{"items": [{"author": "Jess", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802", "anchor": "fb-827723273802", "service": "fb", "text": "Great changes. But I'm not noticing a big difference with the comments. I guess the indentation has been reduced slightly?", "timestamp": "1480267693"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827723867612", "anchor": "fb-827723273802_827723867612", "service": "fb", "text": "→ The comment text is larger. Here's what it looks like without the change:", "timestamp": "1480268061"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827724586172", "anchor": "fb-827723273802_827724586172", "service": "fb", "text": "→ Here's a live view where you can see the changes on a phone: http://www.jefftk.com/concerns_with_intentional_insights
http://www.jefftk.com/ea-index", "timestamp": "1480268136"}, {"author": "Jess", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827726731872", "anchor": "fb-827723273802_827726731872", "service": "fb", "text": "→ Ahh, good, I see now. I would maybe suggest making the font of the meta text (commenter name, date, time, etc.) larger to match the comment text.", "timestamp": "1480269445"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827734037232", "anchor": "fb-827723273802_827734037232", "service": "fb", "text": "→ How's that now?", "timestamp": "1480272892"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827734087132", "anchor": "fb-827723273802_827734087132", "service": "fb", "text": "→ (check the live-site version to see the change)", "timestamp": "1480272926"}, {"author": "Jess", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827776586962", "anchor": "fb-827723273802_827776586962", "service": "fb", "text": "→ Much better! The buttons are also too small to hit easily, both on the top ribbon and also within the comment boxes. (Sorry, should have mentioned that before.)", "timestamp": "1480293125"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827778303522", "anchor": "fb-827723273802_827778303522", "service": "fb", "text": "→ The comment buttons are hard to fix. I messed with them before and it didn't work, but I should try again.", "timestamp": "1480293763"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827778667792", "anchor": "fb-827723273802_827778667792", "service": "fb", "text": "→ Yeah, the comment buttons are sprited with http://effective-altruism.com/static/icons.png and scaling sprited images is annoying.", "timestamp": "1480294118"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827778707712", "anchor": "fb-827723273802_827778707712", "service": "fb", "text": "→ This is possible to fix, but not just by saying \"here's some extra css to use if you're on mobile\".", "timestamp": "1480294158"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827780573972", "anchor": "fb-827723273802_827780573972", "service": "fb", "text": "→ (On the plus side, though, if you were fixing this you could also upgrade the sprites to look good on hidpi displays.)", "timestamp": "1480294340"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827780773572", "anchor": "fb-827723273802_827780773572", "service": "fb", "text": "→ (Ugly voting buttons on retina are the main thing holding the EA Forum back from its destiny of greatness.)", "timestamp": "1480294377"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827904725172", "anchor": "fb-827723273802_827904725172", "service": "fb", "text": "→ I figured out how to fix the sizing for the buttons: \"zoom: 2\"", "timestamp": "1480360926"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827904765092", "anchor": "fb-827723273802_827904765092", "service": "fb", "text": "→ Possibly I should have made more of my changes with zoom on parent elements instead of finding lots of things and boosting their font sizes.", "timestamp": "1480360948"}, {"author": "Jess", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827931112292", "anchor": "fb-827723273802_827931112292", "service": "fb", "text": "→ Jeff, I'm curious: when merging two forks of a project with many differences, is the standard practice to just do a giant multi-file diff? Or is there a more sophisticated structure used?", "timestamp": "1480372039"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827723273802&reply_comment_id=827931316882", "anchor": "fb-827723273802_827931316882", "service": "fb", "text": "→ These forks are so far diverged I doubt merging them would be reasonable. Depending on motivations, I would probably pick one or the other, and then try to port over specific desired features. This would involve more or less writing them from scratch with the code from the other side of the fork as a guide.", "timestamp": "1480372197"}, {"author": "Arun", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827732530252", "anchor": "fb-827732530252", "service": "fb", "text": "I like the changes but here are some suggestions:
1) I think you should add a bit more padding between each post
2) I would increase the font-size of the Outline", "timestamp": "1480272004"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827732530252&reply_comment_id=827740289702", "anchor": "fb-827732530252_827740289702", "service": "fb", "text": "→ I thought that was hard to fix, but actually it turned out to be that android's font-boost was messing things up. I disabled that (which is kind of a hack: set max-height to a very large value) and made things bigger to compensate. Here you go: http://www.jefftk.com/concerns_with_intentional_insights", "timestamp": "1480276023"}, {"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332", "anchor": "fb-827735983332", "service": "fb", "text": "I take it that bringing the EA forum up to date with the Reddit codebase is not as simple as a git merge then?", "timestamp": "1480273670"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827739675932", "anchor": "fb-827735983332_827739675932", "service": "fb", "text": "→ Uh no, no it's not.

