The long-awaited moment... Dracula theme added! Oh also, the site is publicly visible now (but let's not be too noisy about it yet).
Alright, after repeatedly delaying it for various reasons, Tildes is now publicly visible. This means that people no longer need to get an invite to be able to browse the site. An invite is still required to register and participate though (and I'm intending to keep it that way for the foreseeable future).
This should be a huge boon to the overall process—people will be able to check out the site before requesting an invite, which will save a lot of effort giving out "wasted" invites to people that just wanted to look and don't continue visiting afterwards. I want to talk more soon about making the process of getting an invite easier, but this should help a lot for now.
Please don't try to bring a lot of attention to the public visibility just yet (you're welcome to tell friends or small groups though). I fully expect some people to notice it naturally, but I'd like to try to keep it a little quiet still over the weekend. There are still a few things that I'm working on, and I'd like to get a bit more done before we start promoting it too widely. Early next week I'll make a post on the Tildes blog announcing it, and then we can go all out with it.
A couple of other notes about public visibility:
- Logged-out users can select their theme - there's a dropdown box in the site footer for them to change it, but they'll have to set it individually on each device.
- Logged-out users can only see the most recent 20 posts on user pages, they don't have access to the full pagination like logged-in users do. That was discussed a little in this thread. I don't know if it will stay this way permanently, but we can try it out for now.
I've also topped everyone back up to 5 invite codes again. The public visibility may cause some of you to get requests from people for invites, so please let me know if you need more. You can access your invite links here: https://tildes.net/invite
And as one other thing, I've also added the tirelessly-requested Dracula theme. This is the first time I've tried using the revamped theme system that @Bauke set up to add a completely new theme, so please let me know if you notice any oddities with it (or if you think I used the colors of it wrong or anything, I don't use Dracula personally).
Please let me know if you have any questions, concerns, feedback, etc. about the public visibility. This is a huge step in the site's progress, and I'm definitely both excited and terrified about it.
I don't have any particular plans for themes. I don't want to get too crazy with it where people have to choose from like, 200 themes, but overall I don't think it would be a big deal to add a decent number of in-demand ones. They're quite easy to add, this is pretty much the entirety of what it took to add Dracula: https://gitlab.com/tildes/tildes/blob/master/tildes/scss/themes/_dracula.scss
Maybe add theme builder to the user profile so we can each build our own custom theme (and share our designs) then just select User Theme in settings
It wouldn't be very simple to set up user themes, since the themes need to be compiled using SASS. It's certainly not impossible, but it would require a significant amount of work and probably won't be something I look into doing in the immediate future. It's not too complicated for people to use a CSS-customization extension like Stylus to create their own theme, if they really want to do it.
If using CSS4 Variables, these could even be modified directly in the developer tools. Makes it really easy to test new color schemes. Though I'm not sure what browsers you support. Unfortunately IE is still a thing.
In lieu of that, you could probably get SASS compilation working with local overrides.
Easier for who?
I think you missed an important point: this is only easier for developers. Non-developers would have a hard enough time using something like Stylus and modifying plain old CSS, let alone setting up an entire Vagrant environment and using SASS to compile the appropriate CSS for them. And although there are a lot of developers here on Tildes, there is still a substantial non-developer user base.
And, we would hope that Tildes will continue to attract a substantial non-developer user base. We want this website to be welcoming to all people, no matter their skills or work background or interests. In fact, I expect in the future that developers will form a minority of the user base here on Tildes as we attract a lot of different types of people.
Absolutely. As a developer, I do enjoy spending plenty of time in ~comp, if my occasional long-winded topics and comments there are any indicator. But Tildes isn't ~comp, and it shouldn't be, otherwise we wouldn't have all of the non-~comp topics that we do. As much as I like programming, it's only a small chunk of what life has to offer. We need the ~humanities, we need ~life, we need ~music, we need ~food... we need that kind of diversity if Tildes is ever going to measure up to its purpose.
I will always frequent ~comp, but I look forward to seeing how the other communities we have now and will have in the future will grow as our user base diversifies :)
Please try to understand something from a non-technical user perspective: the time it takes to perform a task and the amount of effort to perform a task are not the same thing. It's all about the learning curve.
CSS has a fairly low barrier to entry. You can find any number of the themes that were shared for e.g. Stylus and try tweaking what's already there to try to get the result you want. It still takes a fair bit of effort, but it's simple and accessible. Once you throw VirtualBox and Vagrant into the mix and start expecting users to use a CLI, you're ramping up the difficulty significantly. Users typically don't know anything about CLIs, VMs, port numbers, the developer console, Git, or any of that. Now, instead of just having to tweak some values in CSS that someone has already written, they have to learn an entire suite of tools and an entire new way of thinking just to get a VM spun up. Even during my first year or two of studying programming, just using a CLI alone was really fucking hard, and that was as a CS major.
People like you and me have the technical background to have the luxury of saying "yeah, just spin up the dev environment and use the SASS route, easy peasy", but most people don't.
