OUTERTHOUGHT.BLOG

Research grant for cloud-scale content management

2010-03-11T14:19:11.000+01:00

Outerthought receives Flemish government research grant for cloud-scale content management

The research activity of the Belgian software company Outerthought gets a serious boost from IWT - the Flemish agency for Innovation by Science and Technology, in collaboration with the Innovation Center East-Flanders and University of Ghent MMLAB/IBBT.

Since 6 years, Outerthought has been the producer of the successful open source CMS 'Daisy', which is in use with large and small organisations both locally and abroad for applications such as knowledge, website and document management. The success story of Daisy now gets a sequel with a research project on scalable and searchable storage using NOSQL technology, making Outerthought a innovation leader in this area.

NOSQL is an evolution in the area of data storage, based on the misfit of relational database technology for internet-scale systems. NOSQL sits at the foundation of the storage- and indexinfrastructure of Google, Facebook and Amazon. The new product 'Lily' will be the first content repository going solidly NOSQL. Through a smart integration with sophisticated index- and searchtechnology, Lily will offer a boundless content repository, which combines the interactivity and searchability of web applications with the capacity for growth even beyond enterprise storage and archival systems.

"Lily allows us to rise above the tar pit of Web CMS systems", says Steven Noels, managing partner at Outerthought, "and lets us focus on our strengths of top-technologists and builders of robust infrastructure software. The standards are high for Lily, but our experience is our advantage, and was also the basis of our positive grant evaluation."

Outerthought has considerable experience with open source software development, and this successful grant is also the result of teamwork. The original findings of the research will be validated by the University of Ghent MMLab/IBBT experts against the state of the art in (meta)data modelling. The Innovation Center of East-Flanders assisted with the administrative work of the proposal.

"This is an important step in the institutionalisation of our research efforts," explains Steven, "we want to create a permanent research cel in our organisation which allows us to better showcase our competencies to the outside world. The support of this government grant allows us as an SME to move more swiftly and decisively in this direction."

The first research results and Lily pilots are expected around Summer, and Outerthought is currently on the lookout for partners for commercialisation and exploitation. The open source model allows for maximum flexibility in this respect.

The Flemish Government wants to stimulate innovation in Flanders. Therefore, it grants IWT annually the budgets necessary to finance research and development by and for Flemish companies.

Contact: Steven Noels / +32 9 338 82 20 / Outerthought / Technologiepark 3 / 9052 Zwijnaarde / stevenn@outerthought.org

Videos from the NOSQL devroom, nr. 9: Stéphane Combaudon

2010-02-24T10:17:37.000+01:00

This is the (excellent!) talk on comparing CouchDB MapReduce with traditional SQL by Stéphane Combaudon. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 7: Rob Tweed

2010-02-22T11:07:04.000+01:00

This is M/DB and M/DB:X on GT.M by Rob Tweed. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 8: George James

2010-02-22T17:32:34.000+01:00

This is George James' talk on GT.M and OpenStreetMap. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 6: Benoît Chesnau

2010-02-17T11:31:57.000+01:00

This is the talk from Benoît Chesnau on CouchDB. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 5: Eric Evans

2010-02-16T14:12:55.000+01:00

This is the talk from Eric Evans from Rackspace on the Cassandra Distributed Database. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 4: Evert Arckens

2010-02-15T17:27:34.000+01:00

This is the talk from Evert Arckens from Outerthought on designing the next-generation version of our flagship CMS Lily (formerly: Daisy) based on NOSQL technologies. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 3: Kristina Chodorow

2010-02-15T13:43:03.000+01:00

This is the talk from Kristina Chodorow from 10gen on mongoDB. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 2: Lars George

2010-02-12T15:38:23.000+01:00

This is the talk from Lars George on his life with HBase. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

Videos from the NOSQL devroom, nr. 1: Tim Anglade

2010-02-12T11:09:24.000+01:00

The first encoding and editing result from the NOSQL devroom just came in: the introductory talk by Tim Anglade. Enjoy, and thanks to Hendrik for video captation and encoding, and to Parleys.com for allowing us to host our video material! (If you want to experience the full Parleys e-learning environment, click here)

NoSQL FOSDEM impressions

2010-02-08T13:56:46.000+01:00

Last weekend was the yearly pelgrimage to Fosdem in Brussels again for geeks (and others) from around the world. On Sunday morning there was a good crowd arriving, but it was hard to get an overview due to the layout of the conference. Hearsay learned us that Sunday was as busy as Saturday.

The NoSQL devroom we organized was a big success. Throughout the day we had a packed room (with way too less oxygen - our usual habitat?) and often we had to put up the "Full" sign and disappoint people that still wanted to enter.

There was a nice lineup of international speakers.

