Localization of images' content

[skorasaurus]

Several pages in other languages have images with text written in English.

For example, the images for JOSM sections - JOSM Plugins in French

A quick review of this appears that some of this has been done for the German section - http://learnosm.org/de/beginner/start-josm/

Second related question and probably warrants a separate issue; having a consistent naming scheme for the images which are stored at: https://github.com/hotosm/learnosm/tree/gh-pages/images
The German images look like they need to be renamed to the scheme used by other languages:

So, the 2nd image used in chapter 3 of beginning section for Dutch is:
nl_beg_ch3_image02.png

This is a pretty big task and probably isn't a priority at the moment (at least not for us in Haiti at the moment). We can use this for discussion and breaking the steps for this.

• Standardizing image naming schema
• Taking the screenshots
• uploading them to here
• making the changes in JOSM.

[MappingKat]

+1

[nebulon42]

There is another problem: images can look quite different in other languages and if there are changes they have to be recreated from scratch. We also need to have the source files with layers in the repository.

For example, if the source file would be available to translate an image would mean to add another text layer in a different language and probably adding another background layer with UI in this language. Others than the original image author then also could easily edit the image.

It would be possible to use Inkscape as image compositor (I don't call it editor here as raster image editing capabilities are limited, they would have to be cutted etc. with Gimp/Photoshop etc.). Inkscape is open source, runs on many different platforms and produces SVG. It is also able to link or embed raster images, which would be perfect for this purpose. Then different text layers could be added. The result can be exported as PNG.

@jeffhaack: This could be relevant for your work on updating the guides or in other words: what do you think? :)

[jeffhaack]

I generally use Photoshop and I agree that we need the source images here too for translations. I've started putting any new psd files I create in here. My suggestion is that anybody who creates an image with text overlay like this should add the source file alongside the PNG.

Regarding standardization of image names, I've followed a few threads where this has been discussed. Yet the image names are all following many different suggestions. Personally I don't think it matters too much what the images are named, so long as they are grouped in subfolders according to the section that they are in. So I am thinking about an image directory structure more like this:

images/beginner
images/beginner/ch1-introduction
images/beginner/ch1-introduction/en_any_name01.png
images/beginner/ch1-introduction/fr_any_name01.png

and so on.

Thoughts from @severinmenard, @MappingKat, @skorasaurus ?

[danbjoseph]

Anyone volunteering their time to create/update images should probably use whatever software they have and are most comfortable using. However, having files in *.psd (or other formats associated with proprietary software) limits the accessibility and usability as a source file.

they would have to be cutted etc. with Gimp/Photoshop etc.

Are you referring to cropping? It can be a little more indirect than Photoshop, but in Inkscape you can use a clipping mask. What raster editing options do you use that Inkscape lacks?

If the image for a new translation requires a new screenshot (example, the website being shown has a non-English interface) then the person creating the translation wouldn't save time/effort by using an english source file image. Source files would be helpful for images like the mouse navigation image.

I think division by language should be the highest level of image organization. See pull request #134.

[nebulon42]

Didn't know about the possibility of clipping masks, thanks for the info. Apart from clipping - I'm quite sure - there is not much than Inkscape can do with a raster image. But editing raster images might not be needed for LearnOSM anyway I admit. Of course there is not much won if source files of unaltered screenshots are created. That is not necessary. For the mouse image such a file, as you said, is a good idea.

But there is another type of images suitable for source files: images that contain screenshots as background layer and another text layer (as in chapter OpenStreetMap.org). Here the screenshot and text layer would have to be replaced by another language version, but the text information would stay in the same format (font, size, position). Maybe a single file with all languages in it would get too complex so division by language is a reasonable approach. Since English is the default language source files should be copied from there as it is already done with the text.

[Nick-Tallguy]

I'd like to close this issue now unless there are any objections.

A system for filing the images has been started & is working well. Uploading images by those not familiar with the system is still slightly problematical - at the moment I am getting them emailed to me, which is working well.

Translation - am putting numbers into images where I can, but some will need a new screenshot in the target language.

[Nick-Tallguy]

Closing - see #526 and commit on staging site. - working on other issue now