Goodbye Gravatar, Hello User Icons!
-
Last month, with the release of NodeBB v0.9.0 coming up, I implemented user icon functionality by default. This change came about due to repeated requests to remove Gravatar integration due to some privacy and security issues.
In contrast to the perceived gain of Gravatar, whether it was removed from core or not was a judgement call. It's very nice to have a user register and be automatically assigned a friendly avatar.
Click here to see the full blog post
-
Another amazing release of NodeBB. Amazing work guys, can't wait to see whats next
-
Very sad about the loss of functionality here. This is dramatic enough that we are concerned about moving forward with the 0.9.0 release - as we had provided feedback previously. It's concerning not only that functionality is moving backwards (the 0.9.0 release is actually looking worse then 0.8.2 for us) and losing ground, but the through process behind breaking the existing releases instead of providing a smooth, option upgrade path. Moving functionality to plugins has left feature after feature in the dust, or halfway broken. We are literally working with less functionality than we had a year ago and that it is intentional decision making to not moving the product forward is incredibly concerning. I do not understand why these decisions are made or what the plan is for the product. By the time that 1.0 releases, what will be left to release.
We appreciate the security concerns with gravatar but frankly, that a BS excuse for someone's pet desire to remove it. That's not a "we are making serious software and supporting our customers" decision. We rely on that functionality and saying it will go to a plugin is crap as has been covered before. Just like the image links that vanished, YouTube embedding that has been broken for forever and buggy mobile experience that no one addresses. Development time is being spent crippling what was solved instead of fixing what is wrong or providing new features.
There is logic, customer-centric ways to dealing with Gravatar if security concern was in any way the reason for the decision - like disabling it by default or providing some controls from the site managers. Doing this is something wholly different and isn't how I see quality software being maintained. We need to know that when we commit to NodeBB that NodeBB is going to commit to us. It's not trivial to choose a platform and if that platform is unstable and the developers don't understand how production environments work... that just doesn't fly.
Sure the code is open source and we are free to fork it and go our own way. We aren't looking for a product where we have to develop it ourselves, we would have done that from the beginning if that is what we wanted to have to do.
Now, not only are we being mocked with "removing core functionality" as if it somehow was to the benefit of the customers, but as an additional slap in the face the BS code that takes our existing Gravatar functionality away breaks the upgrade process so we can't even move to 0.9.0 on our test platforms to see how bad the step backwards is going to be.
And all of this after we expressed these concerns and trusted NodeBB to take rational, customer-focused steps to not act like this months ago. We are not happy in the slightest. We've committed nearly two years to this platform and consistently see it slipping more and more into excuses and less and less into being production viable.
-
Have we, using NodeBB in production for a moderately busy forum, simply outgrown the NodeBB vision and are not really a target audience?
-
That's... Damn. I mean, we want to make everyone happy with our product but its difficult if as you said you've been upset for awhile but haven't raised an issue on our github tracker at all like, (Why is X gone? Why does Y not work?). You know we are responsive and try to make things work for everyone.
Specifically for gravatar, we've been planning a seamless transition, so its one of those things where we decided that 0.9 launch shouldn't wait for it, but its not like we couldn't expediate the build if you expressed that you absolutely needed it... I'd just wait a few weeks before migrating that's all.
-
@psychobunny said:
That's... Damn. I mean, we want to make everyone happy with our product but its difficult if as you said you've been upset for awhile but haven't raised an issue on our github tracker at all like, (Why is X gone? Why does Y not work?). You know we are responsive and try to make things work for everyone.
Actually I did exactly that and the answer I got was "we pull things out of the core and send them to plugins, the plugins are not supported so they may or may not work." So any "moving it to plugin" turned onto code for "dropping the feature."
-
@psychobunny said:
Specifically for gravatar, we've been planning a seamless transition, so its one of those things where we decided that 0.9 launch shouldn't wait for it, but its not like we couldn't expediate the build if you expressed that you absolutely needed it... I'd just wait a few weeks before migrating that's all.
@julian and I had discussed the importance of Gravatar weeks ago. Only tonight did we find out that the support for it was being dropped so that there will be no expected seamless path AND that the code that breaks it also breaks more stuff. Why expend effort going backwards? Why not make the product better rather than worse? All we wanted was to be able to force Gravatar and have no local avatar images. What we had today wasn't ideal but it was good enough.
If we wait weeks or months, what does that get us? A plugin that limps along and sort of works but isn't official and isn't supported like YouTube or searches?
-
There will be a seamless path as I first mentioned, and it will be officially supported ex markdown, composer etc
-
Anyways, is there anything we can do aside from finishing what we plan to do with the gravatar plugin?
