It's been a while but today it's time for a little celebration... My previous post was about my new site not being indexed by Google, well this had improved slightly over the last months. A search for my example 'nakamichi 700zxe' still didn't result in any usable results (less then 10 pages or more away) until TODAY!
I checked again and it now gives the #1 result as a hit to my website's page!

My latest project vintagecassette.com is now live for about 2 months. Good search engine rankings are essential for sites such as mine, it needs traffic to produce profit.
A few weeks ago I implemented Google sitemap to get better results in Google, only the previous html page was listed in Google even though it was removed more then a month ago.. Now after a few weeks of the Google bot visiting daily, parsing my sitemap and indexing the site it surprises me that NONE of the spidered content can be found while searching Google! The crazy thing is that sites offering dynamic (ebay) content can be found when searching for a specific product of a brand! Also strange is that Google claims you need links to get higher rankings, these sites are not linked to anything and my site is now liked to some others including some wikipedia pages but still no improvement..
While watching log files I saw traffic comming from msn.com! It seems that Microsoft is actually doing a better job with MSN then Google which seems to be too much busy with making money of it's sponsored links and other activitied then actually providing realistic search results! MSN lists my site ad the #1 for specified queries as 'nakamichi 700zxe'
Go Microsoft! (Did I write that?)

Seeing our holiday plans change drastically we decided to go somewhere... Being spammed with the Amsterdam trams driving around with a promotion by Arke.nl to go to the Dominican Republic for a price starting 399 euro me and my (Dominican) girlfriend were thrilled!
While looking on their website everything changed (once again)... To start the 399 is for a less popular destination, 599 is for a prime destination, ok, we take the 399 anyway... clicking through to the destination leaves you with a list of dates and prices that are for 1 person 2.5 times (!) as much as 399 euro and for 2 persons 2 times as much! This is absolutely insane and makes you wonder why they dont list this as 'starting 1 euro', because everything starts somewhere usually. Imagine you go to a supermarket and see the prices next to a product as 'starting 1 cent' and when you go to the cashier to pay you get the actual price, would you accept this?

To me these people are pure criminals.

For about 2 years I was a member of the MMBase developers list and today I decided to quit. Compared to the Apache.org lists i'm still subscribed to this list was a joke. Basically the whole community consists of a handful developers that have been on this project since it started and are not willing to let new people in their world. My main frustration was that the Lucenemodule I wrote that was not part of their project and is hosted on Sourceforge was copied into a 'new' project by one of the committors... In real opensource projects people help out each other and join forces but not here. Recently after the latest release of  my version the committors of the clone seemed to have simply copied one of the new features I added to my module to their codebase! Last week they proposed a vote to add the clone module to be part of the 'contributions' of MMBase which was not questioned by any of the voters and basically that was it for me....

I've released the 1.7.9 version of my pet-project the MMBase Lucenemodule and these are the main changes:

New Features:

1. Extractor plugins
   Lucenemodule's dependencies on pdf and word content extraction are now handled by Extractors. This makes the module less depending on some external jars. Extractors have to implement an interface and have to be declared in the definition xml file. Extractors are executed by their mimetype by the Lucenemodule. You can register the same extractor with different mimetypes if you wish. The previously used dependencies for word and pdf are now included automatically as the default and can be overruled by defining a extractor with the same mimetype.
2. Support for Excel and RTF
   These formats are now supported as Extractors.
3. StandardCleaningAnalyzers
   This Analyzer which is available for Dutch and English are useful in most cases where you want to search and index word with special characters.
   Words like 'één' can then be found in their original form 'één' and base form 'een'.

Improvements:

1. Cache cleanup
   MMBase cache is now cleaned when indexing of nodes takes place to keep the memory usage low, the behaviour of the 1.8 HugeNodeListIterator was simulated here.
2. SearchTag custom queries
   The custom queries in the search tag are not supported anymore, the taglibs match tag should be used.
3. re-use-index
   This option in the module xml file will keep the index in tact and only does updates on the index on a restart of MMBase. Use this option in combination with a large interval time to keep the module from re-indexing.
4. Usage of Lucene 1.9
   The new version of Lucene is used and the module now uses the new Lucene 2.0 api.

Bugfixes:

1. Various bugs were fixed with the help of Alban Hertroys and others.