Why Wiki , Blogging & Social Computing

Blogging and wiki is something about the knowledge management, one step further than the information management. There are many ways of knowledge management, blogging and wiki are just one of them.

why we bother to blog or wiki it rather than make a phone call, send a email or book a meeting room? Like it or not, social computing is part of how organization works today and an even larger part of how it will work tomorrow.

Social computing is what happens when people selectively consume, produce, and share online content in new ways. Technologies like blogs, peer-to-peer networking, RSS, wikis, and other collaboration tools continue to grow, evolve, and feed the demand for new methods of interaction. The challenges that you face today as a result of access to online information, software, and communities will intensify, with new challenges emerging over time. To maintain influence in this new world, you must embrace social computing, not fight it. You must become a facilitator for this kind of collaboration and a full member of the social computing community.

Develop your interest in using these forces to do your job more effectively, rather than letting them just make your job more difficult. Instead of seeing this as a new onslaught of back-door paths for unwanted change to reach your organization, adopt a leadership role in using these tools to create a more collaborative and productive organization.

Don't fight this trend, embrace it. You should become an advocate and facilitator for social computing. Make sure people can get what they need within the organization so that they won't have to go elsewhere. Make yourself part of the social computing community, and use that insider advantage to help the organization succeed.

As we take advantage of social computing, don't neglect traditional means of communicating and promoting collaboration. Face-to-face idea sharing remains an important tool in building and nurturing critical relationships, and we should continue to use it for communication.

KM Tool - Wiki

Wikis are tools that allow for collaborative editing and publishing and hence the creation of user generated content. Wikis are another example of social software - software that encourages people to work together.

Wikis are a useful tool when you want people to collaborate and work together on creating a common document, e.g. a project document, a manual, and so on or to work on a common strategy or event, e.g. a conference programme.

In corporate and general use they can be used to create collaborative documents, manuals e.g. WikiHow (a collaborative Manual)

The most obvious and famous example of a wiki is of course Wikipedia, which is an encyclopaedia that has been created by thousands of individuals. Every page in Wikipedia can be edited, which means it is constantly evolving and growing and can be kept up to date.

In the commercial tools market there are a number of enterprise wiki tools available, for example, e.g. Confluence

But there are also a number of free tools, for example:

* MediaWiki is the software used for Wikipedia and the other Wikimedia Foundation websites. Compared to other wikis, it has an excellent range of features and support for high-traffic websites using multiple servers (Wikipedia peaks in the 2500+ requests per second range as of June 2005).
* pbWiki is another hosted service and allows you to set up a wiki very quickly, as are Wikispaces and Wetpaint
* If you are looking for a personal, notepad-type wiki to download onto your own computer, then TiddlyWiki is a useful tool

Reference list

Mediawiki Installation & File Upload HowTO
http://www.washington.edu/computing/web/publishing/mediawiki.html

MediaWiki Comparison Matrix
http://www.wikimatrix.org/show/MediaWiki

How to install PHP and MySQL
http://www.ozzu.com/ftopic56303.html

Wikis - and how to create them quickly and easily
http://c4lpt.co.uk/handbook/wikis.html

CVS vs SVN

For CVS, Things like tags and rolling back to older versions is far easier with CVS. And the lack of 'all or nothing' transactions is not a big problem. just add a 'build-****' tag regularly. This makes regressions very easy to track. The 'Revision Graph' in TortoiseCVS is also worth gold.

In CVS, multiple versions can get the same tag, which is logical, but with SVN it would have to 'version' the tags. A tag is a tag in CVS, but it's a folder in SVN.

Operationally speaking, maintenance of repositories and modules are easier in SVN than in CVS, in my opinion.

For development, the Tags & Branches function is very important which CVS is better than SVN. As SVN is actually using different approaches for tagging ( Basically, a tag and a branch, for that matter in SVN is kind of like a copy. However, it's more of a "smart" copy).

CVS Convention and Best Practice
http://tldp.org/REF/CVS-BestPractices/html/index.html

How to install CVS in RedHat Enterprise
http://personal.vsnl.com/sureshms/linuxindex.html
http://www.michael-amorose.com/articles/computers/cvs/6_1.html
http://www.taursys.com/howto/cvs/
http://www.faqs.org/docs/ldev/0130091154_198.htm

CVS Installation for Windows
http://www.devguy.com/fp/cfgmgmt/cvs/cvs_admin_nt.htm

useradd -r -c "CVS user" cvs
passwd cvs
cvs -d /usr/local/cvsroot init
chown -R cvs.cvs /usr/local/cvsroot
cvs -d :pserver:daps@localhost:/usr/local/cvsroot login

CVS Tutorial Link
http://www.cvshome.org/
http://www.developingprogrammers.com/index.php/2005/11/24/cvs-and-subversion-combined-tutorial/

CVS vs SVN reference Link

http://www.pushok.com/soft_svn_vscvs.php
http://www.devmaster.net/forums/showthread.php?t=7394
http://svn.haxx.se/users/archive-2006-10/1382.shtml
http://lists.danga.com/pipermail/mogilefs/2006-June/000336.html
http://wiki.scummvm.org/index.php/CVS_vs_SVN
http://www.gigascale.org/softdevel/faq/27.html
http://users.footprints.net/~kaz/mcvs-svn-comparison.html

SOA, Web Services & BPEL - Tech Forum in Singapore

Recently, i gave a talk and BPEL demo to singapore partners for Oracle BPEL & SOA suites. Introducing them idea of SOA, BPEL and Web Services and how to make use of oracle technology in their applications.

