As promised in my post, "Pentaho Reporting 3.5 for Java Developers First Look", I've taken the time to thoroughly grok Pentaho Reporting 3.5 for Java Developers by Will Gorman [direct link to Packt Publishing][Buy the book from Amazon]. I've read the book, cover-to-cover, and gone through the [non-Java] exercises. As I said in my first look at this book, it contains nuggets of wisdom and practicalities drawn from deep insider knowledge. This book does best serve its target audience, Java developers with a need to incorporate reporting into their applications. But it is also useful for report developers who wish to know more about Pentaho, and Pentaho users who wish to make their use of Pentaho easier and the resulting reporting experience richer.

The first three chapters provide a very good introduction to Pentaho Reporting and its relationship to the Pentaho BI Suite and the company Pentaho, historical, technical and practical. These three chapters are also the ones that have clearly marked sections for Java specific information and exercises. By the end of Chapter Three, you'll have installed Pentaho Report Designer, and built several rich reports. If you're a Java developer, you'll have had the opportunity to incorporate these reports into both Tomcat J2EE or Swing web applications. You'll have been introduced to the rich reporting capabilities of Pentaho, accessing data sources, the underlying Java libraries, and the various output options that include PDF, Excel, CSV, RTF, XML and plain text.

Chapters 4 through 8 is all about the WYSIWYG Pentaho Report Designer, the pixel-level control that it gives you over the layout of your reports, and the many wonderful capabilities provided by Pentaho Reporting from a wide range of chart types to embedding numeric and text functions, to cross-tabs and sub-reports. Other than Chapter 5, these chapters are as useful for a business user creating their own reports, as it is for a report developer. Chapter 5 is a very deep dive, very technical look at incorporating various data sources. The two areas that really stand out are the charts (Chapter 6) and functions (Chapter 7).

There are a baker's dozen types of charts covered, with an example for each type. Some of the more exotic are Waterfall, Bar-Line, Radar and Extended XY Series charts.

There are hundreds of parameters, functions and expressions that can be used in Pentaho Reports, and Will covers them all. The formula capability of Pentaho Reporting follows the OpenFormula standard, similar to the support for formulæ in Microsoft Excel, and the same as that followed by OpenOffice.org. One can provide computed text or numeric values within Pentaho reports to a fairly complex extent. Chapter 7 provides a great introduction to using this feature.

Chapters 9 through 11 are very much for the software developer, covering the development of Interactive Reports in Swing and HTML, the use of Pentaho's APIs and extension of Pentaho Reporting capabilities. It's all interesting stuff, that really explains the technology of Pentaho Reporting, but there's little here that is of use to the business user or non-Java report developer.

The first part of Chapter 12, on the other hand, is of little use to the Java developer, as it shows how to take reports created in Pentaho Report Designer and publish them through the Pentaho BI-Server, including formats suitable to mobile devices, such as the iPhone. The latter part of Chapter 12 goes into the use of metadata, and is useful both for the report developer and the Java developer.

So, as I said in my first look, the majority of the book is useful even if you're not a Java developer who needs to incorporate sophisticated reports into your application. That being said, Will Gorman does an excellent job in explaining Pentaho Reporting, and making it very useful for business users, report designers, report developers and, his target audience, Java developers. I heartily recommend that you buy this book. [Amazon link]

Except where otherwise noted, this content is
licensed under a Creative Commons License.

I was approached by Richard Dias of Packt Publishing to review "Pentaho Reporting 3.5 for Java Developers" written by Will Gorman. (Link is to Amazon.com)

LinkedIn
Richard Dias has indicated you are a Friend:

Hi Joseph,

My name is Richard Dias and I work for Packt Publishing which specializes in publishing focused IT related books.

I was wondering if you would be interesteed in reviewing the book "Pentaho Reporting for Java Developers" written by Will Gorman.

- Richard Dias

After some back and forth, I decided to accept the book in exchange for my review.

Hi Joseph,

Thanks for the reply and interest in reviewing the book. I have just placed an order for a copy of the book and it should arrive at your place within 10 days. Please do let me know when you receive it.

