Carlo Daffara is the Italian representative of the European Woking Group on Libre software, he worked in 7 EU research projects related to FLOSS, including one of the largest migration experiment for European Public Administration (COSPA).
I asked Carlo to join the conversation to tell us more about Open Source and Research.
How did you start your activities as ICT researcher within EU funded projects?
My first Commission activity was with the EU WG on Libre Software, where we [Barahona and I] prepared an article on the economic potential of FLOSS. It had quite an impact, and was also used as basis for many legislative actions and other research activities. I then started working in EU research projects related to OSS; the first one was SPIRIT, for open source in health care, where we prepared one of the first European sourceforge clones. Another influential one was COSPA, where we studied the real TCO/ROI of a migration to open source software on the desktop of European Public Administrations.
Few weeks ago I asked Alessandro Rubini his opinion about “the” community, and as you might know he is quite skeptical about.
What is your opinion about “the” community?
Alessandro is right in expressing disbelief in a generic “community”; there are organized communities that can be recognized as such (Debian or Gentoo supporters are among them) but tend to be an exception and not the rule. Most software do not have a real community outside of the developers (and eventually some users) of a single company; it takes a significant effort to create an external support pyramid (core contributors, marginal contributors, lead users) that adds value. If that happens, like in Linux, or the ObjectWeb consortium the external contributions can be of significant value; we observed even in very specialized projects a minimum of 20% of project value from external contributors.
It is worth to notice that both Carlo and I use the “pyramid” expression. While he is more focused on the contribution side, I used the expression to layer the market talking about Funambol business model with its CEO Fabrizio Capobianco. Funambol addresses users’ needs depending on their level up the value chain, where “free” customers are at the bottom of the pyramid and Carriers at the top – ISV, Wireless Manufacturer and Application/Internet Service Providers somewhere in the middle. Please note that now Funambol is available in two editions (used to be three), a sign that they are keeping moving and refining their business model.
You stated that at least 20% of project value come from external contributions, I guess this information come from early FLOSSMetrics results. What about the project?
FLOSSMETRICS is aiming at the creation of a set of tools and a comprehensive database of metrics related to open source projects, in a verified and stable way. It is planned that this will help future research on the software engineering aspects of OSS, and on the interaction of coding, non-code related activities, and social interactions.
Our area of work is related on the sustainability of open source-based business models, extending the work we have done in the past 5 years in SPIRIT, COSPA and other projects. We will leverage the database of code to find how communities can find sustainable development models, how business can cooperate or start an OSS project in a sustainable way.
Talking about business models, you are working from years on taxonomies and categorizations, would you tell us what is an Open Source firm in your opinion?
My opinion is that code licensing cannot be the only important parameter; for example, a company that pays developers, during work hours, to work on open source projects does indeed benefit OSS in general, and should probably be considered as such. In this view, Google can be considered an OSS company even if it does not release much code under OSS licenses.
It is important to consider that while it is understandable that claiming to be OSS when it is not true is negative for the market as a whole, the strong tension that is actually developing while debating OSSness is due to business reasons, as new entrants are trying to find a niche market (in some cases “faking” OSS capabilities) and OSS incumbents that are trying to magnify their “truer” OSSness for signaling reasons.
Here you are bringing some salts to the discussion, I will soon come back on this, may be asking you more on the subject.
Talking about public funded projects, me and not just me have been quite critical of some European funded initiatives. Do you see any problem in this respect?
The problem with some EU projects is that they are not really “open”, in terms of external suggestions, interactions, criticisms and such. Some projects “blended” very well with development communities, like EDOS, by having a dedicated person that interacts with groups, companies and communities in a structured way; I would expect that any “open source” project should have at least this kind of openness.
I totally agree. I think such approach should be used just everytime a significant IT public budget is allocated, it could be a very effective audit. I would also like to see EC evaluators be proficient with Open Source, and I know from my experience that applying is pointless if you are not in their know.
I am not convinced that there is a need for public funded OS projects, nevertheless I am in favour of public funding in presence of a clear market failure. Do you?
The problem is twofold: first of all, it is important to recognize the market failure, and this is not an easy thing to do; the second point is how to organize a project so that it is self-sustaining in the long term despite the market failure. This is in a sense an open research problem (that we hope to address during FLOSSMETRICS).
Thank you very much for sharing your thoughts Carlo, you are always welcome!
Carlo Daffara is head of research at a small italian OSS consulting company, he has participated in many other working group of IEEE, Internet Society and the EU ICT task force. He worked in 7 EU research projects related to open source and free software. Carlo has also been the Technical Director of the Italian Open Source Consortium for about three years, and he eventually succeeded to the Presidency when I left last year.
Technorati Tags: Open Source, FLOSSMETRICS, EU
Investigating the chicken coop.. by Meffi
Roberto Galoppini, 
gabriele, and 1 other are discussing. 
Ron German are discussing. 
Krissy are discussing. 
Michael 2:42 pm on March 16, 2007 Permalink
Do you think that OSS BI will have any chance?
Gartner wrote that Excel 2007 could be better than Actuate… If this is right, I also see no chance for OSS Bi solutions…
Do you have more information to this topic?
