Over the last few months, we’ve been ramping up the creation of new articles for the EA Forum Wiki. (Some of these articles can be applied to posts: we call those “tags”.)

Today, we’re starting the EA Forum Editing Festival. The festival is an all-out blitz of tagging and article editing, with the goal of:

  • Applying as many useful tags as possible to as many posts as possible, so that tags become a much better way to find relevant content.
  • Creating and updating Wiki articles, so that someone interested in a topic can reliably get a good summary and know what to read next.
  • Improving the way the whole system is organized — including the layout of our tag portal, which tags even exist, and the way posts are ranked within individual tags.

To kick things off, we’ve released an update to the tag portal, showing almost every article on the Forum (not just tags — we may change the title) in a collection of tables. 

While the new portal isn't complete, and we may end up reorganizing it entirely, we hope it inspires you to find an article to edit, create a new article, or suggest improvements to the portal itself.


People often say they’re looking for a way to do some small, useful thing for the community in their spare time.

This is one such way. You’ll make the Forum better with every tag you apply, every edit you make, and every suggestion you share.

How long is the festival?

One month. Until Friday, May 7th, we’ll consider any edits, tags, etc., to be part of the festival. This means they’ll make you eligible for fabulous prizes.

If you’ve already been doing lots of editing and tagging — good news! Your past efforts have not gone unnoticed, and we’ll consider them when prizes are given.

Why hold a festival? What’s with all the hullabaloo?

At the moment, most of the tagging and article editing that now happens on the Forum comes from a very small number of people. 

We really appreciate their work, but we also want to get more people in the habit of adding tags, editing articles, and generally taking part in the crowdsourced bits of the Forum.

So we’re writing a loud, flamboyant post, pinning it to the top of the frontpage, and giving away fame, glory, and money. We hope the festival will help the Forum settle into a new pattern of broad participation.

If you don’t do much editing or tagging yet, now’s the time! See below for ideas on how to start.

Why is this important?

Improving the Wiki

See my previous post. In short, we’d like people to have access to a collection of material that sums up lots of important bits of EA knowledge: philosophy, cause areas, organizations, even memes. Pablo is working full-time to generate and organize content, but he can only write so much — a really good wiki has to be a community effort.

Tagging posts

There are new posts on the Forum every day. And every day, it becomes just a bit harder to catch up on everything that came before, and a bit harder to feel as though you’ve read the “right” content on a given topic.

Tags make this better. Ideally, a tag will:

  • Summarize a concept well enough to give someone a basic understanding
  • Gather together many posts that can help them learn about the concept
  • Sort those posts according to their relevance

I’ve had many people tell me they feel overwhelmed when they try to explore EA. Our massive backlog of content has been a barrier. But properly tagged and sorted, it can also be one of our greatest strengths. And you can help to make that happen!

How can you help?

As I said before, every tag and every edit helps. The Forum Wiki has no perfect articles, and few posts have every tag they should.

That said, here are some easy ways to get involved:

Tagging

  • Tag your own posts with as many relevant tags as you can.
  • Visit the list of posts with 25+ karma and no tags. Tag some of them.
    • Update: People have now tagged almost all of those. But you can take the "Threshold" number at the end of the URL and change it to see untagged posts with any minimum karma number you choose (20, 15...). Consider setting a lower threshold, then tagging any posts you find.
  • See the list of new tags at the top of this page. Choose one. Add it to relevant posts.
    • Some tags are older, but only recently became usable as tags (before, they were "wiki only"). Those tags will have (null) next to their names at the bottom of this page. Adding those tags to posts is also very useful!
  • Choose an organization. Add that organization’s tag to every post about their work, every cross-post from their website, etc. (If they have no tag, create one!) You can find these posts by entering the org’s name into the searchbar.
  • Choose a tag. Search for the tag’s name, and apply the tag to every post that mentions it (if you think the tag actually belongs on those posts).
  • Go to a tag’s page. Upvote the articles you think are most relevant for that tag, so that someone interested in the tag will see those articles first. We see every tag vote, so we notice and appreciate people who do this!
  • Create a tag you think should exist, but doesn’t. Apply it to whatever relevant posts you can find.
  • If you want to use an existing tag but can't use it on a post, that means it's set to "wiki only". Send me a message and I can remove that status. (We've applied it to most of the new articles by default, so that the tag menu remains navigable.)

