Archived website

This online community was active in conjunction with the Digital Agenda Assembly 2012 and is now archived and available for institutional memory. You can now join the discussion at https://ec.europa.eu/digital-agenda/en/community

Open Data topics of EU-wide relevance

digisus's picture
Submitted by digisus on Fri, 2012-06-08 15:17

Having read the posts in this group, I'd like to suggest that the following topics should be discussed at the DAA.

1) Inclusion
Opening up data, government and others, is an important avenue towards more transparency of institutions and digital innovation in 21C. Yet, we [http://www.okfn.org] are convinced that an important prerequisite to make this vision happen, is free access to the tools needed to manage and analyse OD. In other words To maximize (re)use we need to include the maxium number of people and therfore, we need to minimize barriers to participate. Open Source software plays an essential role here.

2) One market ...
It is important to shape Europe's emerging OD market in such ways that the market really is a single market.

a) ... One license
From a legal point of view that means for OD (as for open ource, btw) that license proliferation is prevented as much as possible. Ideally, we are striving towards a single license that captures the "Open Data Principles" in a legally binding manner. We are interested to discuss how work already done, like OKF's Open Database License [http://opendatacommons.org], can help here and how the way forward can look like. Do we eed another "EU Data License" at all? And if yes - which I am not yet sure about -, can we mimick the approach the EU has done in software by making the EUPL a "GPL translated legally into all EU languages"?
As a related question, we need to clarify whether copyright licenses - even if open like the CC family of licenses - are adequate and legally applicable to data set; or whether they do not in fact create new risks around "copy fraud" which could transfrom into liabilities of data-providing PS bodies.

b) ... open standards
From a technical point of view, we should be agreeing (soon) on a minimal set of openly specified data formats to be used preferentially on all the data portals deployed. That is not to force portal providers into specific technical solutions, but to minimize technical frictions for a single European open data market and ecosystem.
We think that choosing data formats that are specified based on open standards [see Wikipedia] are of crucial importance here. The essential criteria are: specified by a non-closed group of data experts, not controlled by an individual market player, implementable at no cost and with no further (or hidden) IP risk by anyone.

I'd hope to get the chance to present/discuss these proposals in one form or other at the DAA and I am looking forward to comments online!

Group audience: 
Interesting!
7 users have voted.

Comments

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 15:38

I would agree with almost everything you said to if Open Apps, i. e. Open Source released as Open Data, are considered a part of the Open Data Universe. Then the needed license issue for Open Data should also address complex data and among them, Open Apps should be there. Currente EUPL is not the solution but, may be, and expanded and modified EUPL license could cover all kind of Open Data, including Open Apps of course. What do you think about it?
I would only add two topics to your list of three ones:
- A "standard" definition of the Metadata needed for publishing Open Apps as Open Data.

- How to manage the big amount of information/data that will appear as soon as Public bodies of Basque Country and even of the whole Spain (and may be other european countries) start releasing the source code of their applications (Open Apps)?

Interesting!
2 users have voted.

digisus's picture
Submitted by digisus on Fri, 2012-06-08 16:36

Olcoz, thanks for the comments. I read the FSFE article too (alas, I do not udnerstand Spanish). From what I understand teh Baseque govt decided to release all software as open source. Which is very sensible decision, I think!
That I do not understand is "open source released as open data" in your comment. I understand software to be program code. So, software is different from data. Software is covered by copyright and you can use open licenses to make it freely available. Data is not covered by copyright law, in fact data as such is not covered by any IP law. The only reason, we need licenses like the ODbL is database law (sth. else we should change, actually).
So, if you want to publish all Basque software as open source and call it "open apps", why not. The open data discussion hoewever, is different from the open software dicsussion, despite many similarities in philosophy.

+1 on the metadata standard.

Why do think it will be a big quantitive problem to release all this software? There are code repository tools available as FOSS today. I think the problem is rather organizational than technical: Should all Spain have one software repository? Or every region or even every city? Another challenge is also organisational (or rather managerial): Just publishing a few TBs of code alone will not help anyone. You need documentation, introductions for new developers, etc. In short, you need a plan for community building. And that is going to be very challenging for the public sector, as I experienced wokring for the City if Munich, because the PS org culture and community culture are very different. :-)

Interesting!
2 users have voted.

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 17:30

Digisus, thank you very much for agreeing on discussing about "A "standard" definition of the Metadata needed for publishing Open Apps as Open Data." during next DAA. In fact, we should try to identify/define a "standard" definition of the Metadata needed for releasing any kind of Open Data, including Open Apps as part of them as I commented in my previous message. Isn't it?