The EA Forum codebase is a fork of the LessWrong codebase, which forked off of Reddit in December 2008 at https://github.com/.../1a01375574f0fa089db20e5fc9881b34a1... Since then the fork has had 2,477 [1] commits while Reddit has had 7,730 [2]. Plus, it looks like the mobile site for Reddit is a completely separate codebase: https://github.com/reddit/reddit-mobile

[1] https://github.com/.../1a01375574f0fa089db20e5fc9881b34a1...

[2] https://github.com/.../1a01375574f0fa089db20e5fc9881b34a1...", "timestamp": "1480275499"}, {"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827742744782", "anchor": "fb-827735983332_827742744782", "service": "fb", "text": "→ aaaaaaaaaaaaaaaaaaa!

OK thanks!", "timestamp": "1480277485"}, {"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827743034202", "anchor": "fb-827735983332_827743034202", "service": "fb", "text": "→ I'd been assuming the fork was only as recent as the forum itself. If it's as old as Less Wrong, oh my God.", "timestamp": "1480277742"}, {"author": "Michael", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827763672842", "anchor": "fb-827735983332_827763672842", "service": "fb", "text": "→ Empirically, well-intentioned people wanting to improve the LW codebase keep bouncing off of it without getting to the point where they can do productive work =(", "timestamp": "1480287534"}, {"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827768513142", "anchor": "fb-827735983332_827768513142", "service": "fb", "text": "→ Michael, are you one of those people? If so how far did you get and what was your experience?", "timestamp": "1480289352"}, {"author": "Satvik", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=827939829822", "anchor": "fb-827735983332_827939829822", "service": "fb", "text": "→ I spent about 4 hours getting the development environment set up yesterday. Still not able to run the manual tests, and one unit test breaks.", "timestamp": "1480376438"}, {"author": "Patrick", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=828314793392", "anchor": "fb-827735983332_828314793392", "service": "fb", "text": "→ I've done a bit of work on the forum, and I've come to the conclusion that at some point in the next few years we'll have to replace it. Here are some of its problems:

* it's based on technologies and practices from the mid-noughties (Python Paste for the server, the last release of which was in 2010; a home-grown queueing framework; sprited PNGs for icons; onclick HTML attributes that submit HTML forms)
* much (perhaps most) of the code is for features not used by the forum (e.g., multiple forums/subreddits, internationalization, meetups, over-18 verification)
* the code quality is poor (huge classes; huge methods; single-letter variable names; an inscrutable database schema; function definitions in templates)

That said, it may be worth delaying the migration as long as posible. It would be substantially easier to replace the forum if someone releases a project similar enough that we can customize. (Telescope and Discourse are examples of open-source projects in this space, though they are probably not a good enough fit for us to adopt them for a new EA Forum.)

In the meantime, efforts like Jeff's are worthwhile. But I think it's not worth sinking a lot of effort into a project that we'll probably have to abandon within a few years.", "timestamp": "1480577741"}, {"author": "Paul", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=828338695492", "anchor": "fb-827735983332_828338695492", "service": "fb", "text": "→ Patrick, this is incredibly valuable, thank you! I'd love to read more about what you think about Telescope and Discourse.", "timestamp": "1480601410"}, {"author": "Patrick", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827735983332&reply_comment_id=828887480722", "anchor": "fb-827735983332_828887480722", "service": "fb", "text": "→ Paul I don't know much about them. Telescope markets itself as \"Build your own Hacker News, Reddit, or Product Hunt,\" which suggests that it may be easier to customize than Reddit was. Discourse seems good, but it's focused on discussions, not articles + comments.

Arbital is working on something too, but I don't have any understanding of how it would differ from other platforms.

