Eduardo Pelegri-Llopart's Blog

Community Archives

Reference Implementations and Production Quality Implementations

One of the reasons for the success of the JCP is that it requires a Specification, a Technology Compatibility Kit (TCK) and a Reference Implementation (RI). The TCK is intended to cover the specification and the RI is required to pass the TCK; together, these deliver specifications that are implementable and testable. Together with the community participation in the Expert Group and in the different Review Processes, all this leads to good, useful, implementable specifications.

A popular misconception is that a Reference Implementation always is a "toy" implementation, unsuited for production use. This is not true. An implementation being an RI just means that it satisfies this specific role in the JCP process; there are all types of RIs. Some implementations are indeed just "proof of concept" but others are production-quality and are used in commercial products.

An example of this misconception applies to Project GlassFish. Every now and then I hear somebody saying that "... but it is just a Reference Implementation". The implementation created by this Community is the RI for Java EE 5 but that implementation is also distributed as the Sun Java System Application Server 9.0. Part of the confusion is because at some point in the past Sun had two AppServers; that changed years ago and it is no longer the case but misconceptions take a life of their own (as politicians, movie stars, and many others know...)

Hope this helps. I'll contact Onno and suggest he add some clarification in this topic to the JCP FAQ.

BTW, if you are interested in Project GlassFish and the Java EE 5, you may want to check a note I recently wrote for TheAquarium collecting 20 Things You Should Know about Project GlassFish.

First 5 months of TheAquarium - Reporting on GlassFish and more...

I've not been very active on my Java.Net blog because I have spend most of my blogging time coordinating and contributing to the The Aquarium. We started TheAquarium on November 30th and I provided a Report after 1 month; since JavaOne is around the corner, I thought I'd do an updated report now, at 5 months, rather than wait for the customary 6 months.

The Aquarium is a News Group blog; we started focused on all portions of the GlassFish Community (including JSF, JAX*, JWSDP, etc) but we also cover several other related communities like Open ESB, Portal, Derby, Web Server and AJAX and Scripting, as well as tools (mostly NetBeans but also some Eclipse) and Mustang. The Aquarium is a major contributor to the list of Frameworks and Applications that work with GlassFish.

In the first 5 months we have written over 570 entries. Counting the original sources (bloggers) is harder, but I did a quick pass and guesstimate it at over 260; some are very prolific, some not. We seem to have stabilized around 5-7 entries a weekday, we are mostly limited by the ability of the editors to keep up with the sources and we are planning to add a planet aggregator to address that. Most of the sources are from Sun but the ratio of non-Sun bloggers is increasing steadly.

We publish two localizations: Chinese - led by Qingqing and Spanish - myself - that provide an additional outreach into specific communities complementing the main blog, which is in english.

We generate a Weekly Roundup of the news and there is also a Search Facility.

Blog readership and impact is always hard to gauge accuratedly, but we are happy with visitors: we are always in the top 10 most popular blogs of Blogs.Sun.Com, often in the top 5 and we have been top 1 several times. Our repeat visitor ratio is excellent, over 30%, and annecdotal buzz is very positive ("Excellent Aquarium", "Great blog"). We believe that The Aquarium has been very succesful as a Knowledge Base and also has worked well increasing awaress of the projects it covers.

Looking to the future, we just added a new editor, Ron, that will focus on SOA, JBI and ESB, and we expect another editor. We also have a Japanese localization ready and we are considering one more. We are also planning some focused coverage of JavaOne.

Overall, our experience with The Aquarium has been very positive and very much welcome any suggestions you may have to improve it.

Time Zones Don't Matter in the BlogSphere

The World is smaller and it only takes a few minutes to get an email message across almost any two points. Geographic distance does not matter, but in one of my earliest blogs I argued that Time Zones Still Matter in the Internet. But that statement was based on email interactions and for the last few months I've had quite a bit of experience in the use of blogs over distant Time Zones. Based on this new experience, I now believe that Time Zones Don't matter in the BlogSphere.

The contributors to the GlassFish Community are from many locations across the globe. Over the last few months these engineers have started blogging with increased frequency, and since late November, several of us have been using these blogs as sources to create a news blog (The Aquarium). Most of the blogs are very informative and, somewhat to my surprise, the geographic origin of the blog - and its Time Zone of origin - is totally irrelevant to its relevance and impact.

I think that what happens is that the communication style encouraged by blogs encourages a careful writeup that is self-contained, which is exactly what is recommended for communication across distant Time Zones. Also, the comments of a thread create a stream of communication that is directly tied to that content, and in most cases, it is quite acceptable to the author of a comment if the response happens many hours after the posting. All of this means that the author of a blog can be many TZs apart from the reader, with no substantial impact on the quality of their interaction. As a typical example, Sahoo is located in Bangalore, and I am located in SantaClara, California but, as an editor and a reader at TheAquarium, he is just one of the good contributors at TA.

