2019/Düsseldorf/urls

''' URLs How? ''' was a session at IndieWebCamp Düsseldorf 2019.

Watch:

Notes archived from: https://etherpad.indieweb.org/urls

IndieWebCamp Düsseldorf 2019 Session: URLs How? When: 2019-05-11 11:30

Structuring URLs, future proof, aggregation,

Participants

 * Sonniesedge.net
 * Webgefrickel.de
 * undefined
 * DanielPietzsch.com
 * – remote
 * remote
 * DanielPietzsch.com
 * – remote
 * remote

Background
URL design

Examples
GitHub's blog post URLs are human readable, but not memorable, e.g.
 * (Charlie showed one that she had bookmarked)

Hackability

 * /things
 * /things/ID
 * /things/ID/otherthings
 * /things/ID/otherthings/ID

E.g.
 * Flickr: https://www.flickr.com/photos/tollwerk/45698290631/in/album-72157703142292825/
 * shows whos photos, photo ID, but also in which album the view is

more notes

 * feels like there is a difference between elements in a path and things in the query string
 * things in the path feel like you have to keep available always
 * things in the query string could be temporal / ephemeral

uppercase vs lowercase
 * ’s URL is case-sensitive, because it uses a path.
 * Charlie’s URLs are case-insensitive. /Bookmarks will rewrite to /bookmarks.

URLs as part of the UI

 * How much control does a CMS give you over the URLs?
 * We are all designers of the experience, we should be designing URLs.

How people use URLs

 * People put URLs in the Google Search field, and expect the first hit to be the thing they want to visit.
 * So do people actually “hack” URLs? Do they understand them?
 * Do people even see URLs? Mobile browsers will only show the domain, some browsers (Safari) have started doing this on the desktop.
 * Omnibar may have made a difference here too: previously browsers had separate fields for URLs and search. Writing a search term in the URL bar would not have worked, automatically informing people more about URLs.

Migrating URLs

 * Charlie kept all the URLs when going from Craft CMS to Drupal
 * If IDs are in the URL, and the IDs can stay the same, it is often easy to change non-ID parts of the URL (e.g. categories) and still have the CMS redirect automatically.
 * On the slug is not part of the unique URL for a post, so the slug can changed or removed.
 * His URLs are also hackable, drop parts from the end and you go back to posts that day, month, year.
 * Posts are stored as YEAR/ORDINAL/ID.
 * 5eb.nl/4zt2 – number of days since epoch, encoded in newbase60: 4zt, and the ID of the post that day: 2.