I have also created a unique link for you. It is http://www.packtpub.com/pentaho-reporting-3-5-for-java-developers?utm_source=press.teleinteractive.net&utm_medium=bookrev&utm_content=blog&utm_campaign=mdb_001537. Please feel free to use this link in your book review.

In the meanwhile, if you could mention about the book on your blog and tweet about the book, it would be highly appreciated. Please do let me know if it is fine with you.

I’m also sending you the link of an extracted chapter from the book (Chapter 6 Including Charts and Graphics in Reports). It would be great if you could put up the link on your blog. This would act as first hand information for your readers and they will also be able to download the file.

Any queries or suggestions are always welcome.

I look forward to your reply.

Best Regards,

Richard

Richard Dias
Marketing Research Executive | Packt Publishing | www.PacktPub.com

Shortly thereafter, I received notification that the book had shipped. It arrived within two weeks.

Of course, I've been too busy to do more than skim through the book. Anyone who follows me as JAdP on Twitter knows that in the past few weeks, I've been:

• helping customers with algorithm development and implementing Pentaho on LucidDB,
• working with Nicholas Goodman with his planning for commercial support of LucidDB through Dynamo Business Intelligence, and roadmaps for DynamoDB packages built on LucidDB's plugin architecture, and
• migrating our RHEL host at ServerBeach from our old machine to a new one, while dealing with issues brought about by ServerBeach migrating to Peer1's tools.

None of which has left any time for a thorough review of "Pentaho Reporting for Java Developers".

I hope to have a full review up shortly after the holidays, which for me runs from Solstice to Epiphany, and maybe into the following weekend.

First, a little background. Will Gorman, the author, works for Pentaho, in software engineering, as a team lead, and works primarily on Pentaho Reporting products, a combination of server-side (Pentaho BI-Server), Desktop (MacOSX, Linux and Windows platforms) and Web-based software (Reporting Engine, Report Designer, Report Design Wizard and Pentaho Ad Hoc Reporting), which stems from the open source JFreeReport and JFreeChart. While I don't know Will personally, I do know quite a few individuals at Pentaho, and in the Pentaho community. I very much endorse their philosophy towards open source, and the way they've treated the open source projects and communities that they've integrated into their Pentaho Business Intelligence Suite. I do follow Will on Twitter, and on the IRC Freednode Channel, ##pentaho.

I myself am not a Java Developer, so at first I was not attracted to a book with a title that seemed geared to Pentaho Developers. Having skimmed through the book, I think that the title was poorly chosen. (Sorry Richard). I find that I can read through the book without stumbling, and that there is plenty of good intelligence that will help me better server and instruct my customers through the use of Pentaho Report Designer.

My initial impressions are good. The content seems full of golden nuggets of "how-tos" and background information not commonly known among the Pentaho community. Will's knowledge of Pentaho Reporting and how it fits into the rest of the Pentaho tools, such as KETTLE (Pentaho Data Integration) and Mondrian (Pentaho Analysis), along with a clear writing style makes all aspects of Pentaho more accessible to the BI practitioner, as well as those that wish to embed Pentaho Reporting into their own application.

This book is not just for Java developers, but for anyone who wishes to extend their abilities in BI, Reporting and Analysis, with Pentaho as an excellent example.

I'll be following up with the really exciting finds as I wend my way through Will's gold mine of knowledge, and, will do my best to fulfill my promise of a full review by mid-January.

You can also click through the Chapter 6 (a PDF) as mentioned in Richard's email.

Thank you, Richard. And most especially, thank you, Will.

Today, SQLStream announced version 2.0 of their Real Time BI solution. SQLStream comes from the fertile creativity of Julian Hyde, who is also the founder of the open source Mondrian OLAP engine. While SQLStream is not open source, it does stem from the open source Eigenbase community, leveraging the user-defined transforms that were originally developed for LucidDB to operate on traditional stored relational data, with SQL:2003-compliant syntax. SQLStream extends this to handle streaming relational data.

In addition to capturing standard, structured data while "on the wire", SQLStream also includes adapters for feeds, such as Atom and RSS, and for Twitter.

