At the moment, sharing functionalities are in the content sub menu "teilen"

Let's include:

• Google Classroom
• Facebook
• Twitter
• Mail

And may be, we should replace the sharing functionalities. E.g. like in Khan Academy (below the head line) : https://www.khanacademy.org/math/trigonometry/trig-with-general-triangles/solving-general-triangles/a/laws-of-sines-and-cosines-review

How can we support translating the strings in the plugins? Don't know the config in athene2-assets for that, but maybe this also works here?

Trashed blog posts are still visible in https://de.serlo.org/blog/

"subject home" and "all topics" are redundant in subject breadcrums. Additionally both sides show a taxonomy overview

potential problems

• at current state, "all topics" is needed in side navigation to build the subject taxonomy

Use case: when merging multiple entities, every entity except the new canonical one gets trashed. We'd like authors to configure a redirection from these trashed entities to the canonical one (so that old links still work).

Even better: merge of entities as a new action so that we have the information about that (e.g. to show the histories of all merged entities)

e.g. email contains link https://www.serlo.org/entity/view/41648 which redirects to de. instead of en.

Interestingly https://www.serlo.org/41648 works, as https://de.serlo.org/41648 tries to resolve the correct instance

To reproduce: create new article, you see the created uuid in the URL, close tab without saving the revision => We have an article without a revision. We should fix that and delete dead nodes that still exist.

On mobile the comments below exercises arent visible. The input field and button to write a comment aren't displayed either.

Show only on de.serlo.org. Also add the "Neu hier" pages of the subjects and /abc

E.g. https://de.serlo.org/entity/repository/add-revision/142516.

I assume that this is because something went wrong when converting? Not sure though. Focusing in https://demo.edtr.io works correctly.

we tried to change the URL of https://hi.serlo.org/--- to https://hi.serlo.org/जुड़-जाना but the hindi letters have been replaced by "---"

The subject hierarchy of serlo is deep. Often one sees three dots in subject bar (id="subject-nav"). My recommendation is to provide a pull-menu to see full navigation branch to subject.

This improves context awareness and thus improves didactics.

e.g. https://hi.serlo.org/-112518

Currently, when you have opened a hint on an exercise and then open the solution, you don't notice that it actually has opened. Proposed solution is to scroll to the opened element.

We already implemented this in serlo/athene2#174 but there were some bugs (see serlo/athene2#402) so we removed it in serlo/athene2#417.
Maybe take a look at the jquery scrollTo plugin: https://github.com/flesler/jquery.scrollTo

reported by @getJack

The items of your RSS feeds seem to be invalid. I tried to open your feed https://de.serlo.org/entity/api/rss/article/5/feed.rss with Firefox 46 which did not display the links of your content. I then had a look at the xml and saw that in your property the links have no valid URL but relative paths.

Please see the result for your self https://validator.w3.org/feed/check.cgi?url=https%3A%2F%2Fde.serlo.org%2Fentity%2Fapi%2Frss%2Farticle%2F5%2Ffeed.rss

If you look up the specification of RSS https://validator.w3.org/feed/docs/rss2.html#hrelementsOfLtitemgt you will see that items require a URL and a path is not a valid URL according to https://www.ietf.org/rfc/rfc3986.txt

Seems like mcrypt_decrypt is deprecated and doesn't work right now (e.g. when using the debugger).

(cc Nish): The magnification effect on images (when clicking) is not noticable for users.

Maybe change the Pointer to a magnifying glass?

When running cypress, i frequently get errors, when the requests are sent too quickly after each other.

Fatal error: Uncaught ErrorException: include(/var/www/html/src/vendor/doctrine/doctrine-module/src/DoctrineModule/ServiceFactory/AbstractDoctrineServiceFactory.php): failed to open stream: No such file or directory in /var/www/html/src/vendor/composer/ClassLoader.php on line 444

ErrorException: include(/var/www/html/src/vendor/doctrine/doctrine-module/src/DoctrineModule/ServiceFactory/AbstractDoctrineServiceFactory.php): failed to open stream: No such file or directory in /var/www/html/src/vendor/composer/ClassLoader.php on line 444

On longer articles you have to scroll all the way down to see the authors description of his changes. It would be easier to review, if this field is displayed at the top.

Currently, we are still relying on some images that are placed in public/assets (e.g. for the random banners and home). We should move these to assets.serlo.org instead.

https://de.serlo.org/datenschutz doesn't work on new system:

An exception has been thrown during the rendering of a template ("The Symfony\Component\Intl\DateFormatter\IntlDateFormatter::__construct() method's argument $locale value 'de_DE' behavior is not implemented. Only the locale "en" is supported.  Please install the "intl" extension for full localization capabilities.") in "static/de/datenschutz/partials/header" at line 23.

https://de.serlo.org/navigation/container/get/1

I tried to change "Neu hier" (under "Bei Serlo Mathematik mitarbeiten") parent from 10 to 1.