I would just like to confirm, as a non-developer myself, that most of @hungariantoast's explanation is gibberish to me - and even the parts I do understand are totally beyond my abilities to do anything about.
Also, quite frankly, I'm a user of software, not a developer of it. Who here can rebuild a kitchen if they want to change its theme?
This ties back into our other brief discussion earlier about how Tildes isn't meant to be ~comp and really shouldn't be. It's easy to forget about the average user when there are so many developers here and so many discussions of a technical nature, but it's important that we not forget about the non-technical users.
Sorry. I was misled by phrases such as "it would be easier for a beginner" and "if someone is a beginner", and the context in which you weren't actually replying to someone who had specifically asked for instructions on how to change the theme, but were explaining in general how someone could go about doing this. I didn't realise you were directing your instructions only at experienced developers.
Have you considered using CSS variables instead of a preprocessor like SASS (or a mix of both)? That way, you could add some color pickers to the settings and users could pick their own colors and therefore make their own themes.
There is, actually! Here's an example:
The important piece is the enclosing within
There's still the problem of e.g. conditional evaluation via SASS that isn't so easy to port over, but I was actually testing out these kinds of teaks last night as a PoC :)
Yeah, it's apparently not a super well-documented bit of syntax, but I managed to stumble across it last night. I was curious to see if it would even work in the first place since I'd noted them as a possibility in another comment. I have most of the replacements successfully made locally and they seem to work, so that's encouraging. There are just some things that SASS does that I would need to try to find an equivalent for to get the last few items moved over.
That being said, if you plan to tackle this one then I'm certainly not stopping you :)
The downside to using extensions is that you don't have access to the simple SASS setup that allows you to define the mapping of otherwise repetitive values and have the styles applied with relative ease.
Of course, that doesn't exactly eliminate the limitations you're referring to regarding the use of SASS. Probably the only way around that would be to write user themes directly to the file system which, in my opinion anyway, is kind of a terrible solution overall. Maybe it's possible to make use of custom properties? If so, then SASS could likely still be used just the way it is now, but with some minor tweaking to the theme system to apply the values to custom properties rather than directly to the target styles, allowing for user overrides.
I'm not really certain of what limitations there are with regards to CSS, however, so I'm not going to speak authoritatively on the subject.
Edit #1: I'll be testing out some minor tweaks over the next few days. So far it looks like custom properties can be swapped for most lines of code and apparently work; however, there are 18 lines of code which, unfortunately, rely on preprocessor-defined values to determine lightness for conditional checks and/or fallback values and would thus be incompatible with user-defined changes, so some additional work would need to be done to handle these correctly. I'll continue digging into this over the next week or two as my schedule permits :)
re: user themes, would there be any way to have something like
this on a user's settings page, where we can define our "background", "text", etc colors, and then have some sort of JS client-side parser/compiler for sass which stores the generated custom css in our browser's cache, taking any workload off of ~ and putting it on users, at the expense of maybe not working on X application which doesn't allow JS or meaning it only works on one machine at a time, and possibly storing only our custom color inputs on our user page so if we log in on a different browser it's re-populated and we just need to click "create" again for that browser?
I have no idea if any of this is either possible or easy, just a thought I had for how it may work without requiring a lot of overhead on the server.
Oh, man. That's amazingly compact! That is really satisfying to look at, and it's clear just by looking at it that it would be fairly trivial for anyone to include a new theme now.
Fantastic work, @Bauke!
Offering a choice that's too big can be very discouraging. I think agree a small choice with the option to add them yourselves and only adding very popular ones is the way to too
Congratulations, @Deimos! It's great to see the site begin to flourish. There's certainly been more activity in the past few weeks and it's been quite nice.
In terms of roadmap looking forward, what's next? I'd really like to see labelling of posts, say for example "misleading", "reactionary", and "low quality". Maybe "misleading" requires an input reason. These would (mildly!) weight posts downwards, and also gives approved users ability to change the title if a threshold of votes to misleading labels is met. Or something along those lines.
When I say mildly, I do mean it. Depending on whether reputation is considered, maybe each label application has a small fractional downward effect compared to a vote. I'm talking quarters, eighths, and tenths here. I'm still of the opinion that comments, activity, and votes should be the primary post ranking metrics.
Another nicety to add would be gravitas/reputation influencing how important someone's applied label is. The latter's probably going to be quite hard to figure out, because you might need to take into account many different parameters; but it'd be nice as someone who's been actively using the site for 6+ months to be able to appropriately apply labels at a higher weighting than a total newcomer who might otherwise not have a complete understanding of the Tildes dynamic.
I'm not totally sure, honestly. There are a ton of different things that need to be worked on. Beyond the ones you mentioned, a few other things I'd like to do in the near future (definitely an incomplete list, no particular order):
I'm still planning to work on a "how to" guide for Tildes on the unofficial wiki (mentioned here and here). Are we going to be stepping on each other's toes, or duplicating effort?
