User blog comment:Kirkburn/How images are picked to represent articles/@comment-22460-20130626185703

We've been looking at this a lot on Memory Alpha recently and have some serious issues with many of the issues chosen for the myWikia app (which seems to use the same logic as described above for categories/etc). The best example on Memory Alpha is the "USS Vengeance" which shows an image collage of several models (used three times on the wiki) rather than the main image at the top of the article that is only used twice on the wiki.

In my discussions with Archduk3, one of the best ideas that we came up with was some sort of "tag" that could be applied to an image to indicate that it was the major/prominent image for the article. This would allow us (for example) to simply add this tag to our template that serves up the main image at the top of the page.

Now, the other issue there is... if the imageserver tool ignores templates entirely (which is also possible), then there's a serious problem, as almost all of our wiki's primary article images are served up within a template.

Edit: I guess the followup to the template bit is... how does the tool "scrape" the articles looking for images? Is it scraping the source text or the output text (post-template parsing/etc)?