The result of the latest works on the OPT source code is complete opt.class.php file which contains (as some of the readers may know) the main parser class. I must admit that this file was more demanding that the compiler. In the second case we can afford with quite a lot. Just an idea for the algorithm and prediction if and how it will influence the rest of the code. In the case of the main class, we must pay attention to the performance. It must contain everything that is run permanently and it must be fast and small. But the number of features is getting bigger and bigger.

The website now has the "Download" page. In it, you can find a new development version of Open Power Template. It also contains a simple script "docgen" which we use for automated generation of the DocBook manuals. The new development versions should appear now much more often than before. In case of many new features, they might appear even every 2-3 days. The reason why this was impossible in the past was the time. The process was not automated and it takes a while to prepare such archive (for example - it takes a couple of hours to publish new OPT 1.x version on the Internet).

Unsually, the final HTML code of the website is composed of several smaller templates containing smaller pieces of it. There are many techniques of combining them. In PHP template engines, the most popular are two forms: manual execution of the templates from the script side and the include directive. Similar solution can be found in OPT 1.x, however in the new version, it should be forgotten - unless we use the quirks mode, the parser does not let us to call the parse() method more than once. In the beginning, I planned to replace it by extending inclusion with "sequences" - a groups of templates to include with their own variables and cache settings. However, one day I was asked on my blog by Coldpeer, whether OPT will support template inheritance. I googled for a while, I looked at it and thought it might be cool. That's how the implementation began.