 "fun" fact for those of you who weren't aware: **the first version of a hashtag that's used is the canonical case** (as in lower vs upper case) that means if you create a hashtag, **camelCase it!!!** on the first use!!! otherwise the autofill will forever suggest the lowercase version and people will often default to the autofill. this creates a less accessible fediverse. 73 comments
@0x00string right???? same same same i went down the same rabbit hole. The #superBowl vs #superbOwl debate too...  @0x00string my best guess at the end of the day is to query for existing casings of the same string, then compare to the canonical case and do {something} to determine which is preferable... @inherentlee if it were me, a dumbass, i would just make whichever is the most used at any given time the auto-suggest one @0x00string I did consider that! but I think the weight of the autosuggest might bias that toward lowercased versions. but use weight is definitely part of the solution 🧐 @inherentlee i would drop the no-cased one from consideration if cased ones exist, idk lol, this is why i dont host websites 😂 @0x00string ah but what if someone made a typo in a single word like tyPo? it's uncommon but not impossible. Or a break in a word that is only questionably a break - I've been running into this with "iframe". do you uppercase the "f"?? @0x00string just the existence of camelcasing can't be enough, there probably needs to be some sort of bare minimum # of uses as well When you say the word outloud do you pronounce it "Eye Frame" then yes, uppercase the Frame. If you pronounce it as If-rame, you wouldn't (not that I know anyone who pronounces that one without the "eye", but you get it, right?) @deirdrebeth @0x00string that's how I lean but my coworker has used the opposite convention! so it can be malleable, is the main point I was making Totes. I figure that the people it was originally created for use text to speech so the sound should outweigh the marketing convention :-) @0x00string @inherentlee R has a package called "janitor" that, among other things, can convert random strings in the column headers of a table to well-formed #snake_case strings. @inherentlee From what I recall it's actually server specific, so if your server saw it elsewhere first then it uses that, but if you're the first user it seems then it uses your version. Because *waves hands* I have no idea. @SRLevine well it'll be "first use in the circle of servers your server federates with" technically, so, yes @inherentlee So it queries other servers for the use to determine the default as you type? I will never understand computers. @SRLevine to my understanding each server does store the canonical hashtags, so you are querying those from your server. But the way those canonical hashtags are *set* is where federation comes in - your server monitors all posts it sees, not just ones that originate from the server.  @inherentlee huh. There used to be a page where you could edit tags if you were an admin, and turn an all lowercase into a camelCase tag, but I can't find it any more... @Anke no shit?? that's really interesting! i wonder why it's gone. that said, this really requires automation to make it truly solvable. editing one by one sounds AWFUL.  I have a python tool that ingests open text medications names and spell checks them vs FDA and also our cloud of misspellings so that it learns "all possible" misspellings and eventually be used to feed back curated matches for all of those strings. If you approached it that way then the up front costs suck but the long term outputs get better fairly quickly I'd think  @anke @inherentlee It's under trends, I think. You can't edit any arbitrary hashtag from there, but you can fix the capitalization of the top 200ish tags that way. I'll occasionally go in and camel case tags that need it. @m750 that is also accessible and perfectly ok! I think technically (this is me being a pedant!) camelCase is first word first letter lower, all following words first letter upper. PascalCase capitalizes the first letter of all words. @inherentlee @m750 yup, either way is better than all lowercase. I like Pascal myself for the same reasons. Would love to see more auto suggestions using either case! @mmdolbow @inherentlee I do lothe the autocorrect to incorrect case, it gets me all the time. @inherentlee can admins do anything about this on a server after the fact? That seems like a worthwhile feature to have (the ability to change the canonical casing of a hashtag) @griotspeak according to this response by @Anke : https://social.scribblers.club/@Anke/113040680826874850 > There used to be a page where you could edit tags if you were an admin, and turn an all lowercase into a camelCase tag, but I can't find it any more... @inherentlee @griotspeak I found it several years and upgrades ago. I don't know if it was removed or if I just can't find it.  @zeborah @Anke @inherentlee @griotspeak It does indeed let you change the hashtag case Tags are very hard to find if they are not trending, in my experience @inherentlee I've wondered so many times why the camel case versions of tags showing up first is so inconsistent from tag to tag! I assumed it was most used shows up first, but it's so interesting to know this is how it works (and good info to spread!), thank you for sharing this! @sodamnqueer most used would make sense, but that would require the server to either (1) regularly survey a hashtag for changes to popularity or (b) survey the whole hashtag whenever the autofill function is called, both of which get quite expensive! @inherentlee that makes sense! I am very much *not* a techy person so I usually chalk most things I'm curious about that don't have clear other causes (ie bigotry influencing choices) up to 'probably caused by something I don't understand enough to get' lol  @inherentlee I do appreciate this, but it might stress me out in the future. 😅 @inherentlee typing in hashtags, having them suggest autocorrect to non-camel case, and me knowing why. I'll be like "why??? Gahhh!" Don't worry, I'll be fine. Just a little strange... @sillyCoelophysis ahhhh yeah that bothers me as well. To the point I went and fucked about in the code! But I couldn't fix it with my limited ruby knowledge @inherentlee There are two species of camel case: #camelusDromedarius and #CamelusBactrianus. My autofill suggested lower case for both. @AlgoCompSynth screen readers already handle both camelCase and snake_case, nothing needs to be taught there  @inherentlee Oh wow this is why my hashtag ArtByWomen (not actually hashtagged it because this is obviously not a relevant post) gets done right by other people as well!  @inherentlee This is such a pain in the ass as I swear every hashtag I use is all lower case and I have to go in and fix it, but I didn't realize that once it's done for the first time it's set in stone. 😞 @retrosponge well, not entirely set in stone - instance admins *may* be able to change it, or we could develop tooling that changes it. but for the average user, yes, the autofill is unchangeable :( @syderiaos the second one you mention is called PascalCase and is also friendlier than lowercase! adding: snake_case (edit: some screen readers will read out underscores, so this is probably the least good) also effective to my knowledge. point is, use something that effectively denotes individual words in some way. @inherentlee  Thank you, good to know! I trained myself to use caps for each word, is camelCase preferable?  @inherentlee I'm a PascalCase fan. I spent too many years as an editor to do the first word in lowercase. @inherentlee As an ex-Tumblrite I have a strong aesthetic preference for all-lowercase tags. Hadn't considered the accesibility aspect before. Thanks for this informative post. (For those not in-the-know, Tumblr tags can function as a lowercase "whisper space" for unloading weird, funny, or unfiltered thoughts one doesn't wish to include in the post body, e.g. "what was i thinking with this will delete tomorrow"). @stranger_stars aesthetically i love lowercase too, but I've come to live with capital letters in hashtags and in longer text blocks - accessibility trumps aesthetics for me. but i get ya! @inherentlee Different platforms can have different tagging standards for a variety of reasons. Seems like a good idea to adapt as necessary. I'm pretty insistent on using book-title-style capitalisation with AO3 tags because that's the standard. It annoys me when people do shouty allcaps or the Tumblr whisper space thing there. But AO3 allows for spaces in tags so caps style is less of an issue in terms of readability.  @inherentlee Thanks for this. Tangentially related: is there a reason to prefer camelCase to PascalCase?  @inherentlee wow! I had no idea and I'm currently doing a project to show a song for each country in the world, hashtagged rather than as a thread. The lovely @salvae23 suggested a Pascal case hashtag which I used. It's going to be many, many, many toots. Initially the autofill was a bit ambiguous, now it's always correct. Thankyou Salva! |
Gregory's Server
@inherentlee now im losing cycles daydreaming of how to write something to automatically detect words and camelcase a string but getting caught up on a hypotehtical bug where multiple words could be cased out of a single string and then writing a different software to specifically locate strings that can be camel cased into different sets of words cleanly and i really gotta get back to work lmao