There are many types of blogs, and I don't want to make a universal statement, but our technical blogs have proven to be quite immune to the Time-Zone problems that are very evident in email. In the new world of global communities, blogs are proving to be a very useful tool.

JavaPosse is neat...

Not too long ago I thought it would be neat to try an audioblog of Java News... Not a very original ideal perhaps, and it turns out that Tor Norbye, Carl Quinn and Dick Wall have been doing a very good job with The Java Posse. They carry a pretty good set of Java News and Interviews and are already in volume 11. Cool!

A pitty that my car does not have an MP3 player... maybe I'll have to finally break down and get one. Anybody knows of a good, not too expensive, solution for my Prius?

JDJ 2004 Reader's Choice - interpreting poll results...

I just noticed that the Java Developer's Journal is holding their annual Readers' Choice vote. At one level this seems a good idea, but, unfortunately, in this type of poll the results can vary substantially depending on who votes, and that can change a lot through a few well-placed hyperlinks. The consequence is that the results are, in my opinion, not necessarily instructive. Polls and surveys are always tricky, as we know from the recent US election, but I find the results from, say, the Evan's North American Development Survey much more interesting, even though it has a relatively small sample.

In any case, if you want to vote on the JDJ Reader's choice, go to the JDJ site, but, if I read the information at that site correctly, they are about to close the vote, so, if you just found out like me, you may be late.

Community Spring Cleaning

We have been doing some Spring cleaning and I want to give you a report on where we are, tell you about some future directions, and ask for feedback.

Home Page: The three panels in the Community Home Page now have different types of content: general news on the right panel, community news on the center, and community resource links on the left pane. We are trying to update the right pane several times a week with news related to this community collected from different RSS feeds.  The center pane is updated less frequently, with news from within our own community.  The left page is mostly static and contains links to different resources, including blogs, Wiki projects and people.

Types of Projects:  Projects at Java.Net can be Hosted or Linked.  Hosted projects maintain most of their content using the Java.Net machinery; linked projects exist elsewhere but they are listed in our directories.  Further, projects may be listed in an incubator if they are not yet ready to be fully advertised to the community.  Finally, we also track products of interest to the community.  I've contacted all the projects in the last few weeks, classifying them and, in some cases, deleting projects.  As with any cleanup effort, keeping it up to date is a never ending effort, but we are reasonably up-to-date. Right now we have about 28 projects in the main directory and about 18 in the incubator, with about 13 additional linked or other projects.  We get a few additional project requests each week; we are listing the new projects in the left pane.

Project Directories: There are multiple 'project directories' at Java.Net, some maintained by the SourceCast machinery, some are updated at Java.Net.  We are now trying to keep them all synchronized.  The directories we highlight in the left pane are Wiki pages that list all the projects and will eventually track their status.  We track separatedly the hosted projects (be them in the incubator or in the main directory) from linked and other projects and products.

Downloads: We are running weekly stats on access to our projects (thanks to Kohsuke and Ryan for this!) and we now have a link of popular downloads on the left pane.

Future Directions: There is a fair amount of effort in just keeping the existing infrastructure working; but, beyond that, there are a few ideas that we want to explore:

• A Technical Forum in this space: I have talked with the Java.Net administrators about moderating a technical forum where invited speakers would kick-off a discussion on a technical topic.  I host similar forums at Sun and I have a fairly good set of contacts to get the forum started.
• Improve Project, Product and Learning Content of Wiki: The wiki can contain much more information on these areas.  Just as an example, the learning Wiki page only lists Sang Shin's online course.
• Community Forums: We are currently not using any of the forums for this community.  We could start using them as a way to build community.
  

Total Cost of Development and Developer Communities

The term total cost of ownership is frequently used to capture costs, sometimes forgotten, involved in owning a system. I've found useful to use the term total cost of development in a similar manner, capturing some of the less common contributors. Some of these costs are:

• Training
• Examples
• Support
• Risk of choosing the right solution

Training, examples and support should, ideally, be targetted to a the specific needs and background of the developer. Many people remember these contributors, but the last contributor is easy to overlook: when a developer chooses an approach to solve a problem (architecture, tools, platform, whatever) she is taking the risk that the approach may not be a good match to the problem. The mismatched approach may increase the cost of implementing the solution, or may even require a restart of the project.