Editing

  • Look at the list of articles. Find one that you know a lot about. Go to that article, and add at least one useful sentence. Repeat for more sentences and/or more articles.
    • At the moment, many articles are single-sentence stubs. If you want to add more detail, those are your low-hanging fruit.
  • Edit articles so that they conform to our style guide.
  • Leave comments where you notice issues or chances to expand an article (even if you don’t want to write the additional material yourself). Pablo (and other editors) will be looking for comments like these to guide their work.
    • Examples of issues you could raise:
      • Possible title changes
      • Splitting one article into several
      • Merging several articles into one
      • Revisions or expansions to content
  • Pick an article and add some related articles to its description. Here’s an example.

Suggestions and sorting

  • Look at the list of articles. The table that sorts them out is very messy. If you can think of a way to make it less messy, whether that means a totally different categorization system or a single post moved elsewhere… tell me!
  • Find articles that you think should be merged, or deleted entirely. Tell us about them. There’s a good chance we’ll agree, because we’ve mostly been taking the approach “build everything, then see what doesn’t need to be there in retrospect”.

What are the prizes?

Once the festival ends, we’ll read through:

  • The edit history of every article
  • Every application of a tag (including upvotes after a tag has been applied)
  • The suggestions people have made for individual articles, the article page, etc.

If we find you, you’ve earned a place of honor on the Forum: In our festival wrap-up post, we’ll list every single contributor in alphabetical order, with a separate list for prizewinners. 

If we see that you’ve done especially good work, or a high volume of generally useful work (e.g. applying tags, which is hard to do “especially well”), we may award you a prize!

Prize amounts will depend somewhat on how many contributors stand out, and what each of them ends up doing. But there will be:

  1. At least a dozen prizes,
  2. Worth at least $100 each, from a total prize pool of at least $5000,
  3. In the form of donations to an EA Funds-eligible charity of your choice.

I'll err on the side of giving out more prizes, so your odds may be better than you think!

Also, there’s a small chance that the Forum might have profile badges someday. If we do, there will be a “festival champion” badge, and it will look so cool. (The “festival participant” badge will also exist, and will also be cool.)

Let the festival commence!

Denizens from every corner of effective altruism flock to the once-in-a-while Editing Festival.
 

Remember those easy ways to get involved?

Why not pick one, then mark a time on your calendar to do it?

Or you can wait for the mood to strike you and go on a crazed late-night tagging spree, which is something we’ve only heard of and has definitely never happened to us.

Either way, we hope you enjoy the festival — both as a participant, and as a beneficiary of the better-organized Forum we’ll soon have.

97

29 comments, sorted by Highlighting new comments since Today at 6:32 AM
New Comment

Number of tags applied to posts the day before the Editing Festival was announced: ~30

Number of tags applied the day after the announcement: ~300!

Good start, everyone 😊

Maybe I'm missing something, but it seems like the wiki isn't labelled as such - as in, there isn't a part of the site called the 'wiki'. There's also the 'tags portal' which refers to the 'EA Forum Wiki', but as I understand it that page essentially is the wiki. The language is confusing.

Should there be a section of the site called the 'wiki' that lists all these pages? Or maybe even consider renaming tags to wiki - where posts on the forum can be 'tagged' with a wiki article.

(The URLs for tags should probably more conventionally be in the format /tags/<tagname>, not /tag/<tagname>. Going to /tag gives a 404.)

We've gone with a different naming convention than LessWrong (they say "Concepts" while using the same "tags" URL), but given the amount of code our sites share, it will take some time to disentangle our terminology from theirs. 

I'd like our "Tags Portal" page to eventually have the name "Wiki", and people should think of that page as the "homepage" of the Wiki. 

Or maybe even consider renaming tags to wiki - where posts on the forum can be 'tagged' with a wiki article.

The "Tags Portal" contained a section I'd titled "Articles" but have just retitled "Wiki Articles". The first words of that page remain the same:

This page displays a list of articles in the EA Forum Wiki. 

Some of these articles are also tags that can be added to posts, so that people can find posts on certain topics. You can upvote or downvote a tag for a given post to move it higher or lower in the list of posts on that tag's page.