The page reloads but nothing changed.

@Knorrke suggested to always ignore trailing slashes in the urls.
We need to check first which pages that rely on that trailing slash to display different pages.
(e.g. mathe/ )

Current example
https://de.serlo.org/nachhaltigkeit/ should redirect to https://de.serlo.org/nachhaltigkeit
instead it displays "Page not found".

• Privacy banner
• Mathjax thingy in bottom left corner
• Spoiler should be expanded

Database changes:

1. Table language: ( You can find the language codes here: https://www.unicode.org/cldr/charts/latest/supplemental/language_territory_information.html)

INSERT INTO language (code, name, locale) VALUES ("ta", "Tamil", "ta_IN");

2. Table instance:

INSERT INTO instance (language_id, name,subdomain) VALUES ((SELECT id FROM language WHERE code="ta"),"Tamil", "ta");

3. roles and permissions (see below)
4. taxonomy

File changes:

1. (serlo.org/public/server): Add language to php config in Dockerfile https://github.com/serlo/serlo.org/blob/7f26740906d6e42de6058717a585df4d6da5e47f/packages/public/server/docker/php/Dockerfile#L7-L11
2. (serlo.org/public/server): add folders and files in src/lang and src/lang/routes
3. (serlo.org/public/server): add tenant branding configs to src/config/autoload/global https://github.com/serlo/serlo.org/blob/7f26740906d6e42de6058717a585df4d6da5e47f/packages/public/server/src/config/autoload/global.php#L41-L72
4. (serlo.org/public/server): Add the instance with subject config (at least /community) to src/module/Subject/config/instances.config.php
5. (serlo.org/public/cloudflare-workers): add tenant route
6. (serlo.org/public/static-assets): Add the static assets for the tenant, e.g. home/about-serlo.svg
7. (infrastructure): Add tenant specific secrets (e.g. tracking), increment build version of athene2 and add tenant to cloudflare (see https://github.com/serlo/infrastructure/pull/94)

On serlo.org (After everything is set up!)

1. Add license through backend (/license/manage)
2. Add pages for subjects through backend (/pages)
3. Add subjects through backend (/taxonomy/term/organize-all)

• It is not visible whether there are any open discussions
• For grouped-text-exercises, it is not visible if there are discussions after scrolling to the end of the opened solution.

Sometimes (especially on ChromeBooks), CloudFlare bot protection kicks in when trying to save content (showing the reCAPTCHA every time and not saving the content). That shouldn't happen, especially for signed up users.

Add tests for the following combinations:

• All content types + taxonomy term + user profile
• created with legacy editor + created with editor
• rendering + editing
• legacy stack + frontend

This is a big task. Please make sure that you’ve got enough capacities in the Team and that there are no other important priorities before starting.

Vroni (SEO consulting) said that the images’ URL are very important for SEO. Improve the image URL structure to get URLs containing the image keywords and, also, the file path where they are deposited. For example, the URL should look like this:

• https://de.serlo.org/mathe/funktionen/funktionsbegriff.png

and not like this:

• https://assets.serlo.org/legacy/56fe95af016c7_bc84b18113f67b5ff4f2443da8eee5accd5ea901.png

E.g. https://de.serlo.org/9081. Exercise c) f(x) = x is not licensed by cc-by-sa-4.0 but there's no way to see that.

This is a big task again ;). Please make sure that you’ve got enough capacities in the Team and that there are no other important priorities before starting.

Vroni (SEO consulting) said that collecting the URLs in XML-Sitemaps is very important for SEO. But once we collect them it’s imperative to update them regularly, so we should do this task just if we can make sure that we have enough capacities.

There seems to be a problem with notifications since ~28.7.

We have three blogs at serlo.org: https://de.serlo.org/blog

But we want to combine the different topics in one blog.

Can you please merge the blogs and copy the postst from https://de.serlo.org/dev-blog/ and from https://de.serlo.org/mathematik-blog/ to the main blog: https://de.serlo.org/blog/
(we can do the copying manualy, if you want?)

I also think we can reduce this unnecessary complicated structure and have just one /blog per language tenant (that was originally my fault, sorry)

Probably fixed by taking the "minimum" of datetime from database and "now", see e.g.

