, those work weeks are tiring, and I cannot afford adding
more geek stuff during this weekend :-(

Too bad, and apparently, the max # people was reached already.

I hope we'll have other occasions... and in the meantime, I'll try and
move on on implementing stuff on bugs and RDF ;)

Best regards,

[Olivier Berger]

Le vendredi 18 septembre 2009 à 12:30 +0100, Chaudhry Muhammad Aftab
Iqbal a écrit :

>
> We submitted a paper last year in SWESE 2008(Semantic Web Enabled
> Software Engineering) workshop where we have proposed the idea of
> customizing / enhancing web applications (e.g. Bugtracker) to provide
> useful information (in this case : bugs) which are not explicitly
> mentioned or interlinked.

You probably refer to :
http://www.abdn.ac.uk/~r01srt7/swese2008/pdf/swese2008_submission_15.pdf ... will tryand read it ASAP.

> We have developed a prototype version for JIRA bug tracker which
> enhance JIRA bug tracker by providing small fancy widgets which
> collects related information about a bug (e.g, mentioning of a bug-id
> in a comment or in summary or in description of other bug) and as
> well as look for dependent bugs which are explicitly mentioned accross
> bug tracker and provide all these information as an integrated view in
> a bug tracker against a bug. please find an attached snapshot that
> will give an overview about it.
>

The plugin is that one :
https://dev.deri.ie/confluence/display/romulus/JIRA+Widgets I presume...
also will try and respond later when I've had time to read that.

>
>
> I would highly appreciate for furhter discussion and possible
> extension or usage of such interfaces or applications on big datasets.
>

Will respond back to you soon with more detailed response.

Best regards,

[Chaudhry Muhammad Aftab Iqbal]

hi,

On Fri, Sep 18, 2009 at 6:37 PM, Olivier Berger <olivier...@it-sudparis.eu> wrote:

Le vendredi 18 septembre 2009 à 12:30 +0100, Chaudhry Muhammad Aftab
Iqbal a écrit :

>
> We submitted a paper last year in SWESE 2008(Semantic Web Enabled
> Software Engineering) workshop where we have proposed the idea of
> customizing / enhancing web applications (e.g. Bugtracker) to provide
> useful information (in this case : bugs) which are not explicitly
> mentioned or interlinked.

You probably refer to :
http://www.abdn.ac.uk/~r01srt7/swese2008/pdf/swese2008_submission_15.pdf ... will tryand read it ASAP.

very well spotted :). Yes, this is the paper but unfortunately, it accepted as a Poster paper rather than Short paper and i dont think, you will find much details in a 2 page paper. please find the attached eight page paper which have some more details about the work mentioned,
 

> We have developed a prototype version for JIRA bug tracker which
> enhance JIRA bug tracker by providing small fancy widgets which
> collects related information about a bug (e.g, mentioning of a bug-id
> in  a comment or in summary or in description of other bug) and as
> well as look for dependent bugs which are explicitly mentioned accross
> bug tracker and provide all these information as an integrated view in
> a bug tracker against a bug. please find an attached snapshot that
> will give an overview about it.
>

The plugin is that one :
https://dev.deri.ie/confluence/display/romulus/JIRA+Widgets I presume...
also will try and respond later when I've had time to read that.

yeah, thats the plugin wiki page where i just summarise the plug-in and put the screenshot of the plugin for bug tracker. I might think it to be a good idea to have such light weight plugins accross different bug trackers which query, gathers and show related information accross bug trackers.
 

>
>
> I would highly appreciate for furhter discussion and possible
> extension or usage of such interfaces or applications on big datasets.
>

Will respond back to you soon with more detailed response.

looking forward for your feedback.

regards
Aftab Iqbal

[Olivier Berger]

Hi.

I enjoyed reading your 2 papers : "LD2SD: Linked Data Driven Software
Development" (http://sw-app.org/pub/seke09-ld2sd.pdf) and the extensive
version of "Sindice Widgets: Lightweight embedding of Semantic Web
capabilities into existing user
applications" (http://groups.google.com/group/baetle/attach/f5607ec43fd643b1/SindiceWidgets.pdf?part=4).

I think the first one is very close to something that we've been
describing in our short paper presented at WOPDASD 2009 : "Weaving a
Semantic Web across OSS repositories: a spotlight on bts­link, UDD,
SWIM" (http://libresoft.es/activities/wopdasd2009-olivier-berger.pdf)

I'm glad such uncoordinated efforts reach similar ideas.

I think we might try and collaborate in reaching a wider implementation
of the ideas we've both emitted.

For instance, we're working on trying to extract bugs facts as RDF from
Debian's UDD and Mandriva's bugzilla, which could enrich the facts
available to LD2SD (we're starting with the bugs, but lots other facts
about packages, etc are available).

On bug links, that you mention in one of the papers, between different
bugs in different bugtrackers, that's the very goal of our current
efforts on the ontology, for instance.

As for the use of Sindice, that's something we hadn't thought, but which
may fill a gap in trying to implement a bug facts retriever that we want
to develop.

Also, for the enhanced Web UI of the bugtrackers, I have some ideas,
among which the use of an "overlay" annotations browser in the form of a
Greasemonkey script that might help prototyping applications even if the
current bugtrackers don't change, by adding javascript "on-the-fly" in
the browser's rendering of the UI.

There are probably many other interesting areas to investigate, and the
best may be to try and setup some voice discussion to try and go
further.

In any case, we're actively working on such matters in the frame of our
Helios project, and will for the next 11 months at least, trying to
reach some live service usable by people (and also contributing to LOD
as much as possible), and made as FLOSS only.

Looking forward to reading from you.

Best regards,

Le vendredi 18 septembre 2009 à 12:30 +0100, Chaudhry Muhammad Aftab
Iqbal a écrit :

> Hi all,
>

>
> We submitted a paper last year in SWESE 2008(Semantic Web Enabled
> Software Engineering) workshop where we have proposed the idea of
> customizing / enhancing web applications (e.g. Bugtracker) to provide
> useful information (in this case : bugs) which are not explicitly
> mentioned or interlinked. We have developed a prototype version for
> JIRA bug tracker which enhance JIRA bug tracker by providing small
> fancy widgets which collects related information about a bug (e.g,
> mentioning of a bug-id in a comment or in summary or in description
> of other bug) and as well as look for dependent bugs which are
> explicitly mentioned accross bug tracker and provide all these
> information as an integrated view in a bug tracker against a bug.
> please find an attached snapshot that will give an overview about it.
>

SNIP

> I would highly appreciate for furhter discussion and possible
> extension or usage of such interfaces or applications on big datasets.
>
> regards
> Aftab Iqbal