Methinks Julian and I need to schedule another lunch soon, so that I can learn more about how this unstructured data, especially from Twitter, can fit into real time analytics provided by SQLStream v2.0.

BTW, you can follow me on Twitter as @JAdP.

Except where otherwise noted, this content is
licensed under a Creative Commons License.

With the opening of the Apple iPhone AppStore on iTunes and with the iPhone2.0 software, I decided to take a look for iPhone BI apps.

The first, from Pentaho, is open source. "Pentaho's BI extension for iPhone works with Pentaho Open BI Suite 1.7. Download and configuration instructions, as well as a short, recorded video demonstration are available..." from Pentaho's iPhone page. Matt Casters has more on his blog, "pentaho and the iphone".

The second is not open source, but is free from iTunes, but "Requires the licensing of Oracle Business Intelligence Suite, Enterprise Edition Plus, and Oracle Business Intelligence Applications, Fusion Edition..." Wow.

Even though I have the latest iTunes7.7, it's showing that my iPhone1.1.4 is up to date. We'll have to start playing with these as soon as iPhone2.0 is generally available - later today maybe.

I recently joined Twitter. I must share the following:

Roebot: #e20 note to organizers: If your panelists do NOT know what SOAP and REST are they prolly shouldn't be on a mashup panel!!! WTF!! about 5 hours ago from twhirl
-- Aaron Roe Fulkerson on Twitter

To which I responded:

Joseph_di_P: @Roebot wiki(SOAP) is what you use in tub to get clean; wiki(REST) is what you do in tub when not using SOAP :-D Easy, yah! about 1 hour ago from Hahlo in reply to Roebot
-- my response on Twitter

I know, I know, all the important stuff happening in the Open Source BI related world this week, and this is what I blog about. Is it a sign of dementia when you crack yourself up?

Here's some of the more important stuff that's been happening:

• Executive movement at EnterpriseDB, as told by Andy Astor
• The Pentaho BI Platform goes GPLv2, and Nick Goodman talks about it
• SnapLogic v2 released with webinars, one was today and the next is on June 18
• The new JasperSoft CEO blogs a bit
• The Pentaho community meetup in Mainz, Germany is about to start
• MuleSource is demoing the new Mule components working together
• Your iPhone will no longer need to be jailbroken, after iPhone 2 firmware, to work with myFunambol and even better than Apple's mobileMe, myFunambol is on track to sync Notes, per the press release

There's much else to do, including some additions to our linkblog with open source for MDM and more open source communities. But, I'm tweeting.

Today I had the good fortune of speaking with Doug Moran, Founder and VP of Community, and James Dixon, Founder and Chief Geek/CTO of Pentaho about their OpenScrum methodology. Doug had responded to my question on LinkedIN, "What has been your experience with Master Data Management and SOA, SaaS and Agile, in any combination?"

In creating the OpenScrum Agile software development method, Pentaho faced many of the same challenges that we've been facing with some of our customers:

1. Extending Agile methods to be used by distributed workgroups, in Pentaho's case this includes every inhabited time zone on this planet for their extended community and several time zones in the USA and EC for the core group
2. Adjusting Agile methods to work with several products or modules or projects in parallel
3. Documenting the communication among team members even when "The most efficient and effective method of conveying information to and within a development team is face-to-face conversation" isn't possible - quote taken from the Agile Manifesto Principles
4. Dealing with the concepts of timeboxes and rhythm
5. Deciding on who best fills the roles of product owner, scrum master, and the various levels of commitment among the core group, the extended team and the community at large
6. How best to involve QA/Test - a subject on which Scrum gives little guidance

The OpenScrum methodology deals with much of this, as does our ever evolving N Dimensions of a Project methodology, which has gone from 5D™ in the mid- to late-90's to the 8D™ iteration that we're currently documenting.

The most interesting point to come from our discussion is that Doug and James have come to many of the same conclusions and principles as have Clarise and I, even though our "experimental sample" is the Pentaho open source communities on the one hand, and the various internal IT and SaaS software development distributed workgroups on the other.