Interesting!
0 users have voted.

mjaque's picture
Submitted by mjaque on Sat, 2012-06-09 07:46

Nice thread, great ideas, congratulations.

Really I can't find much differences between data and code. Both are bits of information with and associated license to allow them to be used.

Considering software apps, especially those owned by public admins, as data, will simplify openning processes. The same procedures and tools could be used for them.

The goal is also the same for both: let people get the the best out of public digital resources, be this figures, images, texts, code...

I really believe DAA should consider software apps as one more kind of data, together with pictures, geographical information, historical data, reports...

Interesting!
2 users have voted.

digisus's picture
Submitted by digisus on Sun, 2012-06-10 14:56

mjaque, as I said earlier, your concept of "open apps" is mixing things that should not be mixed because program code and data are not the same legally speaking: The first falls under copyright, the second does not.
Now, if we have an area in the whole digital world that for once is -not- already causing difficulties by being covered by some type of IP law, why should we ever change it?? Legally, data is more free and accessible BY DEFAULT than software. And we should keep it that way because that is what "open data" is all about.

Interesting!
2 users have voted.

Olcoz's picture
Submitted by Olcoz on Sun, 2012-06-10 17:31

Let's focus on the standardisation or harmonisation of Metadata needed to release Open Data, right?
After doing so, do you mind if we also propose to extend a little bit this standard Metadata definition in order to use this simple releasing mechanism of Open Data to release Open Apps too?
The proposal is to have a Metadata definition to release Open Apps as an extension of the Open Data's Metadata and not the other way around. Don't worry about it.
You could stay at first level (where it seems you are currently focus on) and leave other people go with us to sencond level (where Open Apps are a kind of Open Data), if you want to do so, isn't it?
So, we would like to participate in first level harmonisation/standardisation process of Open Data's Metadata (you should also look at http://daa.ec.europa.eu/content/open-data-not-always-public-data) and we invite you to do the same with us in second level, for Open Apps, if you would like to do so, of course. Finally, we would like to do so in DAA and in other forums too. ;)
Kind regards

Interesting!
0 users have voted.

mjaque's picture
Submitted by mjaque on Sun, 2012-06-10 20:09

Marcus, you can always finds reasons for and against any idea. The question is: what makes this society better. Because that's the goal of DA.

Are there legal differences? Yes.
Can they be overcome? Of course.

So the question is: do we, as society, need to consider open apps? And I'm sure the answer is yes.

Differences between data and code are really tiny. Code and data can be handled the same way and can be opened and shared with the same tools.

So let's focus on similarities instead of pointing differences, we need to open both.

Interesting!
1 user has voted.

Rose Lockwood's picture
Submitted by Rose Lockwood on Wed, 2012-06-13 11:48

In the "Big Data" (defined as large volumes of unstructured information) corner of the data world the distinction between "data" and "software" is even more blurred. "Tools" (software) and "resources" (data) are almost indistinguishable when the computing techniques used (e.g. for search, sentiment analysis, language translation, etc.) rely completely on large datasets (e.g. the Web!). Alas, this type of data is not unambiguously "more legally free and accessible" - ask anybody who is trying to compete with Google in the field of automatic translation. IP concerns - real or imagined - are a huge barrier to opening up the Web for large-scale multilinguality. The DAA Data Group could make a significant contribution by promoting legal clarification of the status of multilingual data on the Web. We need an Open Data principle, perhaps similar to "fair use", that defines segmented, extracted, text elements as data, distinct from "content" with its complex IP constraints. You could think of it as "Open Data Pull" versus the more visible "Open Data Push" in the Public Sector. In my opinion Europe will never achieve true inclusion, in a fully integrated single market, unless this kind of "language data" is open and accessible.

Interesting!
2 users have voted.

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 17:27

Digisus, the code of an application is also data, right? In thos case this code, and particularly its source code, is under copyright rights, that's the reason why we need to release it under EUPL license, in order to put in free and public access while guaranteeing the four freedoms of Open Source.
Until now, releasing Open Source used to be done through a forge, in which a community works and so on and so for. Our approach is different, we are separating the releasing process of Open Source from the reusing framework/forge in which to work with it and to do so, we need a different mechanism to release Open Source than current ones. We decided to consider Apps as a kind of complez Data with their corresponding complementary inofrmation/documentation plus the license/s under which we open them. Then we can release Open Apps using similar mechanisms currently used to release Open Data.
On the other hand, if we could consider Apps as a kind of Data (maybe a complex one) and if were a need of a common european license to release Open Data, I wonder why we could not try to extend/modify EUPL in order to be used not only with Open Apps, i. e. a subset or kind of Data, but for any kind of Data. In fact, such an extension should also fix the issue that for releasing Open Apps, besides the EUPL for the source code/data, we also need CC By-SA licenses for the complementary information/documentation, in order to provide an EUPL* covering all data and metada associated Open Apps. Does it clarify our approach and clear your points?