There would probably be duplication, but I'm happy to procrastinate on it and let all the other people that say they're working on docs/intro-posts/etc. do all of that kind of stuff, and then maybe I can build some of it into the official docs. That's why I think we need an integrated wiki, it's hard for people to coordinate all of that now (and needing to do it all through GitLab is difficult for people that don't already have experience with similar systems).
Tell me about it!
What kind of improvements and additions are you wanting for the docs site? I'm a technical writer for $megacorp and can lend a hand if needed. I poked around the static site repo, but didn't see any specific contribution requests other than not touching legal and privacy stuff.
There's not really anything in particular, but there's a ton of decisions/knowledge/etc. that's all very scattered right now and just getting passed around through the community as relevant threads arise. It would be nice to try to organize it all so that there's an easy way for people to look up information, and get caught up on past decisions and such.
Hey so a bit of a noob question: Im here to contribute and really like the site so far, but Im overall relatively new to the "collaborative coding" scene in general. Is there somewhere I should discuss ideas I have before starting to work on them? Or should I just start doing stuff, and toss up merge requests on Gitlab and discuss the ideas there?
Have a read through this file, it has a section about this: https://gitlab.com/tildes/tildes/blob/master/CONTRIBUTING.md
Let me know if you have any more questions after that though.
Will do, thanks :D
Obligatory plug for invite links. Those should be rather fun.
Yeah, we need a wiki. We need rather a lot of things. :P
One potential issue with the 'long-time' users having more influence is that it may contribute towards circle-jerking. Someone with a 1 week old account may not 'know how the site works', but their opinion on what constitutes a low-quality is probably more valid than older users that are used to the general low quality. Think about old subreddits for example that have settled into a stagnant meta to imagine what I mean.
A similar thing for activity - high activity doesn't necessarily mean a more 'correct' opinion, it just means they spend more free time on the site, or are more talkative/have stronger opinions that they feel they need to defend. I would almost say more time on the site is actually a negative, but that would just encourage multiple accounts, so that isn't a useful metric to use.
But what if the older users are used to the general high quality in a internet forums? Would their opinion on what constitutes a low-quality post be valid then?
Remember that Tildes has started out, as much as possible, with an emphasis on in-depth content and high-quality discussion. That's the culture that we've been working to develop over the past 8-9 months, and that we'll continue to develop as the site grows. So, how can a newcomer - say, someone who signed up only a couple of days ago - be a better judge of what's low quality and what's not acceptable here than someone who's been around for a few months?
How so? How is it negative to be involved in, and engaged with, developing this site by posting content and/or participating in discussions?
You're basically saying that newcomers know the culture and expectations of Tildes better than people who've been here for months. Would you say that about anything else? Would you say that someone who has never visited France knows the French culture better than someone who's lived in Paris for years? Would you say that someone who has never a member of a university social group knows the culture of that social group better than someone who's been a member for a few semesters? How does inexperience and ignorance trump experience and knowledge?
There is a lot of truth to what you say. I guess I was focusing on the problem of communities getting stale and wrapped up in their own meta. My thinking of having the new members having more say than older members is that they may have a 'fresh set of eyes', and can generally see problems that have become the norm to others. But I didn't really think this through and it would come with many new problems, some of which you give. The culture we are trying to develop would be harder to maintain against a constant influx of new members for example.
As for your second point, I was trying to figure out a way of preventing a few people that are active on the site having more weight and importance by having not only a higher quantity of posts, but all of their posts being more important too - I felt like it was double dipping in a sense. Also this particular metric only encourages quantity, not quality.
The suggestion you were replying to was "Another nicety to add would be gravitas/reputation influencing how important someone's applied label is."
You may have noticed that there's a 'label' link underneath all comments on Tildes - except your own. You can apply labels to other people's comments, such as "Exemplary", "Noise", and "Malice". It is extremely important to note that you can't apply these labels to your own comments, only to other people's comments.
In this context, the suggestion was made that we could add a similar labelling scheme for topics. And, if a long-term user applies a label to another person's topic, that label would carry more weight than the same label from a new user. So, if I, a long-term user, labelled your comment or topic as "Exemplary", that would carry more weight than if a newcomer labelled your comment or topic as "Exemplary". It is extremely important to note that I am not able to label my own comments as "Exemplary", and would not be able to label my own topics as "Exemplary" - so I can't do anything to influence how my own topics are viewed by others.
The actual definition of "long-term user" would be more like "user who's been on Tildes for a while and has made generally good contributions to the site". It's not solely about time served, or even quantity of posts. There will be metrics to measure how good people's contributions are. The term used around here is "trust" or "reputation": Tildes will have a measure called "trust" or "reputation" which keeps track of a user's overall history on Tildes. To name just a few examples that have been suggested over the months:
If a lot of your comments get labelled as "Exemplary", that would increase your trust score.
If a lot of your comments get labelled as "Noise", that would decrease your trust score.
If you generally vote on topics that end up getting high vote scores, that would increase your trust score. (You're showing that you know what is good content.)