There are other contributors to the TCD (tell me in the talk back), but what I find interesting is that all these costs are very well addressed by developer communities. The example that I often use is Struts . I don't want to minimize the technical benefits of Struts, but I believe that the main reason why Struts is so succesful is the very strong developer community built around the code which help reduce the tocal cost of development by addressing the contributors I mention above.

There are a number of developer communities, from our own Java.Net communities (I belong to the WS & XML and the jwsdp communities) to those provided by many vendors, includinng Microsoft.

Buiding a community is hard and takes time to build, but when they do a good job, they are extremely valuable. I think we are still, collectively, learning how to make these communities best serve their members.

Distance in the Internet... Time Zones and Geos

I recently attended a presentation on a study corelating the mode of interactions (face to face, phone, email) between participants with the distance between them. The study used geographic distance and reported that people geographically close to each other would use face to face communication and that they would start using phone and email as people got farther away . This may be correct for the community used in that study but the results do not match my experience: we use email much more often than that and I believe geographic distance is not a good way to measure distance in the internet.

I interact with many people, some have offices in my building but others are located in other sites in the West Coast (of the USA) and elsewhere in the world. Most of my interactions with other software engineers are through email, regardless where the recipient is.

Email is the prefered medium for our engineering community. Email is asynchronous, fast, and can be scanned quickly. We complement email with (synchronous) meetings where people are seated around a virtual room, some face-to-face, some teleconfering using phone, video and VNC. Some groups also use IM and chat rooms.

My experience is that geographic distance is much less important than time-zone distance. When the time-zone distance is small, email and even teleconfs hide any geographic distance. You can engage in a sequence of email messages and solve a problem. Or you can arrange for a virtual meeting and do high-bandwidth exchanges. Sure, there are some problems with virtual meetings and one needs to pay attention to the non-local participants, but the problem is manageable.

Start increasing the time-zone distance and communication becomes more complicated: the number of mail exchanges in a day are reduced; there are fewer overlapping hours in the work day. Increase the time-zone distance, or include multiple sites (say India, USA and Europe) in the conversations and email exchanges slow to one a day, and meetings just can't happen.

So, what works for collaboration over large time-zone distances?

Here are some ideas that we have used:

• Design out the problem: rearrange the tasks so that where fast turnaround is required the time-zone distance is low.
• Complement with additional interaction mechanisms: I've found web sites, Wiki and bulletin boards to be less susceptible to time zones as they are more static and capture snapshots of exchanges. These mechanisms are not a complete solution; if you have suggestions and/or expererience, please share it.
• Kick-start the relationship. The effect of time-zone distance makes it specially hard to "figure out" the other party so one can translate idioms (typed or vocal) into their intended meaning. One trick that we have used is to ask for a picture and for some short biographical blurb of the other party and then we post this in an internal site. Then it may be easier to add other bits of information to that core to "figure out" the party. Another, more traditional mechanism in the business world, is to fly some people across the time zone.




Java.Net for non-English speakers - a Call for Action...

Posted by pelegri on June 13, 2003 at 04:10 PM | Permalink | Comments (1)



I was born in Spain (in Barcelona, Catalunya), and grew up in Venezuela where I went to college. I moved to the San Francisco Bay Area for my graduate work and eventually came to work in the Java platform group at Sun Microsystems. Now I'm used to being at the center of the Java universe, and for many of my interactions I either walk down the hall, pick up the phone, or have a quick email exchange (too much of that, but that is a different topic!). Living here it is easy to forget that for some people it may be very hard to find information about Java, specially informal information, and specially if you are not very comfortable in English. And that it can be hard to find a developer community.

Yesterday I was at the java.net BOF that JBob put together. JBob says that the oversea response to java.net is extraordinary. To me that makes a lot of sense: I believe that java.net will help all Java developers, overseas or not, to be better connected to each other. Some of that connection will be in English but not everybody is fluent in this language, and I want to encourage us to remember the needs of those developers. We can do this in a number of ways: by using the community to localize documentation, by creating forums around specific languages and by writing localizable applications. It will take a while, but let's get moving.

As an experiment in this direction, and carrying the hat of manager for the Web Services and XML community, I've created two forums in there: one is intended for conversations in Spanish, I will participate there (although I'm going to take a post-J1 break starting Monday); the other is for conversations in Japanese, I hope Kohsuke Kawaguchi will participate there. If you are fluent in the WS & XML community and are fluent in the language drop by and give us a hand. Let's see how it goes; we will report back.

Muchas gracias,

- Eduardo Pelegrí Llopart / Eduard Pelegrí i Llopart





This work is licensed under a Creative Commons License.



Powered by
Movable Type 3.01D