Captain Jordan

Mahz's Dev Journal

I need a fourth rating option, and I need it to be silly.

@Awson

Maybe a option?

@TheMaster99

If you do, I will probably click that on every single post just to troll you guys. You have been warned.

Mahz's Dev Journal

Downvote is too easily abused, especially by trollers, flamers, and the aforementioned vengeful wronged middle schooler. @Captain Jordan, I see the sense in what you say, and I would agree that in a writing community it is better to write a proper response than simply click a button. To be honest, I'm pretty ambivalent about a post rating system. Like you, I'm just giving my two cents.

@BBeast

Fine, in the future I'll give my three cents. Then mine will be worth more than yours! *insert maniacal laughter here* Once I'm in a rating mood, I'll like your post for use of the 'wronged middle schooler'. Let's make that phrase a thing!

Mahz's Dev Journal

I just find the whole 'thanks' and 'laughs' part so subjective. So are 'likes', sure, but at least you can imagine a broad spectrum for which someone liked the post. Seeing that someone has a lot of 'laugh' ratings, and then finding their posts completely unfunny, for me that would make the ratings suspect and hard to believe the next time I saw something like that. Same with helpful posts. There are so many naive problem solvers out there (like those of us in this topic, hey-o!) who produce suggestions and tips that aren't really all that helpful. Except, to the OP, they may just be considered helpful, or at least they sound helpful. When someone with real knowledge in that field comes along, then, and sees all the 'thanks' ratings on a post that is completely off-base, it also makes such ratings suspect. This kind of thing shields users from true and accurate evaluation by masking their posts with 'likes' and 'thanks' and 'laughs' from people who may or may not actually get the intent behind the post, or who naively believe the post was helpful. I know we're in a society of instant gratification, of 'click and move on', and so forth, but this is a writing community. Having ratings such as these which trend towards simple categorization of a person, rather than writing out a thoughtful response, is kind of defeating the point of this community. Which is why, if we must have some form of rating, let's make it as simple as possible. tl;dr: I think there's too much subjectivity in the specific ratings to base generalized opinions of a person on those ratings alone, and makes it a poor argument for having three ratings.

@Captain Jordan

The rating system is useful as a giver->receiver token system, not as a character-evaluation model. The aggregate stats you see on the profile are interesting but rarely useful. You're pretty much confirming why nobody uses aggregate ratings to evaluate people. Maybe the word "rating" and the rating-table on the profile confuse the intention, but someone with a high rating in a +1 system isn't guaranteed amazing, either. They've simply provoked a lot of positive ratings usually because they've just been around a while with a mediocre-friendly disposition (example: my Reddit profile ). A "+1/-1" binary system is even more subjective with even more useless aggregates, but I'm not convinced that's an issue. Worst case scenario, the granular rating system is trivially degraded into a "+1" system or removed entirely.

@Mahz

Pretty much, which is why I wasn't convinced that it was a reasonable argument for having three similar modes for rating. And yes, reddit is a good example of why it fails to be useful, even with all the trickery reddit pulls, people still gravitate towards upvoting a well-upvoted post and downvoting a well-downvoted post. I seem to notice that a positive-only system seems to at least avoid the problematic system of "who would downvote this?" or worse, actually going after a downvoter in revenge like some wronged middle schooler. At worst, an upvote-only system creates the uncertainty when a lack of upvotes occurs, is the absence of votes due to dislike, apathy or simply a post that someone missed reading? I still think the rating system should exist as a simpler "like" only (or +1 or upvote or thumbs up or whatever).

Mahz's Dev Journal

I think the three distinct ratings are useful. While a single rating is fine for individual posts, it does not help much in determining what kind of poster the user is as a whole, as it is here:

Mahz

If there was only one rating, we would not be able to distinguish between users who write lots of helpful posts to users who write lots of funny posts, or if they do a mix of both. These three ratings, when viewed from the user profile, provide a suitably diverse rating of the user, and does not blur the reasons they were up-rated (quality, helpfulness and humour being the three main categories here).

@BBeast

Perhaps. I just find the whole 'thanks' and 'laughs' part so subjective. So are 'likes', sure, but at least you can imagine a broad spectrum for which someone liked the post. Seeing that someone has a lot of 'laugh' ratings, and then finding their posts completely unfunny, for me that would make the ratings suspect and hard to believe the next time I saw something like that. Same with helpful posts. There are so many naive problem solvers out there (like those of us in this topic, hey-o!) who produce suggestions and tips that aren't really all that helpful. Except, to the OP, they may just be considered helpful, or at least they sound helpful. When someone with real knowledge in that field comes along, then, and sees all the 'thanks' ratings on a post that is completely off-base, it also makes such ratings suspect. This kind of thing shields users from true and accurate evaluation by masking their posts with 'likes' and 'thanks' and 'laughs' from people who may or may not actually get the intent behind the post, or who naively believe the post was helpful. I know we're in a society of instant gratification, of 'click and move on', and so forth, but this is a writing community. Having ratings such as these which trend towards simple categorization of a person, rather than writing out a thoughtful response, is kind of defeating the point of this community. Which is why, if we must have some form of rating, let's make it as simple as possible. tl;dr: I think there's too much subjectivity in the specific ratings to base generalized opinions of a person on those ratings alone, and makes it a poor argument for having three ratings.

Mahz's Dev Journal

Last night I created a post-rating system that's inspired by a popular Xenforo plug-in called Post Ratings. Some first-effort screenshots: - You can only rate a post every 30 seconds - Once you rate a post, you have 30 seconds to undo it Some things that might not make it into the first release: - Receive notification when someone rates your post - A button that displays who rated the post - Profile tab that paginates through your ratings - On desktop, the rating buttons only appear when you hover a post (Keeps them out of the way) - On mobile, clicking a "Rate Post" button pops up your three options (Prevents sausage-fingering ratings while scrolling) - Reduce the visual footprint of a post's ratings so it's not so in your face Thoughts?

