Discourse is slow on Android. Why?


  • Banned

    @cartman82 said:

    The feeling on TDWTF is that this is a jungle. And that you should stop tinkering with interface and new features, until you get the number of bugs down to a manageable amount.

    One thing you can do to help us out here, is make sure you like the bugs that are affecting you, it helps float the ones that are bothering you most to the top of our list.


  • Grade A Premium Asshole

    All dickweedery aside, that seems like a counterintuitive response. Why would we "like" something that is negative? That seems like those out of place likes that inevitably surface when someone posts something about someone's dog dying on FB.



  • @Polygeekery said:

    All dickweedery aside, that seems like a counterintuitive response. Why would we "like" something that is negative? That seems like those out of place likes that inevitably surface when someone posts something about someone's dog dying on FB.

    It's part of the way they dog food Discourse into being their bugtracker. More likes = higher priority, in theory.



  • @abarker said:

    dog food Discourse into being their bugtracker

    TRWTF, as it has been from the beginning.


  • Grade A Premium Asshole

    So we should REALLY like it when private user information gets dumped into the JSON feed? I don't like that. But I have to like it if I want it fixed.

    Discologic.


  • Discourse touched me in a no-no place

    @sam said:

    One thing you can do to help us out here, is make sure you like the bugs that are affecting you, it helps float the ones that are bothering you most to the top of our list.

    The one that irks most right now is the way that things hang occasionally, making the browser window stick in the “Connecting…” state. That might be a deployment bug though, or some bad interaction between the software, the deployment and some users/bots. (I include “problems with the hosting” as deployment bugs.) No idea how to report it, or whether to report it to you since I've no idea where the fault lies.

    But it really stinks for usability fail. ;-)



  • @Polygeekery said:

    So we should REALLY like it when private user information gets dumped into the JSON feed? I don't like that. But I have to like it if I want it fixed.

    And in those situations, we should probably create sock puppets to artificially inflate the likes on those bugs to really make sure they get fixed.


  • Grade A Premium Asshole

    This seems like a good use for a bot army.



  • @Accalia! Make the bots like all the bugs! Please.


  • FoxDev

    @sam said:

    make sure you like the bugs that are affecting you

    that concept never fails to give me a bit of cognitive dissonance....


  • Grade A Premium Asshole

    I would like to suggest a naming scheme. MrBurns0001 through MrBurns9999. No need to be subtle. ;)


  • Grade A Premium Asshole

    @accalia said:

    that concept never fails to give me a bit of cognitive dissonance....

    It is one of those statements that gets in your head and could potentially cause an aneurism.


  • FoxDev

    hmm... tricky.....

    could you open a github issue so i can keep track of the outstanding bot requests?


  • BINNED

    @accalia said:

    could you open a github issue so i can keep track of the outstanding bot requests?

    Only if we can like the github issues!


  • Grade A Premium Asshole

    @Onyx said:

    Only if we can like the github issues!

    You should be able to, how else would you know which issues are most pressing?


  • FoxDev

    well there isn't a like but you can star them i think.....


  • Grade A Premium Asshole

    Is a star > a like? More cognitive dissonance.


  • BINNED

    I never understood stars. What does it mean when I star a post, or and issue, or a website even? Am I commending it?

    "Good issue, here, have a gold star!"

    Is that where the Opera devs got their fear of bookmarks recently? Then again, it's not even starring in Opera, it's stashing! The fuck was wrong with bookmarks? Did people forget what a book is ALREADY?

    (To be fair to Opera devs, they somewhat fixed bookmarks after much whining from the userbase).


  • ♿ (Parody)

    @Onyx said:

    Did people forget what a book is ALREADY?

    NERD ALERT


  • Grade A Premium Asshole



  • In that case, I claim MrBurns0000. After all, I need a bot to do my bug liking for me. ;)


  • BINNED

    @boomzilla said:

    NERD ALERT

    *looks around*

    Yup, still among the people capable of starting a flame war about where curly braces should be placed. Not concerned with that label.


  • Banned

    Bringing this back on topic, a brand new Nexus 9 with the new hotness Nvidia chipset and Android 5 pre installed scores.. 750ms to 850ms. (Mine does not arrive until Friday but others posted results)

    That's about on par with ye olde iPhone 4s, I'd estimate? Maybe a bit better since the 4s would perhaps do 1000ms, 2x as fast as the 4.

    http://what.thedailywtf.com/t/discourse-is-slow-on-android-why/4417/112

    The bug report did get updated a lot by multiple members of the chromium team:

    Here is hoping they can un-suck complex JS perf on android one of these days.


