News/Changes
Follow JMDB development progress on Twitter!
2010-05-14
Again some updates on the development of JMDB v1.40
As you can see above I added a link to Twitter where I post updates on the development progress. As those are only small changes there are several tweets per week while I update the news section several times a year. To know exactly what happens you should follow the Twitter feed jmdbdev (just click the link above).
Now we can look at the JMDB v1.40 pre2 progress:
- Added support for sound-mix.list.
- Added support for soundtracks.list.
- Fixed bug in biographies.list that prevented line-breaks.
- Changed import of goofs.list and trivia.list to remove blank lines.
- Changed output for listings so a bullet character is used instead a "-" or "*". This improved the readability.
- Changed output for genre and keywords into comma separated list of words.
- Reduced memory usage on import (plots, quotes, biographies, movie-links).
- Changed internal database table structure for:
- production-companies
- color-info
- technical
- language
- distributors
- locations
- Added experimental support to display images (the filename schema will change) of movies/persons as you can see below. The images need to be stored in the subdirectory "imagerepository" of the JMDB installation (Supported formats: JPG, PNG and GIF).
- Started to change the order of items shown in the detail window (only movies).
- Some other minor changes...
Screenshot of the "experimental" image display function (JMDB development version 2010-04-20)
As most work has been completed, only the MySQL fallback mode plus some bugfixing need to be done before JMDB v1.40 pre-release 2 will go into public testing. The feedback on the current development version is very good.
2010-04-01
No april fools' joke...JMDB v1.40 pre-release 1 is coming today... has been released!
Well I guess you can't believe it but it's true. I've updated the download section with the all the changes of the JMDB 1.40 development done in the last years.
There are still a few things missing:
- The fallback mode for full text search on MySQL if your search word is to short (≤3) is not yet included.
- Support for the soundtracks.list(.gz) is not yet in included.
- One new problem brought to my attention two weeks ago and which also effects MySQL
(but only if you select 'MySQL >3.23.28' or
'MySQL ≥4.0.0' from the drop-down list in the JMDB-Setup dialog).
The problem is an incredible long title of a series from Japan. As the older MySQL versions do not support that long VARCHARs and JMDB ignores the Exception thrown the internal movieid is wrong. So every data lookup starting at a special point (after 1/3 of the movies) returns the wrong data, as the wrong row is fetched from the database.
A solution is on it's way but as most users moved to MySQL 5.x it's only a minor issue. The next version will include a fix...and this time it won't take that long to get the release out of the door.
There is something else I want to let know. Since a month or so I use Twitter for the
development updates as it's easier to throw out some small bits of progress on the development
this way.
If you want to follow use this link: Java Movie Database (JMDB) twitter development feed...
2010-02-15
Update on the development progress
Finally a new life sign... :)
after a bad start into the new year with a failing harddrive (bad blocks, so
I couldn't start the OS) that slowed me down together with one 66h work week in mid January,
I was finally able to complete some of the work.
The best news is that on Wednesday last week JMDB v1.40 pre1 has been send to some beta users. After I finally have my E-Mail application reinstalled after the harddrive failure, some more testers will get access to the new version shortly. When I don't get seriously bugreports this version will be available shortly to everyone (7-10 days). The only thing I'm going to update are the MovieCollection example files, as pre1 currently contains outdated files.
The enhancements and new features added since December 2009:
- Around Christmas I added the full text search for MySQL and PostgreSQL. At first I
wasn't sure how good the MySQL implementation would work out, because the text
search feature only works with MyISAM but not InnoDB which normally is used and
also supportes transactions. The speed improvement for MySQL and PostgreSQL is
very good but will still improve when I finally complete my backend changes
(connection pooling,...). In case you don't remember the speed improvement for
the plain search here are the old results using (I)LIKE:
- Actor (name) search: 2800-times faster (5.6 sec down to 2.5 msec)
- Movie (title) search: 1000-times faster (15.0 sec down to 15.3 msec)
- The SQL search tab is currently put on hold because of some Java Swing problems I wasn't able to solve yet.
- I cleaned some code parts and added better error handling at several parts of the code
- For other enhancements I made since JMDB v1.36 look in the news of 2009! Most are speed enhancements (import and with the text search also query speed improvments) and some are additional checks for the IMDb errors that showed up a lot last year.
2010-01-02
Happy new year!
