Monday, May 26, 2008

What makes Open Source Communities Thrive






Today I attended a session by David Axmark - Co-founder of MySQL, one of the most popular enterprise level database which was recently bought over by SUN and has been open sourced.

The session was one of the general information session which covered, high level design, road map, community development etc about MySQL. David strongly advocated "open source" strategy for his database technology. I liked his thoughts about open-sourcing and here are some of the take aways ...

Faster Turn around for product defects : Bugs reported on the propritory software take ages to resolve (most of the times). Bugs once reported get into vendors tracking database and most of the times user does not have a mechanism to see the status and contact the designer working on it. On the contrary, bugs filed against open source software are open source as well. Community can see the bugs and anyone is free to pick up the bugs and drive them to resolution. In most of the cases, this is fast, transparent and robust

Faster time to market : Contributions to open source is not bound to any release cycle or does not result out of any market forecasts. Community developers often use the open source code base to add on value added features and support as and when it appears relevant to them. In the bargain, as the community gets active around the product, product gets feature rich exponentially. To cite an example, SQL Server supports almost all languages including the recent once like ruby etc.

Quality Aspects : Open Source products are built under strict community vigilance. 100 balls are watching every moment to any changes being done. This adds quality to the contributed source that finally makes it to the production.

On being asked about what does it take to build a successful open source community like MySQL, he said ....

A. The product must address the real problem for which people still struggle to have a good solution.

B. Resort to short but relevant and quick product updates. Typically , come up with releases that have one big increment and hundreds of small but relevant updates to start with.

C. Complement the offerings with good documentation and easily accessible binary downloads so that the product can be quickly accessed, configured and used. Its a good idea to maintain a 15 or 20 min time threshold to what user would need to get started with the product. Complex configurations and heavy installations are deterrents to early inceptions to some otherwise great products.
Sphere: Related Content

Thursday, May 22, 2008

Sun JavaOne 2008, San Francisco, USA : A picturesque summary

Ever imagined dressing up yourself with an expensive embroidered shirt and then hiding it behind a high neck black coat. This, by no means, demeans the worthiness of the shirt you possess however, your inability to sport it to the outside world definitely makes your pricey possessions lose its steam.

SUN Java One Conference is all about this. This is the confluence where geeks and gurus of technology industry around the world meet up and exhibit their latest and greatest products and inventions on a common platform provided by SUN Microsystems.

I was fortunate to be part of SUN Java One 2008 this year. Java One was held at San Francisco, Moscone center between 6th - 9th May, 2008.

Here are some of the things that I did there :

1. Conducted Pod Duties on the Project Mural Pod.














2. Spoke to Pod visitors at length about Open Source MDM (Master Data Management) offering and role of Project Mural in packaging all these together. Mural Pod witnessed variety of visitors that varied from CTOs, CEOs, Startup founders, journalists, analysts, students, technology freelancers, tech freaks and last but not the least, goodies seekers !

3. Running visitors through Sync Demo put up by the team at the Project Mural pod.

4. Delivered a community cornet Talk on Project Mural with Srinivasan Rengarajan (Mural Architect). The talk was well received by the attendees.



5. Meeting with Students from SJSU (San Jose State University) and their Dean, Dr Dan Harkey, to discuss about the ongoing engagements on research projects with them.

See Students and Mural Community for more details.
Also see Blogs from Sandeep Konchady, Mural community coordinator at SJSU.




6. Attended general session by James Gosling, co-inventor of Java.

7. Proctored following Java One Hands - on - Labs:
  • Plug into GlassFish™ V3 With JavaServer™ Faces and jMaki
  • SIP Performance benchmarking
  • Adding Convergence of media to your Java EE application using NetBeans and Sailfin.


8. Visited other pods and spent some time to understand the offerings from other open source vendors.

Here are some pictures that explains the grandeur and exuberance at Java One Pavilion :


Sphere: Related Content

Sunday, March 30, 2008

MDM Summit Demo: Some Learnings and a Great Team Work ....


Still seas do not make skillful sailors ... that was the message emanating from the Gir conference room at India Engg. Center of SUN Microsystems on the evening of 27th Feb 2008 as team Ahi was working full throttle on the MDM Summit Demo to be delivered around 8:00 PM to Monrovia team that evening.

Though the work had already begun couple of days before on the individual components for the demo, real impetus came in when the integration started with entire team getting involved. The stage was set, goals were clear and the eyes were focussed. The team got into a common room after a pint of wisdom from Ahi. All discussions came to a screeching halt and it was time to execute and build the bones together.

What appeared to be a sprint started taking the shape of a marathon when the team was unable to find a stable Netbeans installer to use for the integration. Version that was used for trial integration was phased out from the build machine. Local copies were also lost. We all groped for an Installer for close to six hours before we got something that looked promising !

Learning #1: Archive the best working software pieces that are used in the integration dry run. You never know when you would need it.

By the time things started getting into the shape at Bangalore, team at Monrovia started coming in and waited for a while before we could send instructions and software to proceed.

Learning #2: Work in a parallel mode with geographically distributed teams. Anticipate latency when huge pieces of software components need to be moved over the web.

As we chugged our way to the dawn, we started facing unusual software issues that required quick fixes and patches. Problems started creeping in when we unplugged the test data that we were working with till then and started using rich data sets created specifically for the demo scenario. The data created had been laying unattended for hours while other minor engg. issues were being fixed at mid-night.

Learning #3: Do early integration (as early as possible) and save time and last minute surprises.

After wresting with issues till 11:00 AM next morning with only occasional forty winks on the respective office chairs, the team started wrapping up gradually. The final bugle was blown at about 3:00 PM and that was when Ahi and Nilesh left the cockpit and ensured safe landing with a successful demo. Srini, went last, ensuring that what we delivered worked fine !

Thats the smile of contentment that keeps the life going .. despite all odds. Part of the game .. eh!







Sphere: Related Content