Note also the second sentence of this post:

Some of these articles can be applied to posts: we call those “tags”.

So we have similar intuitions, but I use the word "tag" to represent an article that can be used as a tag, because saying "I tagged this post with article X" also seems confusing.

We're still trying to figure out the extent to which having "wiki only" articles make sense (and what fraction of articles don't work as tags). If we end up making all articles usable as tags, the distinction between "article" and "tag" disappears, which will lead to more terminology change.

Love it! I'm in 💃🕺💃🕺

Well, I'm out.. did basically nothing, sorry! 😊 (It's a good thing that editing the tag-wiki is still available after the Festival) 

Is there a way to show my appreciation for an edit? 

Often I see excellent edits[1] to the Wiki show up in my Forum homepage, and I would like to be able to show my appreciation to someone[2].  Ideally with low effort and without otherwise adding any value.

Is there a like/upvote button for Wiki edits I'm missing?

--

[1] For example, check out how much information this article on iterated embryo selection is collating and condensing. It was written a few months ago, and is now Google's featured snippet for iterated embryo selection (a sign that Google 'thinks' it's the best, succinct summary of the term).

[2] To be honest this is so frequently  Pablo or the EA Wiki Assistant (I think also Pablo?) that I should probably just send a DM.

Voting on edits is recently in the pipeline.  In the mean time you can comment on the tag, which gives the author public recognition.

Awesome, glad to hear that! Thanks, JP!

Just started by tagging this post :)

For what it's worth, a lot of old posts don't have tags, or have some tags but don't have ones that weren't around when the posts were created. It'd be great to see a lot of these posts get tagged to make them easier to find.

That's exactly the reason I linked to this list of high-karma, untagged posts as a starting point for people to work on. (And... it's already down at least 80% from where it was — good job, people!)

By changing the "threshold" number at the end of that URL, you can see untagged posts at any level of karma. That's a good way to find posts that may be especially worth tagging. (I've added this notice to the "how can you help" section — thanks for the suggestion.)

That seems like a good idea to me. One good starting point would be tagging the posts linked to from Saulius's Collection of good 2012-2017 EA forum posts. (Obviously there are also many other ways one could get started tagging old posts.)

  1. I found tagging buggy. I tried to tag something yesterday, and I believe it didn't get through although it worked today. The 'S-risks" tag doesn't show up in my list to tag posts at all, although it's an article. But that might also be something about the difference between tags and articles that I don't understand? I use firefox and didn't check on other browsers.

  2. Is there a consensus for how to use organisation tags? Specifically, is it desirable to have have every output that's ever come out of an organisation to be tagged to them or only e.g. organisational updates? I've seen the first partly, but scarcely, done and am not sure about my opinion. (I mean things like "This report is published on the EA forum and the person who worked on this report was at org X at the time and wrote it as part of their job")

edit: 3) Just adding this on here...Is there a way to tag everything that has one tag with another tag? (I'm speaking of the 'economics' tag + lots of more specific tags; 'moral philosophy' and 'metaethics' etc.)

  1. Sorry about the issues. On S-Risks, it is a wiki-only tag, though probably we should change that.
  2. I really like the idea of tagging everything that's been officially produced by an organization with the organization's tag. So you might go to the Rethink Priorities tag, sort by top, and see a "best of" list.
  3. [Edit reply] Not to my knowledge, sorry.

I think Chi's point 3 suggestion would sometimes be helpful, and even more so if we could somehow sort-of pre-select a batch of posts for giving tag X to, but then scan through the list and un-select some before the tags are applied. This could be like how many sites (e.g., gmail) let you click one box at the top of the list to select all items in that list, then individually unselect some. 

And ideally the pre-selection could be for all posts with a given other tag, all posts by a given author, or something else or combos. (E.g., I'd have used this for tagging most Aaron Hamlin posts with the Center for Election Science tag.) 

On 2, I share that view, and I'd also add that I think "organisation tags" should also be applied to things about but not by the org. E.g., I think donation writeups that discuss why the person donated to orgs X and Y and considered but ultimately decided against donating to Z should be given the tags for orgs X, Y, and Z. And I think someone's attempt to summarise and critique an org's theory of change and recent outputs should be given that org's tag, even if the person doesn't work there.

