I have my hands in many parts of Drupal and for some time now I make concious effort to make sure things would go fine if I disappear or something. Not that I want to leave the Drupal project, not at all, but you never know what happens.

Not for the faint of heart but I have Drupal that boots without SQL using memcached. It's fast. Very. If you have a focused site, you can gain from running a code coverage script and moving everything to conditional includes that's not used on the focus-pages. You can easily get several hundreds of requests from one small machine.

Again I need to use my blog to answer some really bright guy who thought that the world would be poorer if his wisdom would not pollute the Drupal Planet. He is wrong: Drupal is not WordPress. We do not make compromises in security design. Any ways of updating/installing a module on the webserver which does not ask you a password does make such a compromise. Now, Drupal has an auto update but it stays secure. Joshua Rogers took on implementing that under the name of Plugin Manager as a Summer of Code project. He stayed with us and develops it. There was a call for core inclusion. A port has been made by swentel. What more do you want? I know what I want: Plugin Manager in core. Care to help?

Thanks to Hagen who uploaded Dries' talk at DrupaCamp Koln 2009 to blip.tv. (His link made the video a bit hard to find so I posted direct links)

Comment module is like the ugly stepchild in the happy family of Drupal modules. Noone wants to party with her because she is SO ugly. Traditional solution was, let's use nodes as comments. Woe to your site's performance shall you do that. Of course, sometimes functionality rquires that. So now I think we should provide that as an option in core -- choose between comments and nodes as comments. I have created a series of issues which would clean up comment module significantly and make some parts reuseable as well.

Theoretically, writing an Oracle driver for Drupal 7 is possible. As the SQLite driver proves, you have quite enough points where you can massage the input/output to make the ancient vestige that's Oracle and the leading tech that is Drupal work together. Yes there are issues which would make this easier, but I am firm it's possible. A few important questions remain: where should this driver live? If in core, are patches going to be held because they fail on Oracle? Who is going to maintain this driver?

Somebody thought it was a good idea to use the Drupal Planet for bug reports. Guess what? It's not. Please don't. There is a proper workflow for these things. Oh and guess what again? There is already an issue in the queue, Damien and Dries are both on it so what the heck more do you want? You are not helping anyone by purporting a half-solution. And I reply in a separate post 'cos if I'd write a comment noone would read that.

As we learned from Angie's excellent article it's useful to read the developer's thinking when submitting a patch. However, certain old decisions left no trace on the website. However, as I recently learned from Dries that in the beginning patches were submitted to CVS for review. So if your cvs annotate reveals a very old commit with no issue, you can look there.

Currently a mere 350 million people have broadband access. I expect this to jump a magnitude within less than ten years as smartphones and cheap laptops and 3G data makes broadband ubiquitous. We can add web services to this as another major traffic drive. They sold 13 million iPhones and Android is just starting. The storm is coming -- will your site stand or be blown away?

It's been years since we are dreaming of getting CCK into core. Dries has announced a code sprint, where six people are going to work on this for an entire week. That's 240 (wo)manhours and all we ask is a $7000 donation. This is your only chance to hire the top Drupal coders for less than a $30 hourly fee. If you have ever used CCK please consider donating the equivalence of one hour payment of your time -- it surely saved that many but likely even more.