  • Grade A Premium Asshole

    @codinghorror said:

    Here is hoping they can un-suck complex JS perf on android one of these days.

    You can sit around and wait for someone else to fix your problems, or you can un-complex your JS. Let the browser do the things that it does and you don't have to worry so much about JS slowing down your application.

    Android is pretty dominant in world market share on mobile due to their lower-priced offerings. So most phones out there will have worse performance. You can either fix it yourself or wait on others to hopefully do it for you. I would suggest you bring on a dedicated mobile developer and stripping down some functionality for mobile.


  • Banned

    Nah, I'd rather urge Google to fix this so the whole world benefits from non-awful complex JS performance, forever, on all Android devices. The more we publicize this, the more likely that outcome is.

    Read the updates to the bug if you haven't.

    As you can see there has been a flurry of activity already thanks to us publicizing these inexplicably hideous, awful results on Android. Consistently awful for a year now...

    I like to take the strategy that results in the best long term outcome for everyone. Short term workarounds and hacks don't appeal to me.


  • Grade A Premium Asshole

    I am unimpressed, a scroll through the comments shows a lot of references to Discourse sites and Ember.JS, so it looks like you are causing the stir. It still looks like you are trying to get someone else to fix your problems for you.

    @codinghorror said:

    Short term workarounds and hacks don't appeal to me.

    https://github.com/discourse/discourse would seem to show otherwise. It is all a matter of interpretation though.


  • Banned

    I find your interpretation fascinating! Here's a direct quote from a chromium team member who works for Google

    I took a quick look a deopts at it seems that optimizations really never stabilize - I might be wrong but it seems like we are hitting the same issue I already reported for angular: first several times function deopts due to polymorphism as we inline more and more alternatives, then it stabilizes - but at this point GC comes, collects maps and optimized code and type feedback are thrown away due to "weak-code" dependency and the loop starts from scratch.

    I bolded a part you might find particularly instructive. You can look up Angular and who sponsors that project:


  • Grade A Premium Asshole

    You have a point, but the part that I found relevant while scanning through it was this:

    Which of those benchmarks are you running? There seems to a bunch of DOM related things in those tests which are not just JS.

    While it may be an issue for you and others using massive JS frameworks to re-implement a lot of what a browser does natively, I do not notice any slowness issues on other sites that I frequent on mobile. I consider it to trivial. I care fuck-all about benchmarks. All that matters to me is UX.

    Also, while I have you, I would love it if you could fix the bug that breaks multiple copy-paste in a single reply on mobile Chrome. The compose window just shits itself and highlights all of the text when you try a second paste. It never triggers the paste function of Android due to this.



  • Even if google finds the problem and fixes is, how long will it be before it is in the hand of majority of users? One year? Two years?

    IMO you need some compromise in the meantime.



  • I ran the Complex List test on my Jolla, and the results are as follows:
    Geometric mean: 3482.82ms, Mean: 3476.6ms, Std. Dev: 161.07ms, Max: 3746ms

    I can test on a Lumia 920 later.


  • BINNED

    @ender said:

    I ran the Complex List test on my Jolla

    OT: Saw the site/video a month ago or so. Looks interesting. Is it any good?



  • @Onyx said:

    Is it any good?

    It annoys me a bit less than Android, WP and iOS.


  • 🚽 Regular

    I don't do this kind of web-dev but it seems like these ultra-heavy-weight frameworks in general are an issue themselves due to exceeding the use-cases that the JS VMs were intended for.

    Terrible Car Analogy: It's like strapping a Saturn V to a car's roof-rack because you like the extra capabilities, do you really get to complain about Ford when the wheels pop off? Maybe waiting until cars have heavy-lift rocket compatibility (or mitigating the issue yourself, as suggested) is a better idea...


  • Discourse touched me in a no-no place

    Let's not get into terrible car analogies again.


  • ♿ (Parody)

    @codinghorror said:

    Nah, I'd rather urge Google to fix this so the whole world benefits from non-awful complex JS performance, forever, on all Android devices

    Can't you do both? Why punish your users who may or may not be able to upgrade even after google figures this out (assuming they do). What about the next performance problem? This just comes across as hating your end users.

    Personally, I don't care, because I have an old dumb phone and I like it, and I pretty much only do discourse on my powerful desktop machine.



  • @ender said:

    I ran the Complex List test on my Jolla,

    Jollo?!