If you generally apply labels that a moderator overrides because they were not applicable, that would decrease your trust score. (You're showing that you don't know how to use the labelling system correctly.)
... and so on.
It's not as simple as "You've been here six months so your votes get extra weighting." There's going to be more to it than that. You have to be active for those six months (or however long), and your activity has to result in positive contributions to Tildes.
As an aside, I really want this to work, but seeing some users of the site starting to use comment labels as disagreement buttons concerns me.
The net effect, if extended to the extent of Reddit's misapplication of the downvote button, would be that users who post unpopular—but reasoned or thought out opinions—would become less trusted through a hellish feedback loop. Which would really suck.
I could make a reasoned post in ~comp about why I think Macs & better than PCs. Or talk about how I think nuclear fission is a dead end energy source in ~enviro. On the whole, I'd get a lot of disagreement on those fronts, but if people start abusing the "noise" or "malice" tag to tag my posts as such, then I could end up with a lower reputation than others merely for posting unpopular opinions.
It seems like the first step really needs to be to ensure that there's some incentive (or inversely, disincentive) to ensure they aren't acting as measurements of disagreement before we can proceed with using labels to inform reputation.
I'm not contesting your comment at all, ideally I'd like it to work like that. But there's so many angles to consider.
There's an extremely relevant comment by Deimos in that sub-thread:
This ties in to one of the suggestions I presented in my comment: "If you generally apply labels that a moderator overrides because they were not applicable, that would decrease your trust score. (You're showing that you don't know how to use the labelling system correctly.)"
Deimos seems to have already implemented a form of this: he is already manually reducing the "trust" of certain users who abuse the labelling system, thus making their labels carry less weight than average. This is obviously a precursor to the intended automated "trust" system.
That's exactly what I'm talking about. It's not the people who get labelled maliciously who get punished, it's the people who do the malicious labelling.
Yup, read that already. I'm extremely skeptical this will be able to scale, certainly it isn't possible with one site admin alone. So this functionality either has to be delegated to approved community members or automated.
Deimos will not be the only moderator forever! He has already delegated some "moderator"-type powers to various users and will continue to do so.
This site will be run on community moderation, where the line between "user" and "moderator" will be very blurred. There will be lots of community-based moderators here on Tildes.
And the actions of those human moderators will feed into an automated system, keeping track of everything. So, in the future, you might be one of these community-based moderators, with the power to see and edit comment labels. You're reading a thread and you notice that a comment has been labelled "Malice" even though it's not malicious. Using your special powers, you remove that "Malice" label, and then move on to reading the rest of the thread. In the background, the automated system notes your action. It notes the user who applied the "Malice" label in the first place, and makes a record that this user's "Malice" label was removed (it also makes a record against your name, that you removed the "Malice" label). If this happens a few times to the same user - that moderators remove their "Malice" labels - the automated system will adjust that user's trust score downward. If it keeps happening, the system may even remove the ability of that user to apply "Malice" labels at all.
We, as community-based moderators, will have powers to see things and change them. The automated system will track what we do and update users' records accordingly.
It's possible that a majority of users on Tildes will have some sort of moderator-type power, ranging from being able to edit tags on a post (which some of us can already do), through fixing malicious labels, up to banning people. There'll be a pyramid of moderators, with lots of people having low-level powers, and fewer people having high-level powers - and everyone moving up and down the pyramid based on their activity (good mods go up, bad or inactive mods go down).
It will scale. Have you ever seen Stack Exchange or Stack Overflow? They use a similar system, and it seems to work.
Ehh, I am 50/50 on this. Conceptually, I think SE/SO's implementation of community moderation is well-intended, but in practice actually misses the mark—I have witnessed far too many cold moderator actions with unhelpful or opinionated perspectives (e.g., locking posts without providing a clearly confused person good help, or making opinionated changes to posts). I am personally scared off from contributing there, due to negative experience both friends and I have had.
A critical factor in a successful implementation is something that grinds opinion out of nebulous mod actions—if there is obvious disagreement or conflict on mod actions (while also accounting for account moderator inaction on some comments—think: if a moderator doesn't remove a tag, that in some regards could count as a confirmation of that tag), then it should probably be escalated up to someone "blessed" (for lack of a better term) to act based on opinion rather than protocol. (It goes without saying that these "blessed" users should be those representing the best of the community, not the loudest.)
I agree. Their system is a bit aggressive and off-putting, overly mechanical and fixated on numbers. That's not to say it's bad or that it doesn't work, just that I think we can kick it's ass and make something much better.
I love the idea of the comment labels, and something like them for topics (that's at least one tier higher to unlock) doing the work. That way it's impossible for just one person to do something. Just like Exemplary tags work here now, they all count up. You hit a threshold, an effect happens. Enough people use malice for example and along comes a paddlin'.
This keeps it in the hands of many people - aggregate action, not individual action. That will actually work better when this place is larger than it is now - more eyes, more tags. Let the tags do the work.
