The NEW Official Unofficial Discourse bug tracker!
-
-
-
He only received one there, too, you'll need to spread your @blakeyrats out over threads@blakeyrat
-
I really don't know why you're wasting your time. I especially don't understand why you've spent real money on trying to educate <a @twatwood & Co.
The Discodevs™ don't think their tools are broken, and don't want
your bugs. they might possibly accept their process is broken, but only partially:My issue here is you are all complaining, "you need a new tool" when the issue you have is "you need different process"
Tools do not dictate process.
…but then…
@sam said:OK, can you show me another open source project similar to Discourse in scale (5 man team) that fixed 4 bugs in the last 24 hours using their bug tracker
If their tool is so fucking good as a bugtracker, let the cunts trawl through meta/bugs over here.Remember:
WTF has been on Discourse almost 60 days now. We believe most of the major issues are resolved, so we're now closing the door on this chapter of our involvement in tdwtf -- the Discourse team members will move on.
In the future, we won't be around on this site except extremely sporadically, so please escalate bugs to http://meta.discourse.org if you feel they are serious enough to warrant our attention.
So long -- and thanks for all the bugs.
We should only be posting security bugs over at meta.d. I mean that. the rest should just get an <a @mention to each of the discodevs.
-
Tufty, imagine a day when discoursebugs.com shows up over discourse.Org
You should head towards that day.
Or treatit like a learning site of things not to do
Pick your poison.
-
Joined the fun and created my first issue.
-
Sam already admitted fault to that
-
Oh. Ok. Who raised it with him, wasn't me I think.
-
-
My poison is wathing this train wreck unfold, day after day, as we uncover bug after bug after bug, all caused by the same underlying bad design decisions. I'm actually getting a huge kick from watching the discodevs entirely fail to "get it".
Maybe that makes me a bad person. I don't know.
Anyway, it's your money, feel free to spend it how you wish, but you're fighting a losing battle. You're doing half the discodevs' homework for them (and paying for the privilege), but without tests for regressions, you can never really flag any given bug report as truly closed.
So what you've created is a Schrödinger bug report system, unwanted and unsupported by the discodevs. You're not helping the situation. Indeed, about the only thing that's gonna help the situation is finding another low-level-access hole, preferably one that leaks personal information, exploiting it on every Discourse install out there, and making the results public.
-
The idea is that if they have historical information so that they can see regressions, maybe they will up their testing? Really though, when possible you should institute a test for every bug before fixing them. Then regressions are caught every time.
Oh well. Fuck it. No sense starting TDD practices now.
Also, @blakeyrat.
-
https://meta.discourse.org/t/the-badge-disaster-is-my-doing/19907
Notice that although this is clearly a bug, this is just a post under "Meta" directly. There is also no link to GitHub with the fix.The bug itself got reported here, but Sam seems to be replying "as new topic" away from this bug report. What is this? Paperchase?
And then they scold us for telling them their processes are wrong.
-
, unwanted and unsupported by the discodevs.
I much prefer @Matches creation to having bug reports sprawling in random spots on tdwtf
-
I much prefer @Matches creation to having bug reports sprawling in random spots on tdwtf
<sarcasm>
But I thought Discourse was a top-rate bug tracking system? Maybe even Awesome.
</sarcasm>
-
But I thought Discourse was a top-rate bug tracking system? Maybe even Awesome.
Have a great Friday night and awesome weekend.
-
I'm actually getting a huge kick from watching the discodevs entirely fail to "get it".
Maybe that makes me a bad person. I don't know.
If the Discourse devs would be a couple of high school kids working on their first large project, yes, getting a kick from watching them fail would make you a bad person.However, since most of the team are senior developers, and combined with their attitude of "we are doing it right, stop critizising our processes!"... I don't know.
I could argue that "getting a kick out of watching someone fail" always makes you a bad person, but that would mean taking a rather high moral high ground, and I would prefer not to be called up on that myself.
-
TWIAVBP. It's Friday morning here, which is nice, as I have a day off and am going to head off into the mountains (no more bug reports or snark for a couple of hours). And as I get to work over the weekend, probably nothing for a few days.
That said, you probably want to up your bug-fixing rate, as we're discovering them faster than you're fixing them.
BTW, you realise that the word "awe" incorporates the concepts of "wonderment", "fear" and "dread"? Perhaps your goal is complete already.
-
Hmm, I guess I'll chime in seriously.
Unsurprisingly, I think discodevs are right in using Discourse to track bugs. Sure, looking purely from a bug tracker perspective, there are better choices. But looking holistically from a business perspective, there are some strong counterpoints.
-
Dogfooding
There're only so many working hours in a day. Discodevs need to ration development, community interaction, bugtracking and dogfooding using this limited budget. By using the same software for both community interaction and bugtracking, they can effectively dogfood/test the forum while doing these, and use the saved hours on other tasks. -
Community interaction
Discourse needs a place to track bugs and a place to discuss features. If they had a separate bug tracker, many of their users would go only there, where communication is superficial and highly formalized. By funneling them to their meta.d site, they get to involve the people who are only interested in tracking bugs in their wider community, and also to pitch them plugins, feature proposals and maybe even paid services in the future. -
Information control
Bug tracker makes it very clear the number of open bugs. This can be embarrassing in a very public software you're trying to pitch to serious players (see the recent Twitter dev forums victory). By mixing the bugs in with other communications, they get to fudge the bug numbers by either deleting some posts in the name of regular "forum moderation" or treating only the first page as "valid" bugs. Bug trackers are usually designed to keep you more honest, with all sorts of statistