Closed corysimmons closed 2 years ago
Hey @corysimmons thanks for opening an issue
Definitely I think there's room for improvement you raised a valid point here 😊. The main concern I see is that we're all biased based on our own use case and context. Which means that the "basic" emoji intro set can be completely different for everyone.
Which can lead to more complexity in the end for someone that is starting 😅
Yup, it'd be up to you and your community which ones constituted a "basic" set. I've worked at a bunch of places that used Angular's commit types, but I'm unsure if that's as common as I think. https://gist.github.com/brianclements/841ea7bffdb01346392c#type
Hey, @carloscuesta, by any chance, could we add to gitmoji-cli a quick report on which emoji is being used just because metrics are awesome? ~Like a opt-in cuz ppl are very protective about telemetrics for no good reason (imo haha)~ Maybe with this kinda of metric, we could more easily determine which ones should be more heavily considered to be included in the "basic" set, or even a frequent, or even if we end-up doing this, maybe we could even add a "Popular" sorting to the site. In fact, we could (also or instead) add this kind of metric gathering to the site as well, like which emojis are being copied. If you give an ok to this idea, I can probably create a simple system in firebase for we to track this stuff. I love the idea :D
EDIT: (actually, instead of doing this simple system myself, I guess we could very well just use something like amplitude to get the metrics from the website at least, donno how we'd do it for the cli, but then again, maybe not necessary, but certainly interesting to see the information)
I like the idea of an opt-in telemetry prompt (explaining the exact reason for the collection). I don't like the website analytics idea because 1) 99% of devs have adblock, 2) a ton of devs will only use the website to look-up emojis they don't know (so your metrics would tell you that some obscure emoji is getting clicked on a lot).
Hey! Thanks for sharing your input 😊
I've considered many times building a simple telemetry system for gitmoji-cli to have a clear vision about the usage of each gitmoji. However I think it's a very sensible and conflictive topic.
I've been postponing this decision over the time and as of today I still think that I would like to not include any telemetry into gitmoji-cli
even if it's optional.
Why? Mainly because I would not feel comfortable about being asked to be tracked in a cli
where you write commits that can contain sensitive and confidential information. I would feel a certain level of insecurity as soon as the question was asked.
Plus if most of the people is using an Adblocker, would you think they would optionally opt-in into a feature that tracks the way they use emojis? I don't think so, I would never enable a telemetry like that and I always reject cli
tracking myself (if possible).
An alternative solution that I can think of is "tracking" the amount of times a gitmoji is copied into the clipboard in the website. It's true that adblockers can block this request, but we can avoid that easily just by sending the track server-side using a servless function in Next.js.
Not sure if it will be worth the time but can be a fun project 😊
Plus if most of the people is using an Adblocker, would you think they would optionally opt-in into a feature that tracks the way they use emojis? I don't think so, I would never enable a telemetry like that and I always reject cli tracking myself (if possible).
An alternative solution that I can think of is "tracking" the amount of times a gitmoji is copied into the clipboard in the website. It's true that adblockers can block this request, but we can avoid that easily just by sending the track server-side using a servless function in Next.js.
I think my point against this, again, is that likely most people will be going to the website to copy/paste obscure/infrequently-used emojis.
tbh this isn't a huge deal to me anymore. It was just one of those "It'd be cool to separate the site into common emojis (like from Angular) and weird emojis". Outside of something like that it seems like any solution is overkill. 🤷
what if we made an script that crawls the git log of people searching for the emoji count on the messages and ask people to run it and submit a json to us ahah, or we run on public repos :shrug: public info is public info, I guess, the only problem with that last approach would be we find repos that use emojis but not gitmojis.
Let's close this then! 😊
Hello @carloscuesta :sunglasses:!
I think the idea of Gitmoji is great, but learning what all the emojis represent is pretty overwhelming at first. Do you think it might be a good idea to simply separate the website into having like 7 emojis for the basic/intro set, that cover almost all use cases, and the rest below it in a different section marked as "Extended Gitmojis" or something?