-
@psychobunny said:
There will be a seamless path as I first mentioned, and it will be officially supported ex markdown, composer etc
@julian and I were discussing and he was unsure how seamless it would be. And since it would be unofficial - what's the plan? Even if it is seamless today, will it keep working and keep being supported? If so, what's the reason for going to a plugin since it was explained to me that plugins would not get support and would not necessarily be reliable (which we've seen as we have to keep switching or dropping plugins.)
What we want is to not lose functionality that we have today. Is that definitely coming someday? What is the commitment to keeping it around?
-
Whats the problem with gravatar? Whoever is interested in your forum will have time to upload some custom avatar...
-
@julian said:
This change came about due to repeated requests to remove Gravatar integration due to some privacy and security issues.
It needs to be understood that while there might have been some requests for this, there were also requests to specifically not take this action as well. I feel that there was a reasonable middle ground that only involved changing defaults and not changing functionality. This is where the true concern stems from. The security fix did not require putting functionality into a plugin to satisfy both the security concern customers as well as the Gravatar using / continuous functionality customers.
-
@exodo said:
Whats the problem with gravatar? Whoever is interested in your forum will have time to upload some custom avatar...
The problem is removing it.
And we have found the latter sentiment to not be true at all. Even very active forum users will do nothing to upload an avatar. Using Gravatar was an important part of that functionality for us.
Also, Gravatar acts as a CDN for many of these forums. Remember that we can't use services like CloudFlare because of Websockets. So having the avatars at a third party was pretty important.
-
You'll notice, for example, that I have no avatar here even though I took the time to set one up. So the same loss of functionality has happened here. So the theory that people will go back and fix them after having done it one and having them "Just disappear" doesn't hold up even in this conversation.
Also, I thought that Gravatar images were going to be converted to local. But my avatar just vanished?
-
Okay let's just back up a bit.
and it will be officially supported ex markdown, composer etc
Officially supported, seamless transition. You don't have to upgrade until the plugin is ready. I'm not 100% sure what the problem is? As I said earlier, aside from actually getting this done which we plan to do soon, what else is the problem that we should look into?
-
Actually I did exactly that
Post GH issues, I'll be happy to reopen if they were unjustly closed and have a second look
-
@psychobunny said:
Okay let's just back up a bit.
and it will be officially supported ex markdown, composer etc
What's official supported versus not officially supported in plugins? When I asked long ago about this specific process it was made clear to me that any plugin functionality was "use at my own risk." Any, specifically coming up in response to things that "had been core and used to work" and were "now in a plugin that didn't get updated to work with the current version."
-
@psychobunny said:
Officially supported, seamless transition. You don't have to upgrade until the plugin is ready. I'm not 100% sure what the problem is? As I said earlier, aside from actually getting this done which we plan to do soon, what else is the problem that we should look into?
I'm happy to wait, that's not a big deal. Although I think there should be heavy visibility into loss of core functionality changes - I'd have had no idea that things were expected to break had Julian not IM'd me to stop me from upgrading voluntarily (he knew that we relied on the Gravatars from the previous conversation where I had requested that this functionality not be removed.)
What we have today is that users sign up and get Gravatars automatically. No extra work. We have local avatars disabled as we do not want a mix of things and we do not want to be serving out images from our NodeBB servers. In a seamless transition we would like it if the users do not lose access to the Gravatars that they have now and that new users continue to have Gravatars by default without having any idea that they need to disable them, upload anything, etc.
There are some simple upgrades that we'd love to see... like removing the interface to upload an avatar when the functionality is removed (currently they are confused because they click it and it doesn't work) and instead having a link that takes them to Gravatar instead. But that is all extra and pretty trivial.
What we don't want to see is current working Gravatars lost and/or new users not having Gravatar work transparently without needing to do anything special.
-
What's official supported versus not officially supported in plugins?
Anything we have removed from core and moved into a plugin (which allows anybody to build their own solution) is officially supported
One example: https://github.com/NodeBB/nodebb-plugin-composer-default
Removing the composer from core allowed us to have the redactor WYSIWYG composer as an option. The idea is that building hooks and making things modular so people can build their own solutions.
As I said on twitter, I understand the frustration, but we're trying to build a solution that makes everybody happy, and we are working on one that makes you happy too. Just keep your thoughts on GH and we will try our hardest.
-
@scottalanmiller said:
@exodo said:
Whats the problem with gravatar? Whoever is interested in your forum will have time to upload some custom avatar...
The problem is removing it.
And we have found the latter sentiment to not be true at all. Even very active forum users will do nothing to upload an avatar. Using Gravatar was an important part of that functionality for us.
Also, Gravatar acts as a CDN for many of these forums. Remember that we can't use services like CloudFlare because of Websockets. So having the avatars at a third party was pretty important.
You can use cf and websockets. Just set another subdomain with grey cloud only for websockets, its not ideal but it will work fine.
Gravatar cdn is slow++
Removing functions is always bas but i,m sure there will be a plugin for it.
dont need to hurry