Separation of powers can help too. Cliques form, but if no clique has access to all features, then cliques can't take control. Having lots of mods dilutes the cliques. It also makes appeals more fair since it can go to a different couple of mods.
There is wisdom in the crowd but also mob mentality too. So equally as important, IMO, is level headed admins who (unlike reddit Inc.) aren't afraid to actually step in when even the trusted crowd gets it wrong and/or abuses the system and give them a stern warning or paddlin' too... which based on my experience interacting with Deimos over the years, I have complete faith he can actually manage.
That'd be the people who make it all the way through the trust system to the top, imo, and yes, sometimes someone has to be there to tell the mob, 'shame on you'.
Just like 1% of the users post all of the content, I'll wager 1% of the users cause all of the problems. All we ever need to do is adjust the thresholds up so that it takes more people to cause actions, and that 1% won't have enough clout to get the system to do anything. They'll just be wasting their time.
This is it, exactly.
We'll also have threads longer than war and peace every time a new feature or tool is implemented in ~tildes.official to get feedback on these systems, and tweak them until they work (as we did with comment labels). At the same time, ~tildes itself will become a place to discuss and propose and critique potential future mechanics. It'll inform the direction of the development. These discussions are crucial to getting these features and tools right.
Paying people to moderate will never work - it'll take more people to do it well than anyone, even Facebook, can afford. Using automated bots to manage things will lead to the mess we see on reddit where automated rules cut down tons of content that should stay up just to try and solve one problem - the bots have no ability to make judgements. Users always get caught in the crossfire.
Narrow AI and machine-learning systems have similar limitations, at best they might be able to play a role sifting through large amounts of data to find things to bring to a human's attention. We do not ever want them making judgements on what actions to take. They aren't good enough for that, and Zuckerberg's insistence that 'the algorithms will save us' to congress is pure bullshit. Human communities evolve and their behaviors change; this problem is permanent and ongoing. It cannot be solved, it can only be managed, and only managed properly by the people living in it and thinking about it, not paid clerks who aren't invested and think of it like a shitty job.
The only solution left is for the users to all moderate themselves. Wherever you are, whatever you are reading on Tildes, being there reading it means you are responsible for moderating it, almost like it's a civic duty (and perhaps it is). Rather than having to check people's history or link them to rules or any of a thousand other annoying, time consuming traditional moderation tasks, it should always stay as simple as 'apply a label' with one or two mouse clicks.
Let the site just aggregate all of that information, and simply adjust people's access to the tools. Use the simple everyman tools well (with 'well' being judged by those who used them before you got here, not code), and get trusted with access to a bit more responsibility. Before you were labeling people's comments. Now you're able to edit tags on submissions... and so on, but only in the places you use the tools (the places you care about), not in the groups you never visit.
The challenge here is figuring out what tools to build, and how to make sure they end up in the hands of people who won't let their ego or bad day at the office get the better of them all the time... and definitely not in the hands of people who are paid to use them to push an agenda. That means transparency. Those things can't be prevented, but they can be detected and mitigated.
Yup, I buggered up. my first post was long ago enough that i actually forgot the details, making my opinion based on a wrong premise. I forgot it was labels, not posts we are talking about.
I'm also hoping that the system used can't be gamed so that simply following the crowd doesn't give an account more trust than it rightly deserves.
Definitely agreed. There's a balance to be struck here; we're dealing with certain metrics & parameters and we're trying to map them into salient indicators for who's more trustworthy with respect to the site's long term goals. It's going to be a fuzzy match regardless of which metric combinations you use, but there are certainly "acceptable" combinations which would work well.
To counter the "long time user" effect, there might need to be some kind of ceiling on where trustworthiness no longer increases with account age. The difference between 1 day and 1 month could be quite large, but 3 months and 2 years might be minimal at best. Same with the "activity measurement" metric.
Then there's a whole discussion you could have around how much weighting should be applied overall as a relative measurement of a base weighting. Should high trustworthiness has an effect of 1.1x times that of an average user, or 4x that of some average user.
There's lots of ways this can go for sure :)
a meta note, but i always find mechanic discussions like this interesting because, in a lot of ways, what we're doing is basically establishing a framework for how communities could possibly operate and organize themselves--one which, perhaps eventually, could become the norm. it's neat. i sorta think of it like the internet equivalent of a constitutional convention.
Y'all need to read Clay Shirky's talk, A Group Is Its Own Worst Enemy. That talk is 16 years old and still every bit as relevant. Every group will self-destruct over time unless the core members get some kinda magic to help them hold it to its goals. Circlejerking is a risk, but remember, the posts will bubble up into the parent groups someday, and the groupthink will get challenged by the parent group when that happens. It really should help.
I'm wondering if different communities could use different weights too. Old players and frequent posters in a gaming community are probably well known players who have real influence in the game (both in their clans/guilds, and within the greater community). But as a community becomes more general (and larger), then it seems to actually become a negative. For example the jokes subreddit did this a few years ago where the comments all became low effort pun threads because there was no 'community' in a group of that size; that subreddit could really have used some new blood.