@Mahz

I guess I'm confused at why we need three different pieces of karma. Shouldn't a simple "Like" or "Thumbs Up" do the trick? It can mean different things based on the context of the post and the rater, but at its baseline the result is a metric for good posts. I would think that a post you like, a post that has good information and a post that makes you laugh would all qualify as 'good', making three different ratings redundant. If you're going to have specific ratings, it seems a bit useless unless there's a description as to why you 'liked' something as opposed to 'thanking' it or 'laughing' at it. And then we get into feature bloat, when all of it could be simply described by a single rating system. My $0.02. As a side note, I do have a suggestion. If someone edits their post after receiving a rating, should the rater be allowed to (without a time limit, unless the time limit is related to viewing the edited post somehow) remove their rating? It's probably a niche scenario, but some people may decide to craft a post to get a rating, and then edit it so that the raters look foolish. That's definitely not Fonz Cool, but it's hard to detect this kind of thing unless people notice the edit and take action. Removing ratings after an edit could be that action, but I suppose a simple post reporting could also do the trick (but people seem more loathe to get someone in trouble, aka snitch, than they are to simply and passively remove their involvement in someone's trickery).

Mahz's Dev Journal

Could I suggest an "upvote" option?

@Melkor

This is actually something I'm working on!

I'm aware of the fact that you do not. However, is it currently impossible for the server code to accept the image redirection, and then before displaying it, resizes it in the same way it would an uploaded image? It could cause an extra few calculations, but what prevents the server from essentially saying "This linked image is too large, so I'm going to scale it down before displaying it" and output the resized version?

@LegendBegins

Because it's an expensive operation and the resized image needs to be saved somewhere (instead of doing it every time). In other words, it involves implementing precisely the "hard" parts of a direct upload system. Whether the user is sending a request to my server with their image (i.e. file upload) or another server is sending a response with an image (i.e. my server making a request to someone's remote avatar URL), I have to do pretty much the same thing. Though in the time I've spent talking about this system I could've probably implemented it.

@Mahz

Kind of a Fry moment? Shut up and code my feature?

Mahz's Dev Journal

No. Right now, you can go to massivewallpaperimages.com and paste the URL of an 8192×4608 image into the Guild's avatar box. Since the guild's code right now is 100% naive and trusting of the client (always a bad idea), every user will have to download that 8192x4608 image. If massivewallpaperimages.com has no-cache headers on their images, then everyone on the guild will have to download your avatar image every time they see it. That's why I depend on people to just follow the rules right now. Now, when I implement a system where users can upload their avatars to the guild and I can process their image, then the guild will resize and downgrade that 8192x4608 image into a 150x85 image. In other words, I will be able to enforce the 150x150 max-size constraint since I control the images. I currently do not control the images.

@Mahz

I'm aware of the fact that you do not. However, is it currently impossible for the server code to accept the image redirection, and then before displaying it, resizes it in the same way it would an uploaded image? It could cause an extra few calculations, but what prevents the server from essentially saying "This linked image is too large, so I'm going to scale it down before displaying it" and output the resized version?

@LegendBegins

That's an awful idea for folks on mobile and limited dataplans, especially considering Mahz's example.

@Captain Jordan

According to my idea, all the data load would be on the server, since nothing more than our current 150X150 pictures would display. Perhaps I'm explaining my intent incorrectly.

@LegendBegins

What's the benefit between implementing this and implementing an image upload system? I'm not sure how fetching from a 3rd party site is going to be any cleaner/safer/less open to abuse than uploading.

Mahz's Dev Journal

No. Right now, you can go to massivewallpaperimages.com and paste the URL of an 8192×4608 image into the Guild's avatar box. Since the guild's code right now is 100% naive and trusting of the client (always a bad idea), every user will have to download that 8192x4608 image. If massivewallpaperimages.com has no-cache headers on their images, then everyone on the guild will have to download your avatar image every time they see it. That's why I depend on people to just follow the rules right now. Now, when I implement a system where users can upload their avatars to the guild and I can process their image, then the guild will resize and downgrade that 8192x4608 image into a 150x85 image. In other words, I will be able to enforce the 150x150 max-size constraint since I control the images. I currently do not control the images.

@Mahz

I'm aware of the fact that you do not. However, is it currently impossible for the server code to accept the image redirection, and then before displaying it, resizes it in the same way it would an uploaded image? It could cause an extra few calculations, but what prevents the server from essentially saying "This linked image is too large, so I'm going to scale it down before displaying it" and output the resized version?

@LegendBegins

That's an awful idea for folks on mobile and limited dataplans, especially considering Mahz's example.

Mahz's Dev Journal

Maybe Verdana? The 'I' (i) is distinct from the 'l' (L), which seems to be your main gripe.

@TheMaster99

I support Verdana. I'm not a big fan of serifed fonts on the web, and Verdana is a good compromise. Also check into the Google Webfonts, there's some good ones in there. This author goes over some top fonts. Source Sans in particular has a straight uppercase I(i), with a small tail on the lowercase l(L) for differentiating the letters.

Mahz's Dev Journal

Oversized avatars can be draining on mobile users and people with slow connections. Just resize your avatar with 2 clicks and be done with it. Though, on that note, I also need to remember to add a "Hide Avatars?" option for users.

@Mahz

But won't that problem still persist when you include the option to upload pictures to the site directly?

@LegendBegins

From what @Mahz was saying, that feature sounds more like an if.