Tim Anglade had the honor to start the day with his presentation NoSQL for fun and profit, where he gave us an overview and some history of the different NoSQL technologies. Including the SQL-NoSQL-NOSQL debated, or is it a non-debate?

Then it was time for some technology introductory sessions. A first one on MongoDB by Kristina Chodorow giving us a clear and detailed view on what MongoDB is about. Lars George followed by explaining the ins and outs of HBase for which he could rely on years of experience with HBase in his 'My life with HBase' talk.

Eric Evans closed the morning sessions with an introduction on Cassandra. It was (quoting @yannski on Twitter) "a very good balanced presentation between high tech and real world".

The afternoon started with a presentation on CouchDB by Benoit Chesneau. CouchDB clearly being a popular topic as the audience kept pouring in for the start of this session.

Next up were two presentations of technologies that use the high-throughput key-value database GT.M as underlying database. First Rob Tweed with M/DB, an opensource emulation of SimpleDB which he wrote in just one week, and M/DBX an open source native XML database. George James followed with an explanation of how they are using GT.M with OpenStreetMap to extend its query capabilities.

A second session on CouchDB attracted again a good crowd. Stéphane Combaudon gave a very clear and informative comparison between MapReduce in CouchDB and SQL in RDBMS. Glad to hear that he (being a SQL DBA) grew from being a stressed CouchDB experimenter to a 'Relaxed' CouchDB user.

I myself had myself the pleasure of closing a successful and informative NoSQL day with a birds-eye view presentation on how we are designing a scalable and available content management system (Lily) on top NoSQL technologies like HBase.

According to the reactions we got from the audience and the speakers, and from what we read on Twitter, this has been a great and fruitful day for all.

Thanks to all the speakers and to the Fosdem staff for organizing this great conference!

Evert.

Call for Participation: NoSQL devroom at FOSDEM 2010

2009-12-01T21:22:57.000+01:00

February 6-7th of next year, the tenth edition of FOSDEM takes place in Brussels, Belgium. FOSDEM is a weekend-long open source/free software fest organized by the community, for the community, and gathers up to 4000 people from all around the world. Best of all, it's free due to a lot of volunteer work and a number of nice corporate (and volunteer) sponsors. The setting is the Université Libre de Bruxelles, and FOSDEM hosts a famous beer night as well. All things FOSDEM can be found at http://fosdem.org/2010/.

This year, we managed to reserve a NoSQL 'devroom', which essentially is a mini-conference-inside-a-conference, a topically-based gathering of people presenting, using and/or hacking on NoSQL data stores. The devrooms are a pretty scarce resource so I'm really happy we managed to get hold of one. The idea is that we can set up and organize a full day for ourselves - NoSQL folks - on the Sunday of the FOSDEM weekend (Feb 7th 2010 - 9AM-5PM).

I'm looking for people wanting to present their favourite NoSQL tool or war story, or to present/clarify academic background papers on NoSQL storage/retrieval/distribution theory, any contemporary NoSQL theme basically, be it CouchDB, Cassandra or HBase, riak, Redis, Voldemort or all the other ones I forget to mention now.

This will be the first EU-based larger-scale NoSQL gathering, and I'm really excited to act as your interim servant/event planner. We have a 70 person capacity room (without squeezing), network/wifi, at least 7 full hours of NoSQL talk time, a beamer and the really nice FOSDEM atmosphere: all great ingredients to make the best out of the day.

I'm looking for speakers, attendees, participants in any kind (video capturing?) and perhaps some sponsors as well. All information - updating as we speak - can be found at http://nosqldevroom.pbworks.com/ which will be mirrored to fosdem.org once we've got a finalized schedule.

I'm enthusiastically looking forward to your responses, ideas, wishes and comments.

NoSQL and HBase at Devoxx - a report

2009-11-25T15:10:07.000+01:00

It's been almost a week ago already that we indulged ourselves with a 48 hour period of full-on NoSQLness. About time to report on that, no?

It all started last Sunday meeting up with Lars George, self-proclaimed EU HBase ambassador to talk shop about HBase and our decision to use HBase as the underlying foundation of our next-generation content store. After sharing a couple of beers and lamenting the rather silent state of the NoSQL movement in Europe, we felt we could be part of the solution rather than the problem so there's a good chance we'll try and organize a NoSQL meetup somewhere in Spring of next year, hopefully being able to share some more in-depth war stories from our own experiences.

After beers and a good night of rest, we spent a good part of the day discussing HBase and Lars' own experiences with it, and I must confess being impressed with Lars' gut to go down the HBase route at a pretty premature stage of its infancy - good to hear it has mostly been living up to its promises so far. Here are some meeting notes from our meeting (application/pdf, 11.9 kB, info) - verbatim.