But of course, the weighting system really has to be a one-size-fits-all for all communities, because it's such a nitty gritty part that many mods would probably just ignore it, or not understand the implications, therefore making it worse. All while adding complexity which is never good.
Yes, I am good at coming up with problems without answers.
Hey that's the first step. A lot of the time identifying the problem is even harder than finding a solution.
Why is that exactly? I frequent a lot of smell subreddits. When I see low-effort content posted to a subreddit that discourages it, it's typically by an account made just a month or two earlier. The older users know better, but new users seem much more willing to post memes and other low-effort content. It's as if they've not bothered to take the pulse of the community, or read the sidebar rules first.
I don't know if that's elitist, but I'm skeptical that new users are more likely to post good content. I'd bet on the opposite being true.
Interesting. It has only been a day or two, and Tildes is already being crawled by Google's search engine.
Yeah, I've noticed Googlebot cruising its way through a lot of the site already. Even just searching for things like "tildes games" brings up ~games right near the top.
Is there a setting to disable user page crawling like Reddit has? I'd prefer my page to not be crawled.
Although it should be said that there might be another option here in the future... the ability to post comments and topics disassociated from your username. If/when and even how exactly it will be implemented is unknown though as there are plenty of things to consider first.
That's truly a shame.
Why? It accomplishes literally nothing other than to imbue people with a completely false sense of privacy. Whether your profile can be crawled or not, every comment you make will get indexed through the topic in which you make them and every topic you submit indexed through the groups in which you submit them.
Exactly. I'd rather have people see me through my individual posts rather than through my profile. Maybe there should be an option to only allow registered users to see my profile.
There basically is that functionality already. Logged out users, including crawlers, can only see the last 20 total things you posted on your profile (comments + topics). Only logged in users can see a paginated view of your account history.
But back to the original point, even if Tildes blocked crawlers from profiles in robots.txt and was using rel=nofollow whenever linking to a profile page internally, if/when Tildes gets popular enough your history will get archived on some third party site that chooses not to obey those rules... whether you want it to or not... and there is nothing any of us can realistically do to stop it.
Hence the "completely false sense of privacy" statement. You only "feel" like the option on reddit is helping you... but in reality it accomplishes absolutely nothing. And if you don't believe me, just go to https://redditsearch.io, select posts and comments, switch the time-frame to all, add your reddit username to the author field and click search. Your entire reddit history is there to be perused and combed through... irrespective of if you had/have "don't allow search engines to index my user profile" checked in your reddit account preferences.
It locked right in to the tags and labels, even managed to deduce that the . in tags like ask.survey was a delimiter. It figured out the ~groups are forum-like objects. It also locked right on to user pages, which is why I think limiting them to 20 comments for logged out users (like bots) was a good idea. We can't keep a public forum private, but we don't have to make it easy to search and index externally either. If you want to invade people's privacy you should at least have to do the work. :P
Yep, even ~games is #1 on Bing under
tildes games. Not that I use Bing.
Now that the site is public, you now have search engines to worry about. I noticed that tildes has an empty robots.txt and there's no robots meta tag on user profiles.
Are you going to add an option to disable search engine indexing on user profiles like reddit has?
I doubt it. I don't think a setting like that is effective at all, since all of the same posts will get indexed through the threads they were posted in anyway. I have it enabled on reddit and can still find all my comments through Google.
As I've said many times, it is impossible to post things on a public site and also stop other people (and search engines, and bots) from seeing them and potentially archiving them. If you don't want something seen, the only option is for you to not post it in the first place.
The Dracula theme is amazing. Thank you for your hard work on this site; it's really coming together.
C - thanks for all the hard work.
I like a lot of what I've read at 'announcing-tildes', so I donated.
Thanks cgrrr! I just saw the email from Patreon about the new pledge, I really appreciate it.
You are on to something more useful to me than Digg, /. or most all of Reddit. I'm not a Facebook type, I barely even tweet, deviantART was sold - and I want to see a safer, non-commercial place to connect and learn. This looks very much like a place I wouldn't be embarrassed to suggest to my sister or my co-writers.
Personally I prefer dark themes and those are usually cool colors so this warmer Dracula theme is a nice change
Drunk fellow here, had them other thoughts but will stop deleting and just say this. Been here as of today, it seems cool and calm. I like it. I think others will enjoy seeing it, and if kept mildly down low at least for now, will grow into a larger version of what it is. A decent corner of the internet.
Jesus. Christ. That username has me splitting my sides.
Now that I got that out of the way, welcome to Tildes! It really is fairly relaxed here. We've had a few incidents in the past where things got a bit heated between members (I'll admit that I'm among them), but for the most part we try to avoid or work through that kind of thing. It's definitely a breath of fresh air compared to the toxic environment of reddit and other social media sites. Hopefully we'll be able to keep that up! :)
It seems the background and foreground colors are flipped:
Tildes Dracula theme: https://i.imgur.com/sfpPXYs.jpg
Bauke's Dracula theme applied with Stylus: https://i.imgur.com/3GLO5Vc.jpg
Otherwise it looks great, but those two are definitely reversed. (I have no idea why imgur made those screenshots huge and fuzzy, but you get the idea).