One of the first things we discussed concerned the idea of fixed timeboxes and rhythm. The idea of rhythm is very attractive, but very difficult to achieve in most situations. In many ways, the concepts that Agile methods address: responsiveness to changing user needs and a changing market, coupled with the realities of changing personnel and other organizational and business changes, make achieving a short-term rhythm very difficult. As an organization matures, perhaps a longer term rhythm, such as the quarterly or "seasonal" rhythms of salesforce.com might be a realistic goal. As an organization first moves into an Agile process, and develops an Agile software development methodology that fits with their evolving culture and ecosystem, sprint timeboxes must be flexible, and achieving a fixed release rhythm isn't practical. For our part, we make setting the timebox part of each sprint planning meeting. Once set, the timebox is inviolable, and any further negotiations must be around the feature set to be achieved in the sprint.

In the same area of time, James and Doug shared their observation that the sprint burn-down chart can result in a false sense of security or panic. No work effort follows the nice smooth regression line of a sprint burn-down chart, but more sophisticated modeling, to plan for the stepped nature of real-life work, is beyond simple spreadsheets. It's not beyond the current math modeling state of the art nor the capabilities inherent to the various analytic and data mining modules of Pentaho. While Pentaho currently has no plans for a "Pentaho for OpenScrum" similar to "Pentaho for Jira", it's certainly food for thought for the Pentaho community, and led to some fun brainstorming on our call.

An issue related to time, and space in this instance, is that almost everyone today works in some form of distributed workgroup. This may include the occasional telecommuter, or, as in the case of Pentaho and other open source projects, communities that span the globe. This would seem to preclude most Agile methods, especially Scrum, which puts a great deal of importance on stand-up face-to-face meetings, specifically the daily scrum. One way to get around this is to use instant messaging or teleconferences, especially if remote whiteboarding can be used - though always one group or another is inconvenienced by the selected time. Another way is to use blogs, wikis or forums to supplement the daily scrum, especially for those who might not be able to attend due to the time selected. One other things that we've done is to record the daily scrum, either audio alone, or with video. The take-away here is that some form of asynchronous communication, and keeping an historical reference, is a necessary addition in the face of the flat world of software development today.

The idea of parallel scrums has been batted around within the Agile communities for some time now. The reality is that sprints must often be done in parallel and some individuals are going to be assigned roles in more than one sprint team simultaneously. The reasons for this are many, but the bottom line was discovered by project managers long ago: small teams with sharply focused goals are more often successful than not. Large, complex or broadly defined goals must be managed as programs or portfolios, not as projects. We've followed separate strategic, tactical and implementation tracks for over a decade, with a modified iterative waterfall approach cycling through these three tracks throughout time. This allows us to easily incorporate parallel implementation sprints for each tactical project. The OpenScrum methodology has some good depictions of their approach to parallel sprints. For another perspective, I can draw upon a buffet-line-conversation I had with a senior project manager [who felt that Agile was "crap"] at a recent SFBAC (San Francisco Bay Area Chapter) meeting of the PMI (Project Management Institute) and that is the fact that Agile Software Development methods are not generally applicable Project Management methods. [My opinion and not necessarily representative of anyone else mentioned in this post.] As such, it's much easier to see how parallel sprints can be planned. With Agile methods, we're developing software, not dams, bridges or space shuttles. Software development is much more of an art than an engineering discipline, and the Agile methods allow for this artistry.

The roles of Product Owner, Scrum Master and Scrum team are very important to Scrum and OpenScrum and 8D™ and still requires refinement. Many folk that we've encountered in our consulting, naturally assume that the role of Product Owner is best filled by an existing Product Manager and that the Scrum Master is a lead engineer. But just as your best technical person is often a poor choice for any management role [soapbox time: forcing great techies into management roles as the only path for advancement and "the big bucks" is just plain stupid today] the lead engineer may not be the best Scrum Master. We feel that the Product Owner should be a true representative from the user community, or at the very least a customer advocate. A Scrum Master must be skilled in corporate communication, escalation and, yes, politics, and empowered to clear obstacles that could prevent the sprint from achieving its goal. In addition to these two roles, the Scrum Team should have no more than seven (7) other [full time equivalent] individuals with technical and business skills, and subject matter expertise that assure the sprint's success. This is a guideline, with lots of ways to be implemented.

