Now we can do this also in Wikipedia. I wrote a Perl-script which scan the dumps of a language and sort the title. An other script get via API the first paragraph and the first image of all articles of one page.
Looks like there was a slight duplication of efforts. http://toolserver.org/~dschwen/synopsis/?l=en&t=Synopsis I developed the synopsis script on the toolserver for the WikiMiniAtlas, where it allows a quick preview of the articles on the map. I found the task to be not entirely trivial. At first I tried fetching the raw wikitext and stripping the markup. However Templates (some Wikipedias use templates to insert population numbers!!), Comments, References, Links make this tedious. If you want to retain basic formatting such as Bold/Italic it becomes a near impossible task. So I switched to fetching action=rendered and using PHP:DOMDocument to extract the first paragraph (Minus tables and minus short paragraph elements that contain coordinates and removing internal links to the reference section etc.). Works quite well.