I'd argue that Bauke has it backwards (or at least chose to use the colors differently than I would). He's using the "selection" color for the background, instead of the actual background color. If you look at something like an official screenshot of the vim theme, it's only the current highlighted line that uses the lighter background color - it's intended for an accented background, not the primary background color.
Ah yes, I see. I'd never seen it in that context before. I agree with the argument, but that's not the appearance we've grown used to when using the theme with Stylus.
I just took this screenshot as I reply here: https://i.imgur.com/MubvcPl.jpg
All that soft lavender is relatively low-contrast for reading, yet there's high-contrast in the text entry box, where we're focusing on our own writing. It's a really nice difference and works beautifully. Bauke's theme is popular for good reasons. We're not primarily writing here; more often we're reading.
Edit: one more screenshot. Big low-contrast areas for reading is one of the things that make Dracula so good.
I'm honestly not sure what more to say about it. Bauke applied the colors differently than intended. If you look at the official GitHub Pages implementation, the background color is dark. Bauke's theme may have used the same color palette, but it's not a proper implementation of Dracula.
I guess I could consider implementing a different theme called, uh, "Baukula" or something, but the theme that you're used to isn't actually Dracula.
I understand now. I honestly thought when I saw "Dracula theme added," it meant Bauke's theme. I knew it was based on a theme for coding, but not being a coder anymore, I was unaware there was an official theme you were following. I assumed when people were requesting Dracula, they meant the version I'm used to.
Yeah, so where the situation really gets weird is that it's likely that a lot of them were requesting Bauke's theme. Quite a few people have been using his theme, and it's probably their first exposure to a theme named "Dracula". That means there's a whole group of Tildes users that, when they think of the Dracula theme, think of one with that lighter background. Everyone else in the world thinks of one with a very dark gray (and that's the one I added).
I'm not really sure where I'm going with this. Relativism. It's weird.
Can confirm, I am definitely one of these people whose first exposure to the Dracula theme was Bauke's implementation and only later learned about the official variant when trying to add to colours of the theme myself to another site.
I actually really like Baukula idea even if it was made in jest, because I too find the soft lavender colour for the main body of the site to be easier on the eyes, almost like a compromise between solarized dark and a regular dark variant.
At a later date when all the more important issues of the site are addressed and all the main themes implemented, the addition of Baukula could maybe make some sort of a nice easter egg :)
You mean "coders" think of it that way. :) A lot of people likely only see it, as you said, as a theme written by a guy with the username Bauke. It's fine either way; I'll continue to use Stylus on desktop to use Bauke's theme, and use the native Dracula on mobile. It works out okay. But I wouldn't be opposed to seeing a "Baukula" added at some point.
Fwiw, I very much think of this version you implemented, and am very happy to have it here now.
Congratulations, @Deimos! After taking a little break from Tildes, I came back this week to find it thriving. This announcement reminded me to finally get off my duff and become a Patreon supporter. Keep up the great work, and thank you for creating this community.
How is Tildes doing on donations for server costs vs. anticipated growth/bandwidth costs when factoring in the incoming increase in traffic?
@talklittle got it exactly right—the site costs are easily covered at this point, several times over by the Patreon pledges alone. The current server should be more than we need for a long time. I tried a brief test today of 200x the usual current traffic to the front page (so about the 10,000% increase he mentioned) and it seemed (mostly) fine. And that's without even particularly trying to make it suitable for high traffic yet. The server costs shouldn't increase significantly for a very long time.
So like he said, the next real funding milestone will be reaching a point where there's enough money coming in that it's reasonable for me to consider this a "real job". If the interest in the site keeps increasing, I think I'd probably like to make that goal more prominent, maybe with a visible monthly donation goal meter somewhere. Also, now that the site's no longer private, I'll be more comfortable applying for some grants and seeing if anything comes from that (it felt a bit weird to try to do it when they couldn't easily access the site).
Regardless, there's not really anything to worry about. The expenses are more than covered, there aren't any investors to breathe down my neck, there are no deadlines. Absolute worst case is that I treat this more like a side-project for a while or take some contract work on the side while it ramps up.
If I recall from the last few times this was talked about, the server is underutilized to the point where Tildes could likely handle a 10000% increase in activity before worrying about that. The recent spike was probably only ballpark 50% increase over the baseline. Bandwidth is factored in as it's not a cloud hosted server.
The more pressing problem is Deimos needs to pay rent and buy food etc. which dwarf current server costs. He said he might apply for grants for some added income.
Yeah, from a funding perspective, getting Deimos a salary is the next priority. It's the difference between him doing Tildes full time, and doing Tildes as a side project while he takes other gigs. This is a five star developer who would normally command a six figure salary and he's being incredibly generous donating two years(ish) of time to this so far.
