Discourse is slow on Android. Why?



  • @Groaner said:

    or, when the battery can't hold a charge for more than a couple hours

    You know, those were replaceable last time I checked... but sure, I'll take it.



  • @Maciejasjmj said:

    You know, those were replaceable last time I checked... but sure, I'll take it.

    The implication was that said battery is still capable of holding a charge.



  • From my experience, the GS2 is junk. But my experience is mostly my wife being happy about replacing it.


  • Banned

    @blakeyrat said:

    HEY LOOK, REGRESSIONS! For another fun demo of how crappy this forum software is, try hitting "Dismiss Posts" on the "New" tab, that's been broken for fucking months now.

    I just tried it on chrome on try.discourse and it worked, what browser are we speaking?

    @blakeyrat said:

    your shitty forum software

    I get it, some are you are upset, some hate the software and so on. I banned myself cause I just can't sit around and listen to people yelling at how my team is incompetent and software is rubbish. Its grating.

    I understand that you have your WTF persona on the super ON switch here, I am pretty sure you don't have them on most of the time in real life.

    You fucking idiot why did you serve steak again tonight don't you know, its fucking discussting

    ^ on purpose

    I appreciate what you are doing here and the bugs you find are helping make the software we are building, better, for everyone. We listen to the feedback, case in point the message bus is a lot less aggressive due to work I did on Friday based on your feedback. I thank you for this and many other things you found.



  • @Luhmann said:

    Most VoIP compression is incompatible with modem sounds

    One of my friends managed to get a Minitel to work on a VoIP line, so I know it at least works sometimes...

    I’ll probably try to connect a modem once my home Internet connection starts working again.



  • Brownies are Delicious™


  • BINNED

    @VinDuv said:

    One of my friends managed to get a Minitel to work on a VoIP line, so I know it at least works sometimes...

    I’ll probably try to connect a modem once my home Internet connection starts working again.

    It can work ... but don't expect it. It's not only compression but noise, lack-of-noise (there is always audible noise on analog lines), added noise (noise is sometimes added because users otherwise think the line is dead) on VoIP, protocols (h323, sip, ...), sometimes high tones are simply cut in an effort to cut bandwidth (this is common on certain non-voip comm links too, like microwaves but those are a WTF of their own) and most importantly timing.
    Timing is a major issue. POTS, ISDN and any other comm link type are essentially all time based. In modern words: every single bit is send on the wire in a specific order and comes out at the other end at the exact same order. PC networks are packed switched, packages could arrive at different order and the timing/delay in between packages is not given. Resulting eventually in the modem/fax tones not being correctly detected or send within the protocol time frames. The old analog protocols (even stupid DTMF-tones) work by keeping certain frequencies (tones) for certain amounts of time and then changing the tone or going silent for a time frame.
    Additionally a modern computer mostly uses a rather relative time concept: a computer tick is not an exact amount of ms. CPU throttling, multicore, virtualization, ... all those nice features will screw up timing in a slight but yet unpredictable way. That is why VMWare even has (had? it's been a few years) specific options to improve clock performance on a virtual machine. Sure sending one fax might work but virtualizing an T1/E1 (24-30 channels) fax server sounded way easier on paper then in reality.


  • BINNED

    @Luhmann said:

    The old analog protocols (even stupid DTMF-tones) work by keeping certain frequencies (tones) for certain amounts of time and then changing the tone or going silent for a time frame.

    And this shit is still used. Fortunately, I had no problems with DTMF reaching the PBX so far, but internally I just switch everything to SIP INFO and be done with it. Having problems with DTMF on a flakey network once was enough to last me for a lifetime, tyvm.

    @Luhmann said:

    That is why VMWare even has (had? it's been a few years) specific options to improve clock performance on a virtual machine.

    Asterisk has a separate component just to deal with that as well. Hell, Digium even sold (is selling?) hardware timing devices to deal with that shit. Fortunately, I only have to deal with SIP (any analog devices are conected to an ATA that deals with that shit for me).


  • BINNED

    @Onyx said:

    but internally I just switch everything to SIP INFO

    If Asterisk handles that correctly I remember other PBXs doing SIP sometimes failed at that. Not providing all or incorrect info for SIP calls and the like.

    @Onyx said:

    (any analog devices are conected to an ATA that deals with that shit for me

    I also worked a lot with these to connect servers to ze phones. The multiple PRIs where rather expensive though. Once had a discussion with an Server drone at a client where he claimed that we should replace the board as a test. I proposed to test by replacing the cheapest part of the setup. We ended up testing the board in a different server.