My thinking is that the same people interested in posts by an org will often also be interested in posts about the org but by other people.

But I think we shouldn't do this when a post only includes a very small bit about a given org (e.g. the posts Aaron Gertler and David Nash make which give updates about many orgs at once).

I think it might be good to have a clearly visible policy about how organisation tags are to be used. This goes especially if the norm I suggest is indeed adopted, since in that case we wouldn't want people automatically assuming that all post tagged with org X were by someone from org X writing in relation to their work for org X.

I think of tags as being for "posts that involve X in some way", which encompasses posts written by and about a given organization.

I do think that org update posts are a good way to use an org's tag. If someone is interested in The Humane League, they might want to see what THL was doing in a given month. It's easier to use the tag for this than to make someone filter through all the monthly update posts to see which ones mention THL. (The downside is that many posts tagged with an org won't have much info about it — are you worried about that kind of tag use not being relevant enough?)

The case when I wouldn't use an org is when that org's work is very briefly referenced in a way that doesn't have much to do with them (e.g. someone cites an 80K problem profile as a source for some claim — that doesn't seem like a statement "about" 80K).

As this conversation continues and I arrive at a firmer definition of an org tag policy, I'll try to make it clearly visible in a few places.

I do think that org update posts are a good way to use an org's tag. [...] (The downside is that many posts tagged with an org won't have much info about it — are you worried about that kind of tag use not being relevant enough?)

I think the main downsides I see are that:

  • There are just so many of you and David Nash's org update posts, and it seems like many of the orgs are mentioned in almost every one of them, and each org is only given something like 1-3 paragraphs.
    • So it seems like if we tagged all of them with every org mentioned for at least a paragraph in them, that'd sort of "clog up" those orgs' tag pages
      • But I guess that that problem is reduced by the fact that those org update posts seem to usually get less karma than the average post from/about an org, so they wouldn't show up right at the top of the org's tag page
    • And it'd probably be systematically less useful to someone who wanted to learn about the org than most other things that have the org's tag
      • Though I guess it might be similarly useful per relevant word, so if people realise they should just read the relevant section if that's all they care about, then maybe that's ok
  • Each of those many-org-update posts mentions probably over 10 orgs (I haven't counted), so every one would have over 10 tags, and that just seems perhaps a bit much
    • But I don't think this is actually a problem; it just might look slightly weird

But as became clear to me when I was writing this comment, the second downside just seems "slightly odd" rather than actually bad, and the first downside doesn't seem major. So I think I'd still vote to have a norm against using org tags for those many-org-update posts, but now it's just a very very weak vote. 

On point #2: Here's what I had suggested on my post about organization tags:

Choose an organization. Add that organization’s tag to every post about their work, every cross-post from their website, etc. (If they have no tag, create one!) You can find these posts by entering the org’s name into the searchbar.

In general, I think it's better to have more posts tagged rather than fewer, and I'd consider "paid work by an employee of X" to be "work paid for by X" and thus, in some sense, "the work of X".

Currently, the tags portal has 3 columns, and seems to be arranged as if I'll read the first cluster in the left column, then the first cluster in the middle, then the first cluster in the right, then the second cluster in the left, etc. Basically, left to right then top to bottom (like in a book), rather than top to bottom in a column then on to the next column (like in a newspaper). 

I say it seems to be arranged this way because that's how it's alphabetically ordered and because that's what would make "Global Catastrophic Risk (other)" come after "Global Catastrophic Risk (artificial intelligence)" (which I assume it's meant to, otherwise what "other" means would be unclear). 

But since each cluster is read top to bottom, I found myself naturally reading down to the next cluster within the same column, and so on, like in a newspaper. This meant I didn't notice that the clusters were arranged alphabetically and instead just saw them as random, and that I was confused about why "Global Catastrophic Risk (other)" seemed to come before "Global Catastrophic Risk (artificial intelligence)".

I think I'd recommend assuming the reader will scan the portal like they'd scan a newspaper, though I'm not sure. 

[This comment is no longer endorsed by its author]Reply

The clusters aren't in alphabetical order — only the articles within clusters.

