Suggestions for pasted image credits caption

commonsmachinery / lo-paste-with-credit

GNU General Public License v2.0

1 stars 2 forks source link

Suggestions for pasted image credits caption #9

Open BjornW opened 10 years ago

BjornW commented 10 years ago

Using the 'Zoomed in Kitten' from the MediaGoblin lab environment. Pasted into Writer I get the following image:

screenshot from 2014-03-13 16 18 38

I'd suggest to change this into this: screenshot from 2014-03-13 16 36 30

removed full url (hide in non-printed document, show in print version) with the title of the image
changed the layout a bit to make it more readable(?)
Instead of 'Sources:' Changed this to 'By' and 'Derived from' to indicate the relationship between the two sources.
Added 'Licensed:' to indicate the CC BY-SA 2.0 Unported is a license (in printing the document we might need to use the full url written out so people can look up the license or add the preamble to a footnote?

I'm not sure if this would be the 'final' version, but I hope it makes the caption & credits a bit more clear to the laymen.

petli commented 10 years ago

The missing title is because the title is missing in the metadata markup - I forgot that when creating the image file. So libcredit falls back on the best thing it finds, which in this case is the plain URL. If you copy the other kitten image you'll get a title.

BjornW commented 10 years ago

Ok, but we can make an educated guess from the url or perhaps the filename?

petli commented 10 years ago

Re: Licensed. We thought that people will (hopefully) see so many credit statements that they want them to be short and will recognise the CC license.

We had "Licensed" when we only know the license and nothing else, but changed that to "Terms of use". That would be a better option if we want to add that to a full statement too.

petli commented 10 years ago

I'm wary about trying to find a title out of the link. For some sites we might be able to, but since we'd need that to be site-specific then we can instead find a proper title in the web page.

For the catalog we'll move the metadata extraction code into a library of its own, and that could have a heuristic to e.g. use the title of the web page. But it might be better not to include a title at all in that case.

BjornW commented 10 years ago

Yeah, it's not an easy thing to do technically especially if the website is not using the correct semantic markup, but it would make the plugin a lot more usable and since the caption is easy to change in Libre Office it might be ok to make an educated guess in this particular case.

petli commented 10 years ago

It's not a matter for the LO plugin though, but libcredit and the (forthcoming, right now ffx addon) libgetmetadata. And in the meantime, add an issue over here to discuss it in the right place :) https://github.com/commonsmachinery/credit-metadata-best-practices