  • Discourse touched me in a no-no place

    @Luhmann said:

    PC networks are packedpacket switched, packagespackets could arrive at different order and the timing/delay in between packagespackets is not given.

    FTFY. Sorry, it was bugging me.


  • I survived the hour long Uno hand

    It seems like we were moving toward a workable solution for the culture shift: we would report bugs on meta.d in polite language, and you folks wouldn't come here and complain about us venting in our vent-space.

    But then we were made unwelcome on meta.d.

    We then moved toward another workable solution: reporting bugs in a bug tracker you could look at without having to visit here. How's that working?

    It seems like we're being awfully reasonable in our attempts to communicate bugs without your having to engage in a culture you find distasteful.... I'm honestly sick of hearing about how "awful" we are.



  • @Yamikuronue said:

    It seems like we were moving toward a workable solution for the culture shift: we would report bugs on meta.d in polite language, and you folks wouldn't come here and complain about us venting in our vent-space.

    But then we were made unwelcome on meta.d.

    We then moved toward another workable solution: reporting bugs in a bug tracker you could look at without having to visit here. How's that working?

    It seems like we're being awfully reasonable in our attempts to communicate bugs without your having to engage in a culture you find distasteful.... I'm honestly sick of hearing about how "awful" we are.

    ++

    +1

    Because a single like is not enough for that post.


  • ♿ (Parody)

    @Yamikuronue said:

    But then we were made unwelcome on meta.d.

    I don't think that's exactly right. We just have to abide by meta.d standards, which are relatively straightforward. Granted, some of the moderation was odd and originally unexplained, which strained relations a bit. I think we've gotten past that now.



  • @sam said:

    I just tried it on chrome on try.discourse and it worked, what browser are we speaking?

    Browser doesn't matter.

    You do need a dozen or so unread topics, though. There probably aren't that many unread topics on try.discourse.com

    http://what.thedailywtf.com/t/dismiss-posts-doesnt-update-unread-counter-even-after-full-refresh/3735

    At the time I posted this on the 3rd, I'd already been noticing the bug for a few weeks, it just took me a bit to figure out what caused it.

    And BTW, the reason people like me get upset at bugs like this is, this is goddamned elementary school stuff. How can different tabs have different ideas of what topics have unread posts? Why isn't there a single canonical source of data for that? Why is it so random? It's like the button goes, "well I don't have quite enough time to mark them all as read... I'll just leave a few behind!" How is a bug like that even possible in well-designed software?

    And the regressions. Remember the list numbering being wrong? Remember it being wrong for ages, and Atwood saying it wasn't a bug, despite it flying in the face of common sense? Remember the Markdown standards document (that Atwood himself was a part of!) declared it was indeed a bug? Remember how Riking after all that goddamned fucking drama finally fixed the fucking thing?

    Well, it's back. REGRESSIONS!

    1. test
    2. this
    3. bug

    @sam said:

    I get it, some are you are upset, some hate the software and so on. I banned myself cause I just can't sit around and listen to people yelling at how my team is incompetent and software is rubbish. Its grating.

    Well, there's three issues here:

    1. The bugs we're finding are fucking obvious, and most of them are the result of simply awful design in the first place. Like having two sources of data for which topics are considered "read", and having different tabs in the UI checking different sources of data.

    2. You keep tinkering with shit that has nothing to do with solving bugs, some of which appears to have no purpose except creating new bugs (like changing how the profile image is centered). What was the point of changing that? Were you even aware you broke that, or is it just another regression out of millions?

    3. There are entire HUGE CLASSES OF BUGS related to infinite scrolling that will never get fixed and we don't even bother posting anymore because Atwood is such a douche about that. Here's a bug: the BROWSER'S SCROLLBAR DOESN'T WORK, IT'S NEVER WORKED and as long as Atwood is in charge, IT NEVER WILL WORK. An you guys, who are supposedly software engineers, simply do not care that the standard window widgets don't work.

    Number 3 is the one that really cheeses me off. It makes me feel that all other bug reporting is just re-arranging chairs on the Titanic, as I've said in the past.

    But that aside, I'd be a million times less upset if you did one simple trick before pushing software to production:

    TEST IT

    Test your changes. Test for regressions. Test for the bugs so goddamned obvious we discover them in milliseconds after the update and, just maybe, don't push out the update if it has bugs like that.

    I mean, I know asking for you guys to actually do usability testing is way too much, because this project is too open source-y to give a shit whether the end-result is actually usable or not. But I'd love to see you guys yank a guy off the street and plop him down in front of meta.discourse. Because damn.

    @sam said:

    I understand that you have your WTF persona on the super ON switch here, I am pretty sure you don't have them on most of the time in real life.

    This is real life.

