I’m increasingly struggling with picture descriptions on Space - I think this happens in particular in blog posts on Discuss (example), but it’s also true for email (i.e. “Upcoming Wikimedia events for August 2019”). I think they need to be better formatted so that they don’t look like actual content. I wonder if there’s any config you can fix on your side that makes the text smaller and/or in italics and always places the captions under the picture.
Thanks for this. Yes, the current way we’re doing the embedding of images on the blog causes this, and as far as I know with the current process it can only be fixed manually from the Discuss end. @CKoerner_WMF, do you think there’s a way we can effectively edit the image caption when embedding the image from Commons? Or even have it not pull the caption and have authors manually add the attribution in the caption, as was happening before?
What is the thing that takes the WordPress post and makes it appear in Discourse?
Commons ‘embeds’ in WordPress are
<figure> elements with
<figcaption>s for the captions/license bit, but by the time they get into Discourse they’re just plain paragraphs with no distinguishing classes or anything, so it doesn’t look like there’s any way to style them. I think any fix might belong in the system that does the moving.
This is managed by a WordPress plugin by Discourse.
This allows authors to check a box when publishing a post a new thread is created with the content of the blog post. This is done via the Discourse API. The authorship is carried over via username matching.
Thanks for the link. I guess it makes sense that the structure is lost, because (I presume) the HTML of the WordPress post is being turned into Markdown for here. There isn’t any support for image captions in Discourse in general is there? It seems like the closest it can come is to use the
<img> syntax instead and set the
title attribute, and that’s hardly as good.
One thing that might help is to fix the caption display on WordPress so that it shows the filename, license, author, etc. all in one element instead of often split as it is now (with the language-specific caption in a separate paragraph). This would look nicer on both WordPress and Discourse I think, but doesn’t totally get away from the fact that the caption in Discourse would look like just another paragraph following an image.