The images are too small (e.g. Exercise 1 at https://de.serlo.org/30560) or too big (e.g. Exercise 4 at https://de.serlo.org/58876)

The landing page is cached (and obviously should be), but for an logged out user no flashers are shown because of this. As @arekkas said: A resolution could be to have a dedicated success page.
But a solution has to adress the following different user actions:

• Sign up (old issue: serlo/athene2#477)
• Forgot Password (old issue: serlo/athene2#281)
• Account Activation (old issues: serlo/athene2#615, serlo/athene2#612 )

Considering serlo/athene2#686 and the issues when searching for courses, we should have a view for the course itself that contains the content for all pages. Links to individual course pages should instead go to the root course (w/ JavaScript only showing the relevant page; see also https://support.google.com/webmasters/answer/66353?hl=en). This should solve all the issues (indexing courses, duplicated content). The only thing we lose with this (and let's discuss if that's a problem and if we could get around that), is that you can't get directly to an individual course page from the search (you would instead get to the beginning of the course). IMHO: this is fine since you'd usually won't start in the middle of the course, anyways.

Navigation: Lehrplan vs. Themengliederung

Aus der Gliederung für einen Lehrplan kann ich auf einen Artikel klicken, der im Inhaltsverzeichnis für Themen aufgehängt ist. Beim Sprung vom Lehrplan in einen Artikel geht die Information verloren in welchem Lehrplan ich war.

So kommt man zwar leicht vom Lehrplan zu einem Artikel, aber zurück geht nichts mehr. Für den nächsten Punkt im Lehrplan muss man wieder durch den ganzen Lehrplan navigieren.

questions: how to do this with caching? What happens when following links in the article?

When the editors split e.g. articles in multiple articles, it would be very helpful, if they can get a list of all entities that link to this entity, s.th. those links can be changed accordingly.

Currently we import at least one howler plugin we don't use (sparcial audio).

This goes back to this upstream issue:
goldfire/howler.js#971

We should keep on eye on that and update the library if it's fixed.

Content created with the new editor has to be rendered w/ React for interactivity. We don't do that for injections, yet.

E.g. https://de.serlo.org/entity/repository/add-revision/142439

TypeError: "e.button is undefined"

Automate this:

• Set user id to id 4 (fake combined user for deleted users) for all relevant tables (e.g. notifications are definitely not needed)
• Delete user from user
• Review where we could delete more (e.g. which events are not needed anymore, comments, ...?). E.g. look for usages of events, e.g. trash bin

Handle:

• When rendering username, maybe handle id 4 specifically
• Profile for id 4 (e.g. no activity log / stats), info that this is a deleted user
• When user does not exist anymore, redirect profile to /4 since deleted.

Currently the Alias is only triggered, when editing. Can we trigger that too, when an Element is moved?

Things to consider:

• We should also update child aliases (course pages)
• What if an entity is copied to a different taxonomy? Maybe we don't want to trigger it then, and I think we can't even differentiate between those events.

The error-messages when during login (like "Mit dieser Kombination ist bei uns kein Benutzer registriert.") are also in other language-versions still in german. I think this text-output should also be given by Crowdin, so that the message is in the respective language.
Function that gives the output/code-line: https://github.com/serlo/serlo.org/blob/0269019e8e00d29d0cbfa3e3994da204b30f4214/packages/public/server/src/module/Authentication/src/Authentication/Adapter/UserAuthAdapter.php#L102).

I can't use ">" in taxonomy descriptions. I tried to use it here
https://de.serlo.org/mathe/deutschland/bayern/fos-bos-technisch
but it has been deleted without notification

The change to toHtml() in #16 caused problems in the legacy system (and might cause problems in other places, too):
When rendering the meta description of a course page, we go up to the parent and render that description instead. But when the parent (i.e. the course) doesn't have a description either, the Normalizer just returns the id as description (I assume, to handle the "go to parent" in course pages). Problem though: the id is a valid json (since it's an int) but can't be rendered by the editor renderers.

• In legacy, we still cover this with try-catch and just return the id as a string (after applying the quickfix)
• In kubernetes, we pass it to the renderer which will fail and therefore return an empty response

We should review whether there are other ints we try to render with toHtml that should be rendered. And probably fix meta description handling of courses since the id as a description doesn't make sense.

https://de.serlo.org/community/neue-faecher-themen/medienbildung/datensammlung-big-data-ist-big-data

This happens with every course from https://de.serlo.org/community/neue-faecher-themen/medienbildung/datensammlung-big-data

Other courses on serlo.org seem to be fine: https://de.serlo.org/61911/uebersicht

• redirection to the page you visited before the registration process after completing the registration
• you are already logged in after following the registration confirmation link in the email
• If you singed up and go to sing in it says something like "Your account hasn't been approved yet". Here we should point to the confirmation email.
• improve role descriptions (see comment below)