[Workbench] Social media optimize anonymous collections/projects/pipelines
everyone wants to share their cool datasets on social media, right? :)
This is how Google+ generates those previews:
And for Facebook, we want some 'open graph' meta tags:
Looks like Google+ supports open graph too.
#8 Updated by Radhika Chippada over 6 years ago
Discussed this with Ward this morning and we agreed that:
- Add open graph meta tags to our pages
- Add these tags for all objects (since the search engine won't be able to access private data anyways)
- http://ogp.me/ can be used to read about open graph protocol
#9 Updated by Radhika Chippada over 6 years ago
- Tested using google testing tool at https://developers.google.com/structured-data/testing-tool/ (this seems to validate only basic data)
- The Face test utility https://developers.facebook.com/tools/debug/og/object/ seems to work only with urls and hence was unable to test using localhost url
- Also, manually verified by looking at the loaded page html
- All existing automated tests passed
#11 Updated by Brett Smith over 6 years ago
1be4d28 is good to merge. I just see a couple of small opportunities to smooth things over for future readers:
- Expounding the comment about Open Graph with a sentence or two about what it is and why we include it us might be useful for people who don't recognize the name. (I'm not sure I would've known without being able to refer to the background materials you provided in this ticket. Those were very helpful, so thanks.)
- Why do the conditions rescue exceptions raised by
respond_to?is a method that Ruby defines for all objects, so it really should never raise an exception, and I don't see any other places in the Workbench code where we rescue it. These conditions might be easier to follow without those.
But go ahead and merge whether or not you make changes. Thanks.
#14 Updated by Nancy Ouyang over 6 years ago
From the sprint review a week ago: we could also consider adding either user-upload(able) "thumbnail" images to the collection/project, or perhaps try to generate one automatically (e.g. the curoverse logo, or some permutation of text on top of the logo with the title or size of the collection or something along those lines) or by default if the user does not select or upload an image.