Another area of discussion that deserves weeks of attention, not the few minutes that we could devote to it in our call, is how best to incorporate QA and all the various testing efforts that any engineering effort requires. Well-thought-out test scenarios and automation certainly play an important part, but that's not the end of it. And we didn't get anywhere near the end in our discussion. Most Scrum articles, books, etc. kind-of ignore QA. Other Agile methods, such as Test Driven Development are centered around it.

We believe in incorporating QA and Test as part of the core group, and that code review & unit testing, at the least, must be a part of the automated daily build process. But what about regression and User Acceptance Testing (UAT)? If the idea is that each sprint results in usable code, isn't regression testing a major part of getting to usable code? Isn't UAT the proof of the pudding? But are such activities a part of the Sprint, a separate, overlapping/parallel Sprint, or a separate activity altogether? And what about QA of specifications, documentation, data integrity and metadata?

There are also some issues specific to open source companies/projects related to User Acceptance Testing, and user satisfaction metrics in general. It's one thing to count downloads, but how does one measure the comments, or really, the lack of comments in the forums? Open source companies often don't know who has downloaded their product, and don't whether the lack of comments on a new release is indicative of satisfaction or disinterest or even disgust. One way in which an open source company can get some amount of automated feedback is to have the product report an heartbeat, or other "phone home" technique [see Positive Feedback Enablers in the OpenScrum methodology]. Pentaho has implemented an "opt-in" or "opt-out" system for their heartbeat and has had no negative feedback for implementing this system.

As you can see… many questions. The answers are often dependent upon the supporting infrastructure, maturity and culture of the organization or community.

The culture results from the personalities of the pigs [the truly committed], the chickens [the involved], the sheep [who prefer to be herded] and the goats [who prefer to be led]; and don't forget the penguins, cats and lone wolves either. Whether in an open source community or a corporate team, a manager may know the core group, but might only guess at the extended team or overall community. And, as with any generalization, the devil is in the details. This is why our methodology has gone from 5D™ to 8D™ over the years, and why James has developed both the Beekeeper and OpenScrum methodologies. From the initial framework for success, one must apply the principles of Agile even to the development of your Agile methods. Again and again, we see the need to adapt to specific situation, cultures and objectives.

We talked for well over an hour. One area that we didn't have time to explore was the Sprint Retrospective vs. Sprint Review. The OpenScrum methodology only discusses the Sprint Retrospective [among the truly committed] whereas we also allow for the Sprint Review [among all interested parties] in our "lessons learned" processes. Perhaps we can go into this further on another call or in the comments.

As always, talking with the Pentaho folk is lots of fun. We're always encouraged by their inventiveness in their execution as an open source company.

When I met with John Sichi back in August he let a cat out of the bag. When I was speaking to Nicholas Goodman in late September, he confirmed the cat, and now he's announced it.

Bayon Technologies is back, and Nick has gone from Pentaho employee to Pentaho partner. We at IASC welcome Nick back to the land of consulting and look forward to working with him in the future.

From a press release:

ACTUATE LAUNCHES BIRT EXCHANGE

Developer Community Site Serves as a Dedicated Resource for Eclipse BIRT and the Actuate BIRT Product Line

Actuate today announced the launch of BIRT Exchange (www.birt-exchange.com), a dedicated resource for the Eclipse BIRT developer community and the Actuate BIRT product line. The growth of the Eclipse BIRT developer community and BIRT’s increased use in enterprise deployments has created demand for supported products and services based on Eclipse BIRT.

BIRT Exchange provides a dedicated site where developers can learn about support and services for Eclipse BIRT, download Actuate BIRT products and participate in the growing Eclipse BIRT ecosystem. Developers can share code and expertise with peers, helping to accelerate the flow of knowledge and information within the community.