Greets
Michael
Roberto Galoppini 7:31 pm on March 17, 2007 Permalink
Hi Michael, I agree with Matt Asay saying open source BI players are transforming BI into something else. It might take time to get OS BI similar to proprietary BI, but I am not sure SMEs need them. Disruptive innovation never need perfect clones, after all.
gabriele 5:54 pm on March 20, 2007 Permalink
Do you think that OS BI is becoming similar to proprietary one? Probably some solutions do, following an OS product approach (dual licensing schema, pre-defined engines stack, …). I know something different: SpagoBI http://www.spagobi.org, following a different approach. It’s an integration platform focused on project results (e.g.: customer needs). The OS approach and business model is very important in BI domain. Innovation? Probably, in the near future. Effectiveness? Probably, now it’s the time.
Roberto Galoppini 6:53 pm on March 21, 2007 Permalink
Hi Gabriele, as I already wrote I totally agree with Matt Asay, OS BI players are targeting other customers, offering something else. As far as I understand SpagoBI, based on the Jasper report engine, is targeted to Public Administrations, am I right?
About double-licensing I believe it makes sense, sometimes, but I doubt it fits any needs. By the way BI requires a lot of consultancy, I don’t think you need any extra “lock-in” measure, right?
I would like to know more about your product and your business model, let me know if you are available for an interview any soon.
It might take time to get OS BI similar to proprietary BI, but I am not sure SMEs need them. Disruptive innovation never need perfect clones, after all.
gabriele 10:14 am on March 23, 2007 Permalink
Roberto, SpagoBI is not based on the Jasper report engine, it’s a platform integrating several different analytics tools and engines, with a non-exclusive choice: http://www.spagobi.org. It’s not targeting PAs, it’s targeting all market domains and needs; a lot of interest arises from PAs now. In my opinion the strength of OSS is not that it’s free (or cheaper, following the dual licensing schema) but that it’s more adaptive to the customers’ needs. Customers’ needs are satisfied by the development of software projects using OSS and so they need a lot of consultancy for all business domains, not only for BI. I hope that for all business domains (not only BI again) the competition is not OSS vs. proprietary; the very OSS is something different from a proprietary solution now and in the future. The challenge is for an effective solution. Sometime I talk about an OSS ecology for value, a value-hypernetwork that I think is suitable also for SMEs. No problem for an interview about business domain, if you like.
Roberto Galoppini 11:25 am on March 23, 2007 Permalink
Thank you Gabriele to keep conversating. I thought SpagoBI was based on jasper report engine because is reported in the dependencies’ list, as far as I understand you really ned only if you’re using some SpagoBI components, am I right?
I heard only about SpagoBI used within PA projects, other markets have different needs, please let me know how you cope with that.
I don’t think OSS is different “in nature”. I believe that many OS products are just licensed with OSI approved licenses, and that is proven to not make any real difference for customers. Adaptability is a nice to have, but it is definitely not granted by source code availability. Software engineering states that modifying software authored by others costs from 2 to 6 times, and sometimes even 20 times!
In this respect OS is a promise, not always fulfilled though.
Consultancy it is great for System Integrators and Consulting firms, again one size doesn’t fit all. BI is a good application area in this respect indeed.
General statements are always too vague, let’s talk about your approach!
gabriele 4:43 pm on March 23, 2007 Permalink
In short,
OS (not OSS) is different in “nature” if you think that OS is not only software “OSI approved”, but a different approach (process development, contribution, trust, network ..and more else): not general statements, facts. I’m working inside a Consortium (ObjectWeb) collaborating with other projects, individuals, integrators, SMEs. We collaborate also with projects and actors outside OW Consortium as well. Adaptability is non “granted” by source code availability, but code availability foster adaptability.
Sorry, words are words and using words we build nothing; software realizations are facts and we daily verify what OS developments makes the difference (non something like: install, configure, go live; A lot of failure with this approach). Obviously, not every time, not everywhere, not in all application domains. It’s just my experience and I hope it’s not the unique experience in OS domain. I’ve outlined my approach in Sardinia Convention PAAL 2007 (you can find a my paper in the website – sorry, it’s in Italian and tailored just for PAs).
Roberto Galoppini 6:02 pm on March 23, 2007 Permalink
Gabriele I totally follow your line of thought, let’s talk about facts. Two questions and few comments below.
Could you approximately tell me if your outside contributions are significant and at which extent?
Could you put me in touch with a SME contributing, in order to ask few questions about the OSS ecology of value you mentioned above?
Talking about the install-configure-go live (or nightly-build) method I would say that it worked quite well with few OS projects, I am definitely more sympathetic than critical of. You know why? Because when your product it is ready for prime time a vibrant community is already there. If in doubt ask Alfresco if it is or not an issue.
What makes a real difference is the modular architecture of software which strongly affects capability coordination. As with code availability, modularity is also just a precondition. What is needed to foster “lock-in free” services is a strong commitment for community-based peer production. And the community process has no much to do with the software, it is about sharing business opportunities in the long run (see MySQL and Red Hat approaches).
gabriele 12:59 pm on March 26, 2007 Permalink
Just some examples: contributions to eXo Platform, ServiceMix, Jpivot, Cimero and more… because it make no sense don’t give back a realization of general interest. But also projects integration (e.g.: SpagoBI and eXo, Talend and more ..): it means sharing projects road-map to achieve a bigger solution of general interest.
Roberto Galoppini 1:19 pm on March 26, 2007 Permalink
Gabriele I was asking about outside contributions to your project, though it is interesting to know you are contributing to others’ projects.
No SME is worth to interview to talk about the OSS ecology of value yet?
Keep in touch.
Michael 10:56 pm on June 5, 2007 Permalink
Thank you both for your contributions in this post. I have learned quite bit more since following along here.