Lowercase all tags

[dd32]

It's been pointed out that some plugins have uppercase tags (ABC) or proper cased (Name).

The majority (82%) of tags are however lowercase: abc, name.

Given these are tags, they should be consistently lower-case IMHO.

[dd32]

Additionally, we should consider whether a tag of register widgets for use on your site is a valid tag, I don't think it is, I personally think we should limit tags to maximum of 2 words.

Looking at existing tags, for a data-driven decision..

Word count	# of tags	Percent of tags (running count)	Largest Install count plugin
1 word	265,398	69.34% (69.34%)	5m+
2 words	87,393	22.83% (92.18%)	5m+
3 words	23,460	6.13% (98.31%)	5m+
4 words	4,685	1.22% (99.53%)	800k+
5 words	1,180	0.31% (99.84%)	100k+
6 words	381	0.1% (99.94%)	40k+
7 words	115	0.03% (99.97%)	30k+
8 words	43	0.01% (99.98%)	40k+
9 words	22	0.01% (99.99%)	300+
10+ words	49	0.01% (100%)	70+

Based on the data, anything 4+ words in a tag is the edge-case in length.

But the install count suggests that 5 words in a tag is reasonable, which seems weird.
Here's one of the affected tags: ​https://wordpress.org/plugins/tags/contact-form-7-invisible-recaptcha/ - IMHO that's not a "tag" that's two tags: contact-form-7 and invisible-recaptcha.

Example of a 4-word tag: ​https://wordpress.org/plugins/tags/contact-form-7-db/ ​https://wordpress.org/plugins/tags/increase-file-size-limit/

I suspect the reason we're seeing lots of plugins with multiple-word tags is SEO of the index, as we include tags as a search keyword, and we limit plugins to 5 tags.

[anonymized_14808221]

I’ve never even considered a tag could be more than one word, but that probably stems from:

• html (or else) tag is a one-word thing
• (hash) tags on social media won’t work with spaces
• “a” tag for me is a single thing - even if I’ve to tag a post as “computer programming” I’ll add a dash just because it feels more “tag-like”
• for some reason (probably again due to above first two points) they are all lower case in my mind

So I can’t but agree tags should be single worded lowercase search tags, not long tail SEO honeypots.

[knutsp]

Tags are words of natural language, and should be allowed to be capatialized as such, as from the input.

I can see the need for a quite strict word limit, even length, and only one or two uppercase letter(s) per word.

As a programmer I can sometimes view tags as similar to array keys, but only as long as I am in that mode.

[dd32]

Replying to dd32:

Looking at existing tags, for a data-driven decision..

I was asked for the data-source, and after talking through it, realised I queried data that might not be representative of the plugin directory today.

The data I queried was all tags, where as lots of closed plugins would've had more spammy tags.

Here's that data again, but this time, only for currently published plugins where the tag is used by more than 1 plugin (because we hide tags used only by a singular plugin)

Word count	# of tags	Percent of tags (running count)	Largest Install count plugin
1	157,142	74.5% (74.5%)	10m+
2	44,511	21.1% (95.6%)	10m+
3	8,236	3.9% (99.5%)	6m+
4	875	0.41% (99.92%)	600k+
5	138	0.07% (99.98%)	100k+
6	34	0.02% (100%)	30k+
7	4	0% (100%)	700+
8	2	0% (100%)	60+

It's not a huge difference, but does confirm that once you pass 2 words you're into the long-tail, and that 3 words is 99.5% of tags (where as before that was 4 words was 99.5%).

[JavierCasares]

+1 to lowercase everything (tags are used for search / filter, so there is no need to have them in other way).

And, about the use of more than 1 word, I use for example “site health”. I understand that maybe we can replace “ ” with “-” (so it may be “site-health” to be consistent.