Interesting!
1 user has voted.

digisus's picture
Submitted by digisus on Fri, 2012-06-08 18:47

Well, now I understand what you want to do -- and for all I know about IP law, I think it is legally neither possible nor desirable. And even if it is possible, I would not call the whole thing "open apps" because apps are program code. :-)

Interesting!
3 users have voted.

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 19:22

Well, of course it is possible and legal, and now plan it will be by default. Open Apps is about oppening the source-code/data of an application and releasing it as it is done with any other data plus the corresponding licenses that maybe they are nor always needed for any kind of data. And I thnik there is no doubt at all about that "program code" (as you called) it is data. ;D

Interesting!
1 user has voted.

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 17:38

Digisus, I partially agree with you about the third point. Maybe it could be fixed improving organization of the code/data of the Open Apps, improving design methodology or even defining a design for reuse one as in microelectronics field (there it is quite common, I was working on it while doing my PhD in microelectronics by mid-90s, by the way) or even improving management practices as you also said to. However these approaches could also be used to improve reusing or exploiting any kind of data, isn't it? However, even though this is not enough to exploit the information contained in big amounts of data, and particularly if these data are complex data, isn't it? So, if we could consider Open Apps as a kind of Open Data (Open Complex Data if you prefer), what do you think then about my last suggestion for topics to be considered by Data Group of DAA?

Interesting!
1 user has voted.

digisus's picture
Submitted by digisus on Fri, 2012-06-08 18:58

To be honest, the concept of "open apps" does not make sense to me and I do not see which problem you are trying to solve by introducing such a concept. After all, the fact that we need two different licenses, one for SW and one for data, is not the crucial legal problem in my eyes.
The crucial legal problem is that we use a few licenses as possible and that these licenses make reuse of SW and of data, possible. As I explained in the OP, data IMHO is not affected by any license question (except in whole agreegates/database -> database law).

The processual/managerial challenge, however, to release large amounts of data (or code) for reuse by a large, unknown community population is huge massive. Hierarchical organisations like PS bodies do not have experience with unhierarchical groups of people - with networks. We need to avoid that massive amounts of data (and sw) are released -- and nobody does anything with them. THAT is the biggest danger here, in my eyes.

Interesting!
3 users have voted.

Olcoz's picture
Submitted by Olcoz on Fri, 2012-06-08 20:45

Well, we are just trying to simplify Open Source releasing process and we think we can apply similar mechanism currently used by Open Data. Maybe some extra Metadata has to be defined for releasing Open Apps as Open Data or better to say using similar mechanisms.
In fact, it should be great to standardise these Metadata and that's all I was proposing, in summary. Well not all because this approach also drives to consider Open Apps as a kind of Open Data, at least according this point of view. Then, it's clear that, at least for Open Apps, it is needed a license and if this could be generalised for any other Open Data (if needed) then same approach would solve both cases that currently are managed in a very different way. And I agree with you on having as less licenses as possible, isn't it? So, if possible, why not trying to have a general solution that could be used for Open Apps and for any other kind of Open Data (if needed). I mean the simplest solution we can find for it.
Reagarding the last point, I think it is probably too early to be sure about how to deal with a problem that we can not even properly imagine yet because we don't have it (yet). Maybe we'll cross that bridge when we come to it. So, I have no problem to leave this point out of the discussion. ;D

Interesting!
0 users have voted.

Policybloggers's picture
Submitted by Policybloggers on Mon, 2012-06-11 10:36

great thread. thanks @Olcoz for pointing this out. have shared with others as well.

@ digisus - Do you know if you are actually presenting this and where at DAA12?

So, the crux of @ digisus argument is: we need a single licence for Open source and Open data. yes? and as and @mjaque says - The goal is also the same for both: let people get the the best out of public digital resources, be this figures, images, texts, code...

@Olcoz says: We decided to consider Apps as a kind of complez Data with their corresponding complementary inofrmation/documentation plus the license/s under which we open them. Then we can release Open Apps using similar mechanisms currently used to release Open Data. – I like this ..

