Some suggestions

[igkiou]

Hi,

I'd like to make a few suggestions for the template. Happy to make a pull request (or separate ones per each suggestion) if there is general agreement for the changes.

1. Configure cleveref so that it uses abbreviations consistently–either always or never abbreviate mid-sentence references. Right now, cleveref abbreviates "Sec." and "Tab.", but it doesn't abbreviate "Figure", "Algorithm", or "Equation". (References at the start of a sentence are already always unabbreviated, as they should be.)

2. Include microtype in the packages in cvpr.sty, which helps improve kerning and spacing. Personally I prefer to use it with the default settings, but opinions differ.

3. Include the "inline" option when loading enumitem in cvpr.sty. This makes it possible to use the enumerate* environment to easily and consistently produce inline enumerated lists, e.g., "We achieve this by 1. doing this; and 2. doing that."

4. Configure hyperref to use cvprblue for all links, by changing the current option citecolor=cvprblue to allcolors=cvprblue. Under the current settings, we end up with different colors for citations versus references to sections, equations, etc., versus URLs. The resulting rainbow of colors throughout the text ends up being very distracting.

5. Configure hyperref to automatically produce correct PDF metadata in the camera-ready PDF. This is already implemented in, e.g., the SIGGRAPH LaTeX template, meaning that final PDFs for SIGGRAPH papers automatically have correct titles, authors, etc.

[taiya]

1. this is already controllable from the user via \Cref vs. \cref, no?
2. problem is microtype adds a fair bit of overhead to compile. I have used it consistently, but usually enable it at the very very end, as it allows you to squeeze a bit of extra space. We could add a commented-out line, or to the camera ready version, but I am hesitant setting it to default is a good idea; @cr333 ?
3. seems like a good idea; send a pull request?
4. as far as I remember, sections vs. citations always used a different color. Do I remember incorrectly? This was from a CVPR 2020 paper, for example: https://capture.dropbox.com/0DmIXJ4picOjDTOz
5. if the metadata is wrong, this seems like a great suggestion. Do you want to create a separate issue, and perhaps a PR with a proposed fix? Seems you know about this more than we do.

p.s. this entire post seemed a better fit for the discussion tab :)