  • @FrostCat said:

    @codinghorror said:
    Yep, this is the GamerGate reply.

    Wow, label much?

    Hey, look, everyone, @codinghorror can be a dick too! One of us, one of us!

    It's standard practice when bringing up this subject. See, you have two sides, the side who noticed something unethical and is mad about it, and the side who is mad that the first side includes people who don't like them.

    The standard practice of that second side, then, is to demonize an entire movement based around ethical concerns, and label anyone who disagrees with them on anything else as part of said movement. They don't seem to realize how mad this makes people, and tell continuously tell them that it's their own fault for all being evil.

    For someone who apparently likes rational discussion, siding with a group of people who refuse to speak with anyone who disagrees with them and (wrongly) labels all of them with some of the vilest possible terms in our modern culture is an odd move.



  • This post is deleted!


  • I'm fairly sure he meant 'on discourse' anyway.



  • I meant specifically in that test. I'm very happy with the performance in general.



  • As an addendum, I support GamerGate for exactly two reasons:

    1. I'm making a game in my spare time. I'd like to sell it one day. I don't want to have to hire a female PR person just so that I can use her to buy good reviews from 'journalists' with her body. That is simply disgusting. It needs to not be a thing now or ever again, and I find it offensive that it has been in the past.
    2. I hate seeing people who take part in the movement for reason 1 or any other valid concern labeled as misogynist bastards for it.

    So @codinghorror: Seriously, man. Think before you try to use being concerned about the treatment of other people as an insult. You cannot apply labels to whole communities. You call this one toxic. If you really think so, you apparently have never visited any other online community. People here are very open when they dislike things, and troll each other endlessly. Yet never once have I found this community to be remotely interested in mercilessly attacking someone. You may feel attacked, but the forum was moved to your software against everyone's wishes. You will receive many pointed comments for this, because your software is very much broken.


  • ♿ (Parody)

    @Magus said:

    So @codinghorror: Seriously, man. Think before you try to use being concerned about the treatment of other people as an insult.

    I think you're reading too much meaning and coherency into what he said. Imagine it was a LOLCAT meme instead and you'll get the right idea.


  • ♿ (Parody)

    More importantly, how does discourse perform on this:


  • BINNED

    @Magus said:

    The standard practice of that second side, then, is to demonize an entire movement based around ethical concerns, and label anyone who disagrees with them on anything else as part of said movement. They don't seem to realize how mad this makes people, and tell continuously tell them that it's their own fault for all being evil.

    The misandry thread is ⬅ ⏫ 🔢 🔄 over there.



  • But can it play Crysis Minecraft 1.8?



  • I'm not even trying to cry misandry, because while I do think it's a problem in some cases, the bigger one is that people think it's okay to call anything they personally dislike 'misogynist'.

    I just wish people would stop trying to bring sexism into discussions that have nothing to do with it, and also stop calling people sexist for calling their sexism out. This spiral is insane.


  • BINNED

    I'd say there's quite a bit of overlap between the two things.



  • Honestly, both sides have valid points. This is because there are people screaming 'misogyny' and people screaming 'misandry' who are clearly more interested in making each other mad, fearful, and (preferably) suicidal than actually dealing with issues.

    It's one thing to write letters or post in a forum about problems you have with a situation. It's another thing entirely to shout people down and reject anything they say because they don't think you look very oppressed.

    If you aren't pursuing a path that is remotely capable of accomplishing anything, it's probably not worth continuing. All you're doing is making more people feel oppressed.


  • I survived the hour long Uno hand

    @Magus said:

    I find it offensive that it has been in the past

    You realize that didn't happen, right?

    I'm all for an examination of ethics in video game journalism. Let's start with the relationship between a company spending advertising $$$ on a given publication and the scores their games subsequently receive. But if you're trying to excuse people behaving like scum, you're not likely to drum up much support.



  • I'm talking about what 'journalists' have said, and have no confidence for Grayson's 'but it was like two days later, so it's alright'. I'm talking about the fact that there are no ombudsmen in games journalism.

    People always behave like scum. I don't intend to excuse that in the slightest. What I am against is being told that I have treated people like scum because I don't like what's going on. I don't like people saying that all of GamerGate is male, or that all of GamerGate is misogynist despite that frequently being disproved. I don't support the attacks against the Fine Young Capitalists, which many people on the opposition cheered for.

    It needs to stop, and telling everyone that anyone who disagrees with you 'behaves like scum' is quite frankly exactly what I have a problem with. Both sides are doing it.


Log in to reply