However, @digisus – why do you say “legally neither possible nor desirable”?
I think it’s a great idea(will explain why soon) but want to hear your views.

Again for @digisus:
When you say: “To be honest, the concept of "open apps" does not make sense to me and I do not see which problem you are trying to solve by introducing such a concept. After all, the fact that we need two different licenses, one for SW and one for data, is not the crucial legal problem in my eyes.”
- Ah but it is  I will elaborate more .. but keen to see why you don’t like this idea. This is one of the most interesting ideas I have seen here – hence curious ..

@digisus:
“The processual/managerial challenge, however, to release large amounts of data (or code) for reuse by a large, unknown community population is huge massive. Hierarchical organisations like PS bodies do not have experience with unhierarchical groups of people - with networks. We need to avoid that massive amounts of data (and sw) are released -- and nobody does anything with them. THAT is the biggest danger here, in my eyes.”

Agree .. in fact that’s why I think releasing apps and data together under a single licence would make it easier for people to use .. and hence it solves the problem which you rightly identify.

Now, my view is:
a) Open data and open source often have the biggest problem of apathy ie no one uses it or knows it
b) Open data is good – but often undertaken by govts for the wrong reasons. To save costs. This means – a developer is supposed to work(for free) on an app(when she gets paid a lot more to develop an app commercially). That’s the flaw in open data thinking – ie wrong motivation from govts to save costs
c) IMHO - The potential of @Olcoz is – the possibility of governments to reuse apps from OTHER governments .. think of it as an open source ERP for govenments created as a set of smaller modular apps created organically. This could be truly disruptive.
d) So, the Basque govt releases open source apps and then the say Berlin govt reuses it – you see what I am getting to? Ie the analogy of open source ERP for govenments created as a set of smaller modular apps created organically.

That’s why I was interested in the idea from @Olcoz when I first saw it ..

Interesting!
2 users have voted.

Gudrun Magnusdottir's picture
Submitted by Gudrun Magnusdottir on Wed, 2012-06-13 10:49

Number 2 - One Market - is not solved by any of the suggestions you mentioned nor even enhanced. Data is mostly written in multiple languages, that contains soft information that is difficult to disseminate. This is especially true when aiming at cross border commercial operability. Localization is possible for websites and interfaces but when it comes to making use of content we barely scratch the surface. Commercially this should be the Digital Agendas strategic goal since it happens to be EUs technical forte!

Interesting!
2 users have voted.

Rose Lockwood's picture
Submitted by Rose Lockwood on Wed, 2012-06-13 12:16

Gudrun is right (echoing my point to Marcus :)) that language is the single biggest barrier to One Market, at least in terms of Open Data. Language Technology turns "soft information" into data, and this should be high on the DAA Open Data agenda! This discussion of the blurry edges between data and code highlights the fact that OD/OS, per se, is not the only issue. An Open Source code/data application bundle created in the Basque Country will not be usable in Berlin unless we have the tools/data bundles to cross language barriers. And as I said earlier, we will not get those bundles of language intelligence unless we can free up existing - and future - multilingual language resources embedded in every corner of the information universe...call it Open Multilingual Data. The beauty of this prospect is that once multilingual data is liberated, its usability goes far beyond simple "translation" to enable virtually every emerging information-based application.

Interesting!
1 user has voted.

alorza's picture
Submitted by alorza on Tue, 2012-06-19 13:12

Ok, Marcus. I only would add a more "bullet":

1) Inclusion (strong stress!)
2) One market
a) one license
b) open standards
c) some datasets compulsory to release

Interesting!
0 users have voted.

digisus's picture
Submitted by digisus on Tue, 2012-06-19 16:16

Hi everyone,
just to let policybloggers who asked and everyone else on this thread know that I will not be in Brussels this week to present these ideas (and have not received an invitation anyway). Colleagues of mine from the Open Knowledge Foundation will attend. So, feel free to get in touch. Wishing all a fruitful discussion with good results! :-)

Interesting!
0 users have voted.

People

competencesmarocaines.org's picture
fhardes's picture
fredriklinden's picture
keneastwood's picture
Nicholas Bentley's picture
JacintaArcadia's picture
Loankanassy's picture
Kasper Peters's picture
Kristijan Jakic's picture
lpujol's picture
Digital Agenda Assembly engagement
glqxz9283 sfy39587stf02 mnesdcuix8
glqxz9283 sfy39587stf03 mnesdcuix8
glqxz9283 sfy39587stf04 mnesdcuix8