Long-time reddit user here looking for alternatives. I've just today made use of an invite (late to the party) and have a small group of meta-theory/techie friends who'd love to join. I'm mentioning this because it seems I've just missed the '5 invites top-off' mentioned in OP. Hook a dude up?
Really like the site, btw.
Send a PM to Deimos and he'll load you up with a nice big satchel of invites.
Sweet, will do. Thanks!
Congrats on hitting this milestone!
I wonder why Google is indexing Tildes.net like it is. It seems to healivy favor the user pages vs the topic pages.
Vs. say HN, which lists the post pages prior any user profiles:
This seems to be not ideal for Tildes, right?
Edit: I’d also like to add a huge thanks to @Deimos for creating this site and codebase, and funding it all like he has. Huge respect.
Edit2: and I assume a blank site: search’s result rankings are an indicator of google’s rankings of the site internal pages, correct? Or does that not matter at all in real world use cases?
Google rankings for the results pages of newly crawled websites are always weird and scattershot-like to start but usually start looking like how you would expect them to after a few months. They just need time to collect more related data (e.g. where user traffic on the result pages is more frequently going, where third party links to the site are most frequently pointing, etc) to help them determine the "proper" result order.
Thanks, I was wondering about the timing factor as well.
I wonder if Google is going to have a hard time collecting user data on Tildes as so many of us attempt to not allow tracking :). I imagine Chrome is a large vector for this data?
Or maybe I misunderstand where google gets this data?
Google's ranking algorithms are entirely black box, so this is really only speculation, but yeah I would assume data collected via chrome specifically is not really a factor in determining search result rankings... just general traffic stats, which they can track basically no matter which browser you're using or how privacy-centric you are since, regardless of that, your browser still needs to send GET requests at the very least and likely also sends referrer data (unless you have specifically stripped/obfuscated that), etc. to google when you use their search, which they undoubtedly monitor.
Thanks! I've made Dracula my new default.
Happy to see this community grow as it nears the 1 year mark.
Congrats guys! I myself just signed up a few days ago and I love the community that this provides.
I like the theme, it's nice and easy to read and Congratulations!
The big day has arrived! I'm really happy to see this happen, thank you so much for creating this place!
My hope is that being publicly visible will allow better conversion of potential interests into active users.
I usually prefer light themes but the new dark one looks very nice. Now if only the mobile Firefox tinted the status bar accordingly, it stays white. It's not Tildes' fault, mobile Chrome gets tinted beautifully with the new theme.
I really like the new theme. Good job.
By the way, not sure if this was discussed before, but having to scroll all the way down to write a comment can be a bit annoying on posts with many comments. :-P
This has been discussed many times! :)
The answer is always the same:
it's a deliberate design choice;
it's to encourage people to read other replies before replying themselves;
it's not going to change.
There was talk at one point of allowing the new top-level comment box to be at the top for certain, very specific topics, e.g. feedback threads like this one where the more input the better, regardless of repeat comments. But, yeah, overall it's probably very unlikely to change for the vast, vast majority of topics.
Who cares about minority cases? Pshaw! Generalisations all the way!
There's a downside to this though. On long pages and specially on mobile, you have to scroll a lot to reach the comment form. Just open this page on your phone.
There's a downside to any design choice. Deimos has chosen this downside for this choice because of the upsides he believes it brings.
If you want to have a discussion about the placement of the comment textbox, please feel free to post a new topic in ~tildes to raise your points. There are other people here who agree with you, and they might welcome the chance to share their views.
Fair warning, though: Deimos has heard all these views before, the many previous times this has been discussed, and he hasn't changed his mind so far. You'd need to come up with a totally novel approach, or demonstrate a very high level of popular support for your point, if you want a different outcome this time.
I'm fine with the decision. I just wanted to share my positive view about the new theme and had to scroll all the way down.
Maybe having the form at the bottom helps reducing comments like the one I left, but it will also reduce the number of new comments after threads reach a certain size. That's why I mentioned this. Not a big deal for me though.
I think an important thing to recognize is that however annoying it is for you to scroll down to the form, that's how annoying it is for everyone else to scroll down enough to see your comment - unless they're sorting by new, your new comment is probably going to be all the way at the bottom. The larger a thread is, the less "useful" a new top-level comment is, and putting the form at the bottom emphasizes that.
Yay (although I'll miss being in a super-secret community)! Also, Dracula looks good and is a nice refresh from Solarized Dark! Unfortunately Ubuntu theme is still not here.
It is if you use solarized dark in your terminal that runs on Ubuntu :-)
Curious @Deimos how has traffic been since being publicly visible?
It's up a pretty good amount. A fair amount of it is bots (from search engines and other services) scraping the site though, so it's hard to say exactly how much it's changed. Around half of the site's traffic is generally logged-out right now, but I haven't tried to separate the bots out from that (I could, but it's mostly not important to worry about).
Thank you for Dracula.