    My enjoyment of this community (barely) edges-out my hatred of this shitty software. For many prominent community members, it did not-- and now they're fucking gone, thanks to your crappy forum software. Maybe if your forum software forces a few more people to leave, it'll shift the equation and I'll go too-- then you can be rid of me. And finally meet your goal of "civilized discourse" by simply alienating half the community and shattering it into nothing. The true goal of all collaboration software: to piss-off and eventually destroy its community.

    Hell, I don't even bother reporting 90% of the bugs I come across, because fuck your open source philosophy shit-- I'll test your software for you when you give me some damned money!-- I just the report the ones that make me go, "wait, am I crazy, or what's going on here?"

    @sam said:

    I appreciate what you are doing here and the bugs you find are helping make the software we are building, better, for everyone.

    Riiight.

    If you care about building better software, you need to go back to the drawing board. Infinite scrolling isn't better than what we had before. Your bloated JS libraries aren't better than what we had before. Markdown is goddamned outright awful compared to real WYSIWYG interfaces. (Even Community Server's, yes, the 2007 version we were using, the one that sometimes famously deleted two characters when you hit 'delete' once, is amazing compared to the editor here.)

    You've made some improvements on the "old fashioned" forum software Atwood hates so much, like better image uploads/pastes, and real-time updating of events is nice. But you've attached to those genuine improvements a dozen whackjob-crazy ideas that don't work, and the end result is forum software that most people here at least hate more than 2007 Community Server. I can guarantee in the years I was there, nobody left our community on that forum due to the quality of the forum software.


  • Grade A Premium Asshole

    Beyond all of that, the software itself is completely unsuited for this group. Completely. Discourse is all about gamification to promote: staying on topic, civilized discussion, killing trolling, etc. Basically everything that this forum is based upon. It is just a shitty choice...for us.

    The only way Discourse can ever work, here, is to run all of us off and start completely over. Almost any other forum software would be more suitable for this group.


  • ♿ (Parody)

    @Intercourse said:

    It is just a shitty choice...for us.

    Except, perhaps, for one thing: karma.



  • @Intercourse said:

    The only way Discourse can ever work, here, is to run all of us off and start completely over. Almost any other forum software would be more suitable for this group.

    Yeah; we all told Alex that on Day Fucking One, remember? He's in the pocket of Big Discourse, it's never gonna happen.



  • Hi Sam!

    Glad your back.

    Thanks for trying to help.

    Thanks for not dismissing us with Cat-memes.


    Srsly tho, in your position, why would you read THIS forum with your morning coffee?? Someone's bound to be complaining. Enjoy yourself a little first, before taking that chance. ;)



  • @blakeyrat said:

    He's in the pocket of Big Discourse

    Beware the Discourse Industrial Complex.




  • Ok so I finish reading the threads I might give a shit about (spoiler alert: I didn't give a shit about any of them). So now I go to dismiss posts.

    "Latest" tab shows 20 unread.

    "Unread" tab shows 20 unread. So far so good.

    Hit "Dismiss Posts"

    (BTW, I swear to Christ yesterday there was a copy of the "Dismiss" buttons above the list as well as below-- am I crazy? Or did you guys introduce, then instantly delete, that feature? Using this site is like trying to build an office block on quicksand.)

    Now Unread shows nothing.

    But... Latest still shows 9. Those 9 are zombies I either have to ignore (forever!) or manually click-into to update the read markers.


  • Discourse touched me in a no-no place

    @blakeyrat said:

    BTW, I swear to Christ yesterday there was a copy of the "Dismiss" buttons above the list as well as below-- am I crazy?

    I'm sure I saw those earlier today.



  • They are still there for me. I think they are only there if there is more than one batch (infi-scroll) of posts. It is not like users expect consistency with UI's.



  • @jaming said:

    They are still there for me. I think they are only there if there is more than one batch of posts.

    How many posts are in a "batch"? This is a unit that is new to me.


  • ♿ (Parody)

    @loopback0 said:

    I'm sure I saw those earlier today.

    Last update was 12h ago (or so).

    http://what.thedailywtf.com/t/docker-upgrades/1929/76?u=boomzilla



  • I just edited, but I mean if infinite scrolling will load more unread topics or they are all loaded initially.


  • Discourse touched me in a no-no place

    @boomzilla said:

    Last update was 12h ago (or so).

    Since then, as it was while I was at work. I've noticed them appear / disappear before, just assumed like normal that it was just another useful feature being removed.

    @blakeyrat said:

    How many posts are in a "batch"? This is a unit that is new to me.

    Same.

    @jaming said:

    They are still there for me. I think they are only there if there is more than one batch (infi-scroll) of posts.