In the late afternoon I was expected in Antwerp for my Devoxx "Tools in Action" talk on NoSQL (with some focus on HBase). It went OK, though it's hard to fit any kind of coherent story into a 30' talk (and my time management is obviously pretty bad). I gave the audience (about 200 people) a very short overview of our reasoning behind moving from My- to NoSQL for our CMS platform, presented "the classics" (CAP and BASE) and then gave a very short HBase intro to finish. I got some questions afterwards as well, which in my mind translates as "people found it interesting enough to want more". Oh, and I only went 5' behind schedule! :-)

NoSQL - Tools in Action

View more presentations from Outerthought.

After the talk, I went down to the BOF rooms to be greeted by Lars and my Outerthinkers setting up the BOF room. Evert learned us about the fishbowl discussion technique, which was pretty interesting and a cool way to "warm up" a room full of techies to actually have an interactive group discussion between strangers. I would lie while understating the importance of Lars' presence during the BOF, as the planned generic NoSQL theme quickly converged around HBase-only chatter, however a lot of the common challenges, problems and design constraints are pretty similar among the different NoSQL solutions.

NoSQL BOF at Devoxx

View more presentations from Outerthought.

We didn't mind the BOF becoming HBase-centric, if only that it showed that next year, Devoxx should accomodate a full NoSQL track (or set of BOFs) given the current interest in the subject. With 50 people attending a late-evening BOF session, I'm sure there's interest for more!

Related, don't forget to check out the Xebia blog report on NoSQL/Devoxx as well!

Join us at Devoxx 2009!

2009-09-30T17:21:58.000+02:00

For several months now already, we have been exploring new grounds in the domain of data management, reading oodles of papers and blogs on the NoSQL movement. Cassandra, Google BigTable, Amazon Dynamo, and many, many others, I think we've collected a reading list worthy of a PhD thesis. Making sure we're making the right decision for the future of Daisy, and exploring new (and exciting) technology at the same time. Yeah, we're a engineering shop, so we tend to enjoy bare-bones unpolished code.

We're hosting a BOF meeting during Devoxx - the yearly European Java lovefest - to discuss what we've learned so far, and would be thrilled to learn about fellow NoSQL-folks in the neighborhood. So come and join us, will you? You get the chance to get to know some new Outerthinkers as well.

NoSQL with Cassandra and Hadoop: sharing experiences and insights, November 16th 19-20h in Metropolis, Antwerp.

Update We'll be talking about NoSQL/HBase during the "Tools in Action" track on Nov 16th in the afternoon from 17:25 to 17:55 now as well.

New books @ the office

2009-08-27T09:40:51.000+02:00

It seems as if we're exploring new grounds for Daisy 3.

Be our guest for a cup and a chat

2009-06-29T09:51:35.000+02:00

Are you in the neighborhood and fancying a cup of coffee? We've just upgraded our coffee brewing equipment and offer a nice Nespresso Lungo or Espresso to anyone joining us for a cup and a chat. With complimentary free Wifi!

Outerthought: Spiking ourselves with caffeine so we can work harder for you. :-)

IKS workshop results are out

2009-06-22T16:22:55.000+02:00

IKS - Interactive Knowledge Stack is an Integrating Project part-funded by the European Commission. It started in January 2009 and will provide an open source technology platform for semantically enhanced content management systems.The IKS project has adopted an open communication and development policy to guarantee innovative and relevant results. To support this, IKS will host workshops every six months to consult with CMS vendors, developers and end-users.

The first IKS workshop took place in Salzburg, Austria on 28-29 May 2009. It provided an open forum for a diverse group of CMS vendors and developers to have their say on semantic requirements for the Interactive Knowledge Stack. Outerthought delegates were Bruno Dumon and Steven Noels.

Despite the limited visibility of IKS, (project only launched in January 2009) interest in the workshop was strong across the board of CMS communities, with experts from as far as India and the USA expressing interest to attend. This justified not only the timely relevance of IKS given the current discussions on Semantic Web and Web 3.0, but also how receptive the CMS communities are to the topic semantic information technologies and services.

You'll find all the results of the workshop on the IKS website.

Interview with Steven Noels from Outerthought from IKS Project on Vimeo.

Browser makers, where art thou?

2009-06-19T15:21:08.000+02:00

These days, we're dreaming about Daisy 3. We've just filed an innovation grant proposal together with a nice group of organizations where our research contribution will focus on the design of a scalable content repository, and are now exploring what we could/should do with regards to the user-facing parts of Daisy 3. While much of the focus of D3 will be around scalability (think 'exponent' rather than 'factor'), obviously we also want to provide a nice and user-friendly default content editing and publishing environment - like we do right now.

We've always been proud of our WYSIWYG HTML editor - in the sense that, even though it isn't the fanciest in its kind, it Works As Advertised (tm). Users are able to use both Firefox and Internet Explorer concurrently, while we try to ensure that consistent, well-formed almost-XHTML is saved in the repository.

