Should tags be case sensitive?

[pde]

Moving an issue to discussion here per @zachleat's request.

Currently the tags in the base blog are case sensitive, but I suspect that's not what most users would actually want. In a project I've used them for, I have a mix of tags from different human and programmatic sources, and it's not easy to ensure the human-generated ones are all lower cased.

What would people think about a config option to control this, probably defaulting to "case insensitive"?

[Ryuno-Ki]

Another option might be to convert tags to lowercase for display?

[nhoizey]

I would even recommend to sluggify tags.

All these tags would have the same etiquette slug, the same URL, and "contain" the same collection of items:

• étiquette
• Étiquette
• Etiquette
• etiquette

"étiquette" is the French word for "tag"… 😃

[Ryuno-Ki]

That would have been my suggestion once an agreement on lowercasing would appear.
Perhaps also worth looking into Levenshtein distance (as plugin) to warn for similar tags.

[nhoizey]

I guess Levenshtein distance could be useful in a linter, but similar tags should be able to co-exist.

[edwardhorsford]

I've certainly had bugs on my site where i've had duplicate tags as a result of inconsistent captialisation.

I don't know if something should force them to one or the other, but it could combine both cases so that either works.

[zachleat]

could combine both cases so that either works.

I’m not sure I understand what this means

[ItaiAxelrad]

I believe @edwardhorsford means collections would be merged if they match when case insensitive/lowercase. For example, if you have two tags posts and Posts the resulting collections would be combined into one and accessed either by collections.posts or collections.Posts?

[edwardhorsford]

Yes, what @ItaiAxelrad describes is what I was thinking

[johnridesabike]

It would be great if there was a way to normalize tags, but I think it’s also a bit more complex than this.

For example, I have a site where I would like these following tags to be “merged” into one tag:

- Atlanta, GA
- atlanta ga
- Atlanta, Georgia

...etc.

What if there was an API for normalizing tags? I can imagine a hypothetical function like this:

eleventyConfig.onCreateTag((tag) => tag.toLowerCase().trim() /* or whatever else you want to do */);

Since there’s probably no default configuration that can satisfy everyone, this leaves it in the hands of the user.

[pde]

It seems that @johnridesabike's approach would optionally support @nhoizey's desire for sluggification too. Is there an elegant way to make that approach a switchable default?

Seems there could be two versions of this, one where that line needs to be explicitly present in .eleventy.js, and another where that handler is the default but can be overridden. Are there strong views on which is better?

[cokkto]

Also remeber about non-latin characters - the issue relates to this one. Tag transform function is really demanded, otherwise urls become the mess of latin and non latin letters and whitespaces

[Brixy]

I second this as I currently have problems transferring a website with blog to Eleventy from Wordpress because of an inconsistent usage of tags.

The company uses camel/dromedary case like

tags:
  - camelCase

so the ‘original’ tag form is important here:

<h1>Tagged “{{ tag }}”</h1>

For permalink slugify should transform tags to lowercase only IMHO. (I am using Eleventy 1 beta 8.)

permalink: /tags/{{ tag | slugify }}/

Also, a customSlugify function (with toLowerCase()) does not work with tags although it does with Markdown-It.

[zachleat]

New home base for this is at #2462 and I’ll think we’ll probably land with an opt-in change per this comment #2462 (comment)