    Don't think I had enough for another 'batch' earlier. If that's it, that's still pretty daft.
    @jaming said:
    It is not like users expect consistency with UI's.

    Not after using Discourse for a while.


  • ♿ (Parody)

    @loopback0 said:

    Since then, as it was while I was at work. I've noticed them appear / disappear before, just assumed like normal that it was just another useful feature being remove

    Could be. I don't try to rationalize UI elements around here. I just consume them as the opportunity arises. It's like eating bugs when lost in the forest. You never know when your next chance will be.



  • @loopback0 said:

    Don't think I had enough for another 'batch' earlier. If that's it, that's still pretty daft.

    Perhaps, I did clear my browsers cache though, and I am seeing the buttons up top, but perhaps there is another factor; I haven't experimented with it much.


  • Discourse touched me in a no-no place

    @jaming said:

    Perhaps, I did clear my browsers cache though, and I am seeing the buttons up top, but perhaps there is another factor; I haven't experimented with it much.

    This does seem to be it - tried it over on meta.d - had over a page of Unread posts and got the buttons. Then reloading the page once I'd got under a page, no buttons. FFS.



  • Yeah, I just went spelunking and found this:

    Bonus points for magic numbers!


  • Discourse touched me in a no-no place

    At least it's likely to be consistently inconsistent, then.



  • @codlnghorror said:

    Consistency is overrated. Discoverability is what matters.

    From meta.d (and linked here).


  • ♿ (Parody)

    /sniff. I still miss the Likes column.


  • Discourse touched me in a no-no place

    @ChaosTheEternal said:

    From meta.d (and linked here).

    I remember that. It definitely explains a few things. If only the discoverability of things was better.



  • I miss it, too ...
    Why did they have to kill it?


    Filed under: never forget


  • BINNED

    I know! They should render the list into a div with opacity: 0, and then calculate the required number of posts to show the buttons from their combined (invisible) height and viewport height!

    Sorry, not trolling, just thought of this and it was a WTF-y enough of a solution to post it.

    I'm kinda-sorta almost fine with the constant here tbqh /duck

    Edit: fucking autocorrect


  • ♿ (Parody)

    The real question is when is it coming back? Supposedly it was going to be a configurable thing post 1.0.


  • ♿ (Parody)



  • Considering this scenario:

    • 100% zoom
    • 1080 vertical resolution
    • Hidden taskbar
    • Stock styles

    Only 17 topics show in the list without scrolling. So, if anything, that magic number should be at least halved.


    Filed under: Or, better yet, 0



  • Goddamned you're an optimistic motherfucker.

    Despite what they type in that thread, the real answer is: "we never gave a shit about that and just said it to shut people up, fuck you."


  • ♿ (Parody)

    I don't think it will actually happen. I expect the topic to be closed at a minimum. And possibly (remote) to have my post deleted.

    Still, it might be fun to see what discussion this triggers.



  • So they actually went OUT OF THEIR WAY to hide these useful widgets.

    OUT OF THEIR WAY to make their product shittier. Just showing them all the time, nah, too simple-- too useful! Write more code to make it worse.


  • BINNED

    I did say "almost".

    I mean, given the circumstances it's a semi decent solution to a problem. Whether it's even a problem that needs solving is a different matter.


  • Discourse touched me in a no-no place

    Ummm...



  • Yes, that would be pretty much the exact kind of thing I mean when I was like, "JUST TEST YOUR CODE!"

    I refuse to believe that anybody actively looking for bugs related to a change to the way the "new posts" banner is displayed could have possible missed that. There's no way.

    The ONLY explanation is that they didn't test the goddamned change. At all.



  • My problem isn't really with the "30"–it is the fact that the buttons are optionally there. I don't know... how 'bout just putting them there all the time?


  • BINNED

    I do something even worse: when there are no more pages to show and clicking "Next" would make no sense, I set the button to "disabled" state.

    Then again, I use pagination, so I'm clearly Doing It Wrong™ anyway.



  • @aliceif said:

    I miss it, too ...Why did they have to kill it?

    Because someone - who already hid it on their forum via CSS - complained about how it cluttered the topics page. Jeff decided that it was time to update the UI for everyone by:

    • Hiding the likes column
    • Heatmapping the posts column based on the likes:posts ratio


  • @Onyx said:

    I do something even worse: when there are no more pages to show and clicking "Next" would make no sense, I set the button to "disabled" state.

    It's like you're a prophet ... a UX prophet!

    All hail the Divine Prophet @Onyx!


  • Banned

    I have a repro of the zombie thing, confirm it is mega annoying, having a look.


Log in to reply