RDBMS - Foreign key indexing

To decide if a foreign key needs to be indexed or not, I follow a simple rule:

I always/only create an index on a foreign key whether:

1 - A deletion on the parent table is allowed and it triggers a cascade delete on the child table
2 - There's need to perform JOIN queries from the parent to the child

In the first situation, an unindexed foreign key will force a full table scan for each parent record deleted. In the second situation, a lack of the foreign key index in the child table will slow down join queries.

I rarely find suitable to create indexes on foreign keys in other situations though.

Oracle JDeveloper and backups

I usually write Java code using Oracle's JDeveloper IDE. In my opinion it's a very good product which I've been using for years. I like it. But yesterday it freaked me out.

Yesterday I refactored a project I'm working on, including change of package names, renaming some java classes and functions and so on. Once I had all my sources organized, I thought that would be good to clean the directories where the old compiled class files reside. There's no need to do it manually as JDeveloper can do this for you; just go to the Run menu and execute Clean project (keep reading before trying!). Once finished I continued coding normally, but in the next build, the compiler complained about source files not found. What?! The cleaning process cleaned so deeply that even some (actually 6) of my source files had been completely wiped out of my hard disk.

Luckily, I had a recent backup and the lost files had just a few lines of code changed. I work with a laptop and I confess that I usually wait too much between backups. I've been very lucky this time; be careful if you can be affected by this issue and more important, keep regular backups, even if you work with a CVS.

I'm using JDeveloper build 10.1.3.34.12. It's an Early Access version and I'm convinced this issue will be solved in the final release.

On another note, I just use the IDE; I try to resist the temptation to stick to their frameworks and libraries because, although they're excellent pieces of software, they usually work better/only with a commercial Oracle product in the backend. If you're developing Java applications interacting with an Oracle database though, Oracle ADF (Application Development Framework, former BC4J) is a master piece of software that excels specially when used with complex relational data models.

[Update (2006-06-11)]: I've been using JDeveloper 10.1.3 (With Service Update 3) for some time and it works perfect; the 'cleaning too much' issue is no longer present. Suport for CVS (and Subversion too) is excellent. Nice.

What a wonderful world

"I see trees of green, red roses too
I see them bloom for me and you
And I think to myself, what a wonderful world"

Louis Armstrong's song has come to my mind while visiting the Forests Forever website. You shoudn't miss it; it's worth a visit. You can enjoy really beautiful pictures of some of the most beautiful forests of our planet.

I don't feel like sending out a call for action, but it's good to know that at the moment we all can visit those -still real- places. It depends on us to keep them alive. Forever.

On RDF

RDF or Resource Description Framework is about describing web resources; where a web resource can be whatever (web pages included) you can reach with an URI. RDF is seen as the way to follow to reach an enhanced WWW known as The Semantic Web. But why do we need the Semantic Web? Why do we need RDF?

This weblog post can help you realize why the web needs to be semantically enabled. The author says that if you are looking for a hotel in New York City and you start searching for “New York Hotel” in a search engine, you’ll end up with wrong results like “The New York Hotel in Las Vegas” or “Hotel New York in Rotterdam”.

Web pages don’t have information that states that “this page is a Hotel’s web page”, and that “the hotel has WiFi” and that “the hotel is located in the city of New York” and “the price range is between this and that”, and so on. This scenario can be reproduced with most of our daily searches. The crawlers of the search engines cannot do anything more than scan the whole text of the web pages and do some magic to allow us to search, hopefully finding what we are really looking for.

The promise of the Semantic Web & RDF is an improved Web where we could effectively find what we are really looking for.

But, and this is an important but, before RDF reaches critical mass, there’s a long way to go until we ‘normal’ people start thinking that the Semantic Web, RDF and metadata is something we all need.

Many people argue that the problem with RDF is its RDF/XML encoding; it’s difficult to read and thus stops people from using it. Although I agree completely (I understand RDF quite well but I often fail miserably when I have to write an RDF/XML document by hand), I think that there is another reason: For almost all users of the web, writing descriptive information of their web pages is not cool, metadata is not cool. HTML reached mass adoption because if I change a paragraph and hit F5, it appears the page updated in front of me. With metadata the benefits are not so evident.

In my opinion the big deal is to make metadata attractive for everyone. We should all be metadata addicted for the Semantic Web to arrive.

Having said that, I’m optimistic regarding the Semantic Web. If it doesn’t arrive, then I’ll be pessimistic about the future of the Web though. Some interesting things have happened recently in the Semantic Web arena; for instance the release of a new RDF query language and protocol by W3C (SPARQL) and some curiosities too like the fact that Google is hiring RDF guys...

If you want to read about RDF, don't miss this quick and excellent introduction to RDF.

Ski week

Last winter some friends and I spent a nice week of back country skiing in the stunning ski domain of Chamonix (France). Here is a picture that silently evokes how good were those cold and sunny ski days.


Picture taken with a Canon Powershot S40