Available on BIRT Exchange:

· Demos — a collection of self-running learning tools on everything from AJAX interactivity for BIRT reports, ad-hoc reporting with BIRT templates, spreadsheet reporting, using sub-reports, parameters and parallel reporting

· DevX — a dynamic content hub where BIRT developers can quickly find, share and discover useful BIRT-related technical information including code snippets, report designs, technical tips and tutorials. DevX makes it fast and easy for developers to search, upload, download and rate content items

· Documentation — official documentation for Eclipse BIRT and Actuate BIRT products

· Searchable forums — on Eclipse BIRT and the Actuate BIRT product line, including a mirror of the Eclipse BIRT newsgroup, which is also searchable

· Software downloads and detailed product information — for Eclipse BIRT and the Actuate BIRT product line, enabling rapid report deployment, spreadsheet reporting and ad-hoc end user reporting

· Support Resources — information on support, services and training available for Eclipse BIRT and the Actuate BIRT product line

· Wiki & Blog —facilitates an active dialogue among the BIRT developer community

ATTRIBUTABLE QUOTES:

· Virgil Dodson, Java Reporting Evangelist at Actuate Corporation –“Eclipse BIRT and the Actuate BIRT product line are a powerful combination of tools that let Java developers spend their time doing what they do best: creating innovative and impactful applications. BIRT Exchange gives developers a place to share and learn from peers about the latest and greatest technology and techniques for adding reporting to their applications, so that they can more easily meet their end-users’ reporting needs.”

· Mike Milinkovich, Executive Director of the Eclipse Foundation – “BIRT has been one of the most popular and successful projects at Eclipse since its initiation by Actuate in 2004. BIRT has successfully grown a large community of developers around it, and has helped extend the Eclipse ecosystem into enterprise reporting.”

· Mark Lynd, President, Firescope – “When we first used BIRT to embed reporting within the Firescope platform, we were under time pressure and needed to quickly come up to speed and to understand our different options. BIRT Exchange would have saved Firescope a lot of time and frustration and made life easier. This is going to be an invaluable resource for any BIRT developer.”

RESOURCES:

· Ask BIRT video

· BIRT 2.2 video with indexed menu

· Interview conducted by Michael Coté, RedMonk

· BIRT on Ohloh

· Actuate Del.icio.us page

· Summary of Actuate’s 2007 Open Source Software Survey

· BIRT Exchange Screenshots

· Eclipse BIRT web site

· BIRT Exchange blog

Clarise and I met with Gabriele Ruffatti and Grazia Cazzin, from Engineering Ingegneria Informatica S.p.A. to discuss, and as Clarise wrote, get demos of SpagoBI [a full BI suite], and see the new Spagic EAI tool that was announced at OSBC2007, and the Spago framework [think Spring, only different].

Once he returned to Padova, Italia, Gabriele sent out some thoughts that came to him after participating in the open source business conference. He kindly gave me permission to publish them here.

"I attended Eben Moglen’s speach at the last Open Source Business Conference in San Francisco (May, 22nd). Hearing so many times: “stand up for freedom”, I thought: "I’m here now to promote free (in Europe we say libre as well) open source projects in different domains (Spago, java framework; SpagoBI, Business Intelligence; Spagic, SOA environment; Spago4Q, a Business Intelligence domain specific solutions) at an event mainly presenting commercial open source". While there, I’ve promoted new enhancements of SpagoBI, the Business Intelligence Free Platform, comparing it to other commercial open source choices such as Pentaho and Jasper and I’ve thought again: "what is the key differentiator? Stand up for freedom…"

"A participant said to me: “Guy: you’ve made a mistake. SpagoBI is not the Business Intelligence Free Platform; it’s the Free Business Intelligence Platform”. Was he right? I think no. This is the key: SpagoBI is a free platform, commercially supported, offering a new choice despite many other commercial open source products claiming more effectiveness to the market.

"What I mean by "commercial open source product" is: a solution claiming to be open source, claiming to have a community supporting it, but offering closed add-ons for enterprise adoption with a proprietary approach to the market (i.e.: acquisition of projects and IPs to strengthen its stack, dual licensing approach, aggressive marketing). Is it effective for the BI domain? Yes, probably it is.

