We don't need all of the SE ones (yet), but given the nature of engine specific expertise and packages, filtering on those things is going to be a very common desire. If nothing else as an end user it is _very_ useful to be able to limit searches to a tag of the package or type of thing I'm looking for when hunting for previous questions, and all of these will fit into that usage pattern.

Here is a list to get us started. Feel free to edit this to add...

## TeX Engines

* pdftex
* luatex
* xetex
* context
* latex3

## Bundles of stuff that are their own ecosystem

* koma
* texlive
* ctan

## Common packages

* tikz
* fancyhdr
* graphicx
* biblatex

## Areas of expertise

* math
* fonts
* figures
* classes
* tables
* beamer

## Build tools

* arara
* latexmk
* pandoc

Right now the way to add tags is to give me a list[^1].

We will of course at some point cater for user-added tags, but in the meantime as a community TeX needs to decide how it will be using tags:

1. Just a few, carefully chosen, tags that cover the areas people will want to filter questions by
1. As many as tex.se (about 2000)
1. Something in between

My advice is to focus on the *reason* for tagging. Things like "I want to filter questions by X" and "We don't want every question title to contain Y or Z" seem to me to be good reasons. On the other hand "Tags help Google" seems to be to be very tenuous, and a lot of effort for the community to maintain.

You can of course change your mind later on whatever you decide now.

[^1]: TopAnswers has the concept of 'implied' tags. Each tag can 'imply' a maximum of one other tag.