Found these good links on Architecture related
Recently in one of the articles in a magazine read about Squeak [in an interview with Allan Kay]. Apparently Squeak is truly portable across multiple platforms. Downloaded Squeak and experimented little bit on programming in smalltalk on squeak. It has better metadata management compared to java. Dev environment is pretty cool... There is good documentation on the website which can take you forward if you are interested in looking at the next generation languages.( Jan 03 2005, 03:16:43 AM EST ) Permalink Comments 
Think of a scenario where we need to rationalize three different user management based data schemas to come up with one so that three different applications can work with a single user management schema. Does it sound challenging ?? Wanted to share with you my experience in this exercise years back. "Schema Rationalization" term popped up in one of Realizing SOA articles. That's how i caught on to it.
In this case, the idea is to first get the unified view and the least common denominator of the three schemas. The migration of three schemas to the unified view is the challenge from both angles : schema, data. I remember challenges in migrating the sequence generator ids from three different schemas. Just to elaborate on the challenge, user id can be a foreign key in transaction entities which makes the data migration problem more interesting. Last but not the least, the authorization requirements. Life becomes really interesting if authorization is not aspected out but all over the place. It was all over the place. The idea was to move out the Authorization related logic into Security Manager Module. SSO was kind of implemented and SSO Agent was checking the token (service ticket) generated at each application regarding the user information for Access, role related details. Token(service ticket) generation was separated out from the three applications. Token was quite similar to SAML effort. After authentication against the database, a unique token is generated which has enough information which can be passed on across applications.( Jan 02 2005, 12:01:17 AM EST ) Permalink Comments 
The context with which I started off this thought is where we have two databases with same schemas but the data may change from time t0 [at that time, they had same data]. To complicate little further, source schema is a superset of target schema. Knowing the mapping of which table [source] goes to target table and transformations [if exists]. Row wise and column wise comparison will be necessary to come up with a report on the differences in both schemas. The other way of finding is maintaining the time stamp/version for every row. After the syncup of two databases, versioning needs to be updated thus ensuring that a simple select on version number after merge/sync will retrieve the records which are different in both databases. These records can be compared with each other and verify if there is any merging necessary. This approach will be complete if the deleted records from both databases from intial point t0 needs to be archived.
To extend the thought, if the application can implicitly note the changes from t0 in a separate schema, like table name, row_primary_key, action.[similar to auditing]. Coming up with a diff report will be relatively easy.
To complicate it further, from the point t0, the schema also changed. The book keeping approach mentioned above may work even if the schema changed. Need to think more on this part...Yet to convert this thought to code...( Dec 31 2004, 12:01:22 AM EST ) Permalink Comments 
Releasing Architect-BOK [ebook in word format]. Need to clean up the articles in the Architect Bank of Knowledge. [selected articles from this blog]. Attached below are the compiled links, if you want to browse the BOK online. For offline reading, check out:
Going forward while in defining architecture, one of the key concerns which will be primary will be Compliance. [ Basel II,SOX,GLBA]. Attempt here is how to address it while architecting. Some of the ideas are like enforcing rules while the business processes are executed. For example, while a loan application is filled/processed/approved. the rules will be processed against compliance bounds for different inputs. Events and compliance reports need to be generated on the compliance violations categorized into severe,critcal, medium, low. Document management needs to be have the compliance checks in versioning, checking in processes. Deleting, Archiving also needs to be touched upon. Audit trails will be a key feature in the data view of the architecture. From the user interface view, to raise exceptions on the filled data, data elements and their valid bounds of data need to be gathered and provided. To start with, Compliance Metadata requirements gathering will be helpful in addressing the compliance concern in the architecture. Exception management will be tied in tightly with compliance.
On a separate thread, identifying aspects which are compliance targets and designing them as aspects will be helpful in long run. loosely calling as compliance aspects.. need to hash out more on this thought...( Dec 29 2004, 12:01:15 AM EST ) Permalink Comments 
Started blogging on Dec 28, 2003. It has been a great journey in completion of one year. Forging ahead to 2005, lot more to come..
check out the book and references for getting started on Biometrics. I found the above book great for a start. Other than watching in hollywood, for me domain knowledge of biometrics is very low. The whole process of biometrics capture,identification, verification, scoring and other related features are handled very well in the book. The author has touched upon five vertical markets where biometrics is active and standards which are existing for biometrics. The standards exists like BioAPI, BAPI, HA-API. As mentioned below, there are opensource projects but not much activity is observed. The reference implementation of BioAPI is found at BioAPI org. Microsoft has support for BAPI.
I was reading an article on getting the price right. It is interesting that simulation can really help solve the setting the price right as non linear dynamics of segments of customer, their price expectations, demand-volume distribution. The dynamic value management can be simulated with the segmentation data and demand-volume regressed data availability. Gathering the price fluctuations during this dynamic changes in demand, matching customer expectations will be helpful in analyzing the behavior of the market before releasing to the market. Simulating Competitor moves will be challenging. But currently my thought is to use game theory to simulate this.
On a separate thread, this can be used for introducing new product, expiring old products. I am very sure that these thoughts are already converted to code somewhere..[didn't do a search on accoona.] But I feel the data collection is going to be the key for simulation to be successful. Another thought which comes to mind is to use merge simulation + reality data to more realistic. Fuzzy Logic based control and revising the simulation events or demand-volume information would be more fruitful in tracking, correcting the direction of dynamic value management and new products release into the market... Need to hash out more on this idea to convert into code...( Dec 26 2004, 12:01:56 AM EST ) Permalink Comments 
check out the Java Studio Creator Link. My first attempt is as below. My First impression says "Awesome". But please go ahead enjoy the experience of building a web page (JSF), bean etc....
If you haven't checked out Accoona.. Go ahead and check out what's new from Accoona. On a separate thread, if you haven't visited google related articles... check that out too..( Dec 23 2004, 12:01:45 AM EST ) Permalink Comments 
Check out this book on Ant Colony Optimization. It is an excellent read for a person with optimization or one with algorithmic bent of mind. The whole idea how ants optimize and identify the most optimal path through phrenomes detection, thickness of phrenomes and evaporation part of it is amazing. How it can be applied to real world problems like Vehicle routing, scheduling, of course Travelling Salesman problem. After reading this book with first sip of coffee today morning, i have different perspective about ants and now more respect towards them -)). Last but not least, don't expect code samples etc.,.. Planning to incorporate this metaheuristic in( Dec 22 2004, 12:05:12 AM EST ) Permalink Comments 
Got a chance to look at the latest wireless toolkit, (2.2) after a long time. Pleasantly surprised with latest and greatest features like Messaging, Media API, OTA.( Dec 21 2004, 01:56:29 AM EST ) Permalink Comments 
This is a personal weblog, I do not speak for my employer.