"What I mean by a "free platform" is: free design, free collaboration, free assembling, free adoption. Not just because the license (SpagoBI adopts the GNU LGPL license), but because the efforts are in the software development improvement direction rather than in marketing proposition, in collaborations with different projects and solutions instead of acquisitions, in integrations with many free, open source, and also closed solutions to achieve the most effective solution for the user, maintaining a totally free code base core, instead of closed specific add-ons. Is it effective for the BI domain? I really don’t know, but the market, or a very new market knows the answer.

"Stand up for freedom. Is it “against” a wide commercial adoption? Is it “against” the growth of a strong business ecosystem? I know that Europe, Asia and South America are looking for free/libre software; Public Administrations are looking for free/libre software; new domain specific applications can be built thanks to free/libre software (Spago4Q is a just a first sample of it). At OSBC in USA, I’ve heard not only of commercial open source packages, but also of “building the right solution with an assemble mindset, opposite to a buy mindset”. It’s just a new choice: with an open mind, you can build open applications offering freedom to invent, to share your own knowledge, to assemble the right solution for people’s needs."

-- Is freedom the right approach to the Business Intelligence domain? from Gabriele Ruffatti, Director - Architectures & Consulting, Engineering Ingegneria Informatica, S.p.A.

F/LOSS - Free/Libre… Open… "Free as in speech, not as in beer" is well known. But how open is open? This is a question that continues to be debated in terms of whether or not a company is truly open source, pureblood or mudblood. Andy Astor puts forth the rationale behind EnterpriseDB's strategy around openness [NB: EnterpriseDB is a company that is sometimes cited as a company using open source without being open source]. We tend to be fairly liberal in our acceptance of what is an open source company [including EnterpriseDB], while the market for enterprise open source defines itself and "commercial open source" has more denotation than connotation within that market, and I rather like the direction being taken in Commercial Open Source, cited above, by Carlo Daffara. Others may be more stringent in accepting a company as truly F/LOSS or not.

Most open source projects avoid locking their users into their product by following or providing open standards and open APIs, and/or using a plug-in, or framework, stack or platform architecture. Such an architecture allows the users and the projects community to more easily customize the product to suit their unique needs, and to more easily give-back to the project, without necessarily becoming an approved committer to the core code. We're currently following 47 open source projects that provide components that can be used to build solutions for data management and data analytics, data warehousing and business intelligence. We're following 5 open source BI suites, most of which use or can use the same core components. JasperReports for the reporting engine and Mondrian for the OLAP engine are common but not all-pervasive. All five suites provide a convenient stack of components that work together "out-of-the-box", but provide varying facility for including additional or replacement engines and components. We're still working out what is the real differentiator among all these tools, but Gabriele has a very good point: the more freedom, the more openness, the better.

Personally, I've been working in data analysis for nigh onto thirty years. The reason that we've been looking for open source solutions for decision support, data warehousing, data mining, BI, GIS, EAI, BPM, [insert latest hyped term here], etc, is that this area requires more customization and user involvement than any other enterprise application. The beginning of this millennium saw a trickle of open source projects addressing BI needs, and 2005 saw an explosion in the number of projects. Open source BI provides the flexibility that is always needed in satisfying the user needs for data analysis, visualization and sharing. The greater the freedom, the greater the openness, the greater the flexibility… the easier it is to satisfy the users. This is why data management, analytics and BI is going to be one the most successful markets for enterprise open source, bar none.

Unlike Joseph, I did not attend the OSBC San Francisco. I, however, attended the Networking Reception where I had the pleasure of meeting Grazia Cazzin and Gabriele Ruffatti of Engineering Ingegneria Informatica, S.p.A.

Grazia provided us with a demo of Spago BI. The features of SpagoBI include:

* Portal
* Report
* OLAP
* Query By Example (QbE)
* ETL
* Dashboard
* Document Management
* Metadata Management
* Versioning
* Administration
* Data Mining
* Collaboration
* Geographic referenced information analysis

SpagoBI is positioned as an integration platform that provides BI solutions to the enterprise. It has no professional version that one can purchase. All the functions and product releases provided are available as Open Source. It is as is released under the GNU LGPL license and is hosted by ObjectWeb Consortium.

It is an impressive OSS BI solution. You can try out the demo yourself and tell us what you think ...