SOA stands for service oriented architecture, it is a architecture model.

Talk about architecture, we need to understand basic building block of each application and system.

Like we build a real house, we use mortar and brick. Every application / system was built with business logic and process. Business process weaves all business logics into one application to achieve business goals. Process and logic weave seamlessly and hard coded into a monolithic application. Like we cook soup, we put all ingredients into water and hard for us to separate each out in the end, by which it reduce reusability of standard process and logics.

In traditional application, when people wanna change the process or integrated with other application / system, especially it is a heterogeneous platform or technology integration, that will be very hard and time-consuming.

For SOA applications, applications also build with process and so-called business logic (have been wrapped into self-contained services). Process and services are loosely coupled, people can easily add or remove services from process, not like traditional applications (process & logic fuse into one). Other than that, people can easily design & program the business process with a advanced IDE. Overall experience for process design and implementation will be a drag-and-drop exercise.

SOA process should be standard, easily extensible and composable. When build an SOA application, we start from building a portfolio of services , follow with well-defined business process to orchestrate services to achieve business goal. This is where BPEL comes in for business process implementation and web service for business service implementation.

Many people equate web services with Service-Oriented Architectures (SOA), but web services are just one element of SOA - they're about service invocation. SOA's scope is broader than just invocation. It's a complete software component architectural model, a way to organize, manage, deploy, discover, and expose software components, and to consume these over the Internet. From the software life-cycle perspective, in addition to the typical software development phases of design, development, and deployment, SOA defines web services-specific phases that include service publishing, discovery, binding and consumption, and management.

Note, SOA != web service+bpel

ESB vs BPEL

Both ESB and BPEL can be used for SOA implementation.

ESB (Enterprise Service Bus) implement messaging to enable services to be integrated in a message-based paradigm: both synchronous and asynchronous styles. ESB moves data via different adaptors (WebServices, FTP,File,JDBC etc) and protocols (HTTP,JMS) and enriches and transform data using XSL & domain value mapping lookups.

Oracle BPEL provides a comprehensive, standards-based and easy to use solution for creating, deploying, and managing cross-application business processes with both automated and human workflow steps.

ESB is good for routing messages to multiple destinations. It is also good for doing transformations that have little to no business rules. The footprint is much smaller and incurs minimal overhead therefore the performance is much better.

BPEL is used for bringing together multiple services. There is much more functionality and allows implementation of complex business logic.

• BPEL data transformations enrich and perform complex changes whereas ESB perform only simple transformations.
• BPEL can use Business rules, Human Workflow and Notifications, exception handling can be done in BPEL.

Use ESB when you want a really low cost solution, Only need connectivity, simple transformations and routing.

BPEL is primarily used for servies orchestration, data enrichment and human interaction.

Use BPEL+ESB when, You need to code complex business logic, integration with Workflow, complex transforms potentially involving the database or having long running processes which may or may not be stateful.

With an appropriately layered SOA architecture, one would look at ESB as a repository of services that BPEL orchestrates. At deployment time, this layered approach provides you the flexibility to dedicate servers to individual tiers (BPEL and ESB) and thereby maximize the resource usage for your particular use case.

Io exception: The Network Adapter could not establish the connection

Recently, i did an RAC validation with our oracle partner and encountered this weird exception called "Io exception: The Network Adapter could not establish the connection".

Two client machine were running same java code (using jdbc thin driver & oracle datasource) in the same LAN as RAC servers, but they came out with differen behaviour. One could connect to RAC database, the other one always got io exception mentioned above.

The only difference between two machines was one of clients' host file having rac ip & hostname mapping entry, which is able to connect to oracle rac db.

Something weird is that all db clients and servers were actually sitting in the same LAN. I could ping and telnet db rac with rac ip address (virtual and real ) from client machines, but when i use ip address in the java class to lookup the db connection, i could not get connected to db if rac ip & hostname mapping entries were not available in the client host file.

Here is the sample Codes for RAC FCF connection:

OracleDataSource ods = new OracleDataSource();
OracleConnectionCacheManager occm = null;
ods.setUser("user-name");
ods.setPassword("password");
String cacheName = "cache";
String dbURL="jdbc:oracle:thin:@"+
"(DESCRIPTION=(ADDRESS_LIST=(LOAD_BALANCE=ON)" +
"(ADDRESS=(PROTOCOL=TCP)" +
"(HOST=rac1-vip-address)(PORT=1521))" +
"(ADDRESS=(PROTOCOL=TCP)" +
"(HOST=rac2-vip-address)(PORT=1521)))" +
"(CONNECT_DATA=(SERVICE_NAME=orcl.abc.com)))";
ods.setURL(dbURL); System.out.println("Url=" + dbURL ); ods.setConnectionCachingEnabled(true);
Properties prop = new Properties();
prop.setProperty("MinLimit", "5");
prop.setProperty("MaxLimit", "40");
prop.setProperty("InitialLimit", "10");
prop.put (oracle.net.ns.SQLnetDef.TCP_CONNTIMEOUT_STR,"" + (5000)); // 5000ms ods.setConnectionCacheProperties(prop);
ods.setFastConnectionFailoverEnabled(true);
ods.setONSConfiguration("nodes=rac1-ip-address:6200,rac2-ip-address:6200"); ods.setConnectionCacheName(cacheName);
//occm = OracleConnectionCacheManager.getConnectionCacheManagerInstance();
Connection conn = ods.getConnection();

It looks somewhere JDBC driver looking up db connection with rac hostname instead of ip address and I have to put the rac ip & hostname mapping in the client host file for jdbc driver to resolve them.