Re: Replacing the animal photos in the ‘erase’ activities

Timothée Giet animtim at gmail.com
Fri Feb 25 10:15:12 GMT 2022


Le 25/02/2022 à 08:41, Karl Ove Hufthammer a écrit :
> Timothée Giet skreiv 25.02.2022 00:07:
>> I agree with most criterias of selection you gave, though I would add
>> one more requirement: the images should still look okay (after being
>> cropped) when displayed on a phone in vertical layout (so the image
>> should be cropped in a way that the horizontal center of the image
>> contains a meaningful part of it, like the head of the animal for
>> example).
>
> Is it possible to set the horizontal alignment of the image? Then we
> can define a value for each image which defines the horizontal
> position (in %) of the most important part of the image. If the head
> of the animal is at 75% (i.e., ¾ of the distance from left to right),
> this value will be .75. When the aspect ratio is < 16:9, cropping will
> start from the left of the image and continue so that the 75% part of
> the image is always visible.
>
> I don’t see native support for this in QML, but there is a
> ‘horizontalAlignment’ property, with values Image.AlignLeft,
> Image.AlignRight and Image.AlignHCenter. This might be enough (the
> interesting part is usually either in the centre, left or right of the
> image). Or perhaps it’s possible to use a hack with a manually
> positioned (outer, clipped) box to achieve a similar effect?

I just made a quick test, and dynamically changing the vertical and
horizontal alignment of the images would work (with one of the 3
possible values for each), and would probably be enough for acceptable
results.

This makes it easier, we'll just need to provide the corresponding
alignment values with the images in the dataset.


>
>
>> One more precision about acceptable licenses: for CC-BY-SA, only
>> CC-BY-SA version 4.0 is compatible with GPLv3 license, so any image
>> using earlier CC-BY-SA version should not be used... Also some images in
>> your list are under the GNU Free Documentation License, which is meant
>> for documentation only and is not compatible to mix directly with GPL in
>> a combined work...
>
> We can ask the authors if they are willing to relicense. After all, a
> person who chose to license their image under CC-BY-SA 2.0 probably
> didn‘t do it because they were against CC-BY-SA 4.0, but because the
> CC-BY-SA 2.0 was the current license at the time. And they might even
> accept CC-BY. And I would think that most people who use GNU Free
> Documentation License probably isn’t against (relicensing to) GPLv3(+).
>
> So I would prefer if we first vote on the *original* list of images.
> If some highly rated images have an unacceptable license, I can try to
> contact the authors and ask about relicensing. If they don’t accept,
> we can either skip the image or, if it’s important that we have an
> image of that exact animal, try to find an alternative image with
> acceptable license. I have currently only looked at Wikimedia Commons,
> because the images are nicely categorised and easy to browse, but
> there are also many images with an appropriate license at Flickr. If
> we‘re looking for an image of a specific animal, it’s easy to search.
>
If you're ok to do that contact work, then fine, we can keep the
original list of images to vote on. Also, after more research, I think
it's hopefully safe to use CC-BY-SA 4.0 images, as the compatibility
with a next GPL is just not guaranteed for now until there is one, but
could happen (and we can hope it would).





More information about the GCompris-devel mailing list