Obviously, the base platform of in-browser HTML editing has its specific downsides, namely ... that editing support happens as an after- rather than a forethought (here follows a pun about outer-thoughts). Looking at the current state of browser-based WYSIWYG editing, we see some progress and surely sustained effort, but also must realize that the base run-time environment of such kludges, i.e. the browser, still doesn't natively supports the Writeable Web.

That means we have been developing Daisy for almost five years now, and now must realize that browsers haven't evolved quite as much in the direction we want: a stable and developer-friendly platform for building applications that help us realize the Writeable Web. We're not Google, obviously, which kind of came to the same conclusion and started Chrome as a consequence (hoping the corporate world will rapidly switch over to a new web client). Not that there is no progress at all, quite the contrary, but the speed of change is below one would expect.

Still, the alternate options are equally awkward. Java applets running in browsers? An exercise in self-inflicted pain. Flash? It doesn't ship with a serious web editing component neither. JavaFX? Silverlight? Let's see those gather some marketshare first.

Now people might react that the state of the art already allows a lot, and will point me to some sort of nifty Javascript or Flash library which looks Really Really Good and has a great demo page. However, what we're looking for is a lot more boring and mundane:

serious copy/paste buffer support on all platforms, so that I can copy text from Word or OpenOffice and the formatting is preserved while mapped to XHTML constructs, on Linux, Windows, and Mac OS X
the ability to drag-n-drop images from my Windows Explorer or Mac Finder in my editor and a consistent way to upload those (possibly sets of!) images to a back-end
worse, the ability to copy-paste images from Office or other desktop applications and convert those automatically to a decent web-friendly format (like PNG) *after* scaling them to the size I want (and still preserve the original paste buffer in case I change my mind)
real-world, performant spell-checking and auto-correct
supporting longer documents (beyond what Google Docs currently supports) while still spell-checking (an issue in my Firefox installation)
a unified way to handle temporary local storage (for those people like me who hit command-S after every paragraph) like having Google Gears installed by default on all browsers on all platforms - because I'd like to have access to my data and applications on the road
and all that while saving semantically-squeaky-clean XHTML, while still allowing specific extensions to encode non-Web constructs

Simply being able to open up browser content to the Writeable Web. Wouldn't that be nice?

Daisy 2.3 is out!

2009-06-09T11:56:00.000+02:00

We are very happy to announce the release of Daisy 2.3 today.

While we've missed our usual bi-yearly schedule for this release, we believe there's plenty of goodness in this new 2.3 release to make up for this delay:

a spiffy global search-and-replace tool
say goodbye to boring download-upload cycles with our new document uploader applet, bridging the gap between your desktop editors (Office, Photoshop ...) and Daisy
a vastly improved document browser (including a built-in facet browser), everything configurable without any coding, with also lots of tiny usability fixes

and for the technical Daisy integrators out there:

an inline editor hook that allows you to modify the default document editor to your heart's content, e.g. putting fields and parts on a single heavily customized screen
we're using FOP now (instead of IBEX) across the board for PDF generation, even for books
even more granular access control
a lot of new workflow features, and the possibility to query the jBPM workflow engine from inside publisher requests (e.g. to show status per document)

The list of changes between daisy 2.2 and daisy 2.3 can be found here: http://daisycms.org/daisydocs-2_3/13-cd/591-cd.html.

You can download Daisy 2.3 here: http://sourceforge.net/project/showfiles.php?group_id=176692

The installation instructions can be found here: http://daisycms.org/daisydocs-2_3/13-cd.html

Next up, two things will be happening:

work is already underway on the 2.4-dev branch, for the moment mostly related to book-publishing and single-sign-on-integration
we've also set up a formal R&D activity that is focusing on our next-generation content management framework, currently nick-named 'D3'

D3 will be a quantum-leap revolution:

we're moving from Apache Cocoon to Kauri for front-end development
the ambitions we have with the new repository engine will be far beyond what is possible with Daisy 2 now, in terms of volume, distribution, and whatnot.

D3 is planned for 2011. Obviously, these are pretty exciting times for us.

In the mean time, enjoy Daisy 2.3 and let us know what you think of it.

Thanks, Outerthought - or rather http://www.flickr.com/photos/stevenn/3597192333/

Second Wikipodium event

2009-03-18T20:40:25.000+01:00

We'll be hosting the second Wikipodium event, a loose gathering of Belgian wiki-philes. The theme will be "Wikis and Classification, or Water and Fire?", in which I'll try to defend Daisy's überclassification capabilities against an army of opinionated Wikimedia lovers. We provide drinks and munchies. RSVP at the Wikipodium wiki.