There is generally a problem: if you want to print your cards (professionally), you have to test the given colours (of banners f.i.) and the dimensions of the card (how many pixel width is best) etc.. So (for me) it is the best to have a special template which is optimal for a professional printer shop.
If you only want to present your ideas here, a card image generator i quick and useful.
I think the intention of the card generator was not to produce printable cards but quick mock-ups. And that it solves fine.
For printing, there are at first the differences of text and number positioning in every browser and operating system and of course it operates in RGB coloring. A template in CMYK will always produce better results with printers. You can also see, that the colors in the online implementation are slightly different from the printed cards, as they are optimized for screen color compositions instead of printed physical color pigments.
Back to the fan-card-wiki:At first: Great idea, but I am also not sure if the wiki idea is suited best (but I have no experience in building wikis).
For a custom implementation, I would say the most important features are searchability and filtering for different types, attack types, authors, fan expansions, cost, and so on.
The card image could be uploaded as a file. So every one can still use the template of their choice.
Integration of the generator would be nice, to reduce double entry. The difference here is, that the generator is currently implemented only in javascript and running locally in the browser. A fan-card-wiki needs some server with a database to store all the information.
Additional features could be comments and versioning of the cards. Or simply have referenceable URLs and keep the discussion here as it is.
Yeah. And a custom implementation still has to solve the login and potentially spam issues...
So much for my thoughts. This got longer than expected.
Edit: Removed some typos. Shouldn't write such long texts on my phone...