Hello all
For tomorrow I have an upgrade of themes.drupal.org planned. And after taht I want to change some concepts on short basis.
What will change during/after the upgrade: The themeswitcher http://themes.drupal.org/themeswitcher will dissapear completely. It looks like crap, is not used often (acc to the logs) and it is horrible in performance (no paging, no caching, read s files from disk to generate the listing etc. We should/must look at a PHP page on Drupal.org very soon, or else automate some project.module stuff to link to themes.drupal.org. After all: selecting/finding themes is not the task of themes.drupal.org
We will have a completely skinned down Drupal on themes.drupal.org. Actually it will be a Drupal with only story.module, comment.module, a few published stories and a few comments. The rest are required (admin) modules. I will remove all other modules. This makes a horrible representation of themes/Drupal. Esp because the niceness of themes is not defined by how well they render a story (Drupal is far more then a blogging system). But by how well they handle the forums, aggregators and the admin area. How well they handle popular contribs such as images. We need to find a solutution for this!
We *really* need some process to select themes. Currently I am the only one that manually reads trough theme code and the theme CSS/HTML to see if a theme is fit for themes.drupal.org. This leads to a) a huge lag, themes may be released for months before available b) a big security hole. I have only two eyes, and simply reading code is NOT a security measure. Simply trusting/installing committed and released contribs (themes) is not an option IMO on our infrastructure. In the current pace we will get approx 10-15% of the released themes online on themes.drupal.org without endangering drupal.org too much!
Bèr