The clusters are arranged according to a couple of heuristics that I value about equally:

  1. Try to make the columns of roughly equal length
  2. Have the "other" cluster near the bottom-right of the section (seems natural for that to be the last thing people look at)
  3. Have related clusters close together (e.g. "effective giving" and "career choice")

I'd prefer to have all the cluster names aligned horizontally, as on the LW Concepts page, but our extremely varied column lengths discourage that for now (this might change as we continue to add new articles, look at new ways to sort the page, etc.)

If anyone has an idea for making the page better-sorted and/or more evenly arranged, I'm all ears. Graphic design isn't my forte and the current version is quite rough.

The clusters aren't in alphabetical order — only the articles within clusters.

Oh man, you're right, and I can't see how I thought it was alphabetical except maybe that I noticed one case where one cluster came before the next one alphabetically and then didn't check my theory. Gonna pin that obvious error of mine on jet lag after a long flight.

Those heuristics sound good to me.

As for aligning cluster names horizontally, it seems like this is already done for "Organizations" by just having (sometimes large-ish) gaps at the bottom of some clusters, and it seems like that looks ok to me? So maybe just do that?

And/or you could do a vertical version of the "justified" setting that Word and GDocs offers? I.e., you could have the vertical space between bullet points be larger or smaller depending on whether the cluster is less populated or more populated than those it's horizontally next to?

Currently, the tags portal says:

If you want to suggest a change to the article portal (e.g. a different sorting system, or a particular article that should be moved), send a message to Aaron Gertler. He knows the portal is a mess right now, and he appreciates feedback.

I imagine it'd be better to also suggest people could make their suggestions publicly here (or somewhere else)? That way these could inspire further suggestions, get upvoted or downvoted, etc.

I've added a link to your "tag proposal" thread (rather than this article, which isn't meant to be a permanent resource).

Will Aaron, Pablo, or other people by default look at all newly created tags to decide where to add them in the Tags Portal? Or is there any need for people who create the tags to flag their creation to someone, and maybe make suggestions about where to put them?

I have now updated the Tag Portal so that it reflects the current state of the Wiki, and will henceforth add to it any newly created entries, so it should always remain up-to-date.

I look at all newly created tags every morning, so there's no need to flag their creation. I'll coordinate with Aaron to find a process for incorporating new tags into the Tags Portal periodically.

Currently the Cause Areas part of the tags portal has two sections for GCRs but none for longtermism. I think it's probably now more common to think about the main EA cause areas as including longtermism than to think about them as including x-risks or GCRs; i.e., the framing often primarily focuses on the long-term future, with x-risks and GCRs as big examples of what one might focus on within longtermism. And this seems like a useful way to frame it, given that longtermism can include both stuff to do with x-risks/GCRs and other stuff (e.g., non-existential trajectory changes, or speeding up progress). (Though this is complicated by the fact that one could prioritise x-risks or GCRs for non-longtermist reasons.)

So I think it might be better to have a category for longtermism and then have the GCR stuff all as part of that. 

Alternatively, there could be an additional category for "Longtermism (other than GCRs)", or something like that. 

We have two categories ("Moral Philosophy", "Long-Term Risks and Flourishing") which capture lots of material relevant to longtermism.

As for the cause area section specifically:

  • AI is its own cluster because we currently have an enormous number of articles about it. If we only had one article about AI risk, I'd put it under "Global Catastrophic Risks" and that would be that.
  • The "Global Catastrophic Risks (other)" cluster feels well-defined to me in a way that a "longtermist" cluster wouldn't. When I look at the "Other" cluster, most of the seemingly "longtermist" causes are still things that many people work on hoping to achieve substantial change within their lifetimes, for the sake of present-day people — anti-aging research, land use reform, climate change...
    • If you ask me about a cause area in that section, I can fairly confidently say whether or not it counts as a GCR. In many cases, I wouldn't be able to say whether or not it counted as "longtermist". (And as you mention, many of the areas could be prioritized for longtermist or non-longtermist reasons.)

I think of longtermism as a common value system in EA. Many causes seem especially valuable to work on given a longtermist value system, but few such causes require a longtermist value system to make sense. (But I spend less time thinking about this kind of thing than you do, so I'm open to counterpoints I might not be considering.)