http://lesswrong.com/.../on_the_importance_of_less.../dhzo", "timestamp": "1480813159"}, {"author": "Henry", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827752460312", "anchor": "fb-827752460312", "service": "fb", "text": "Is there somewhere that developers can volunteer for EA causes? I'd be happy to help with odd jobs like this.", "timestamp": "1480281918"}, {"author": "Michael", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827752460312&reply_comment_id=827773503142", "anchor": "fb-827752460312_827773503142", "service": "fb", "text": "→ https://www.facebook.com/groups/dotimpact/", "timestamp": "1480291696"}, {"author": "Hauke", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827752460312&reply_comment_id=827860159482", "anchor": "fb-827752460312_827860159482", "service": "fb", "text": "→ Sam Deere", "timestamp": "1480335017"}, {"author": "Henry", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=827752460312&reply_comment_id=827863787212", "anchor": "fb-827752460312_827863787212", "service": "fb", "text": "→ Thanks, Michael!", "timestamp": "1480338928"}, {"author": "Lucas", "source_link": "https://plus.google.com/117844313637558297240", "anchor": "gp-1480394285613", "service": "gp", "text": "With this approach, tablets (and any other larger-device browsers that respect the viewport meta tag) will get a zoomed-in version of this \"mobile layout\". Similarly, different-sized phones will get text at significantly different physical sizes. That's probably not a huge issue in practice, but it is a limitation of this technique to be aware of.", "timestamp": 1480394285}, {"author": "Lucas", "source_link": "https://plus.google.com/117844313637558297240", "anchor": "gp-1480394570814", "service": "gp", "text": "> I figured out how to fix the sizing for the buttons: \"zoom: 2\"\n
\n
\nI think background-size is the CSS WG's intended property for image sprites at non-px resolutions, but yeah, I expect zoom also works. And I think transform: scale(2) is basically equivalent to zoom: 2, if you prefer to support Firefox instead of old IE.", "timestamp": 1480394570}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1480424990577", "service": "gp", "text": "@Lucas\n\u00a0What I want is something where the browser can tell me guesses about the expected arc size of a pixel, since that's what gives the \"this is too big/small\" experience. I know the browser can't do this perfectly, since you can always hook something up to a giant screen or something, but it should be able to be pretty good.", "timestamp": 1480424990}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1480425032572", "service": "gp", "text": "@Lucas\n\u00a0Yes, it looks like \"transform: scale(2)\" is a much better choice, thanks!", "timestamp": 1480425032}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1480430194014", "service": "gp", "text": "@Lucas\n I just tried it with \"transform: scale(2)\" and it moved the buttons to completely the wrong place. I added \"transform-origin: 0 0\" and it mostly worked, except it moved the rightmost buttons to be out of the box off the screen. Here's what that looked like:\n
\n
\n
\nhttps://plus.google.com/photos/103013777355236494008/albums/6358399270252683569/6358399270840395186", "timestamp": 1480430194}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1480430236163", "service": "gp", "text": "@Lucas\n On the other hand, \"zoom: 2\" degrades gracefully to just small buttons in FF, so I think I'll stick with that.\n
\nhttps://plus.google.com/photos/103013777355236494008/albums/6358399447988056849/6358399448364160786", "timestamp": 1480430236}, {"author": "Lucas", "source_link": "https://plus.google.com/117844313637558297240", "anchor": "gp-1480436676208", "service": "gp", "text": "Fair enough. I forgot that zoom changes the element's layout dimensions (whereas transform doesn't).", "timestamp": 1480436676}, {"author": "Lucas", "source_link": "https://plus.google.com/117844313637558297240", "anchor": "gp-1480438136775", "service": "gp", "text": "@Jeff Kaufman\n
\n> What I want is something where the browser can tell me guesses about the expected arc size of a pixel...\n
\n
\nCSS \"px\" was supposed to be what you want, and it generally is fairly close if you set <meta name=\"viewport\" content=\"width=device-width, initial-scale=1\">. The downside is that you get stuck rewriting all of your layout CSS to work on the smaller screen's width, which is exactly the task you're trying to avoid in this case.\n
\n
\nI guess you're wishing that CSS had an accessor function for the computed value of an at-rule property. Then you could (in theory) ask for the value of @viewport's zoom property and use that with calc() to compute your layout dimensions. I don't expect that to ever happen, though, because it'd blow a lot of the renderer's performance optimizations out of the water.", "timestamp": 1480438136}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=829124316102", "anchor": "fb-829124316102", "service": "fb", "text": "The EA Forum change is live, and I just proposed a similar change to lesswrong: https://github.com/tricycle/lesswrong/pull/597", "timestamp": "1480959900"}, {"author": "Jess", "source_link": "https://www.facebook.com/jefftk/posts/827721691972?comment_id=829124316102&reply_comment_id=829139011652", "anchor": "fb-829124316102_829139011652", "service": "fb", "text": "→ Nice job Jeff!", "timestamp": "1480966999"}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1481749857416", "service": "gp", "text": "@Lucas\n\u200b I figured out how to do it: gate the font boosting with a media query on device-width. Now if you visit an EA forum page on a tablet you'll get the narrow-screen changes but not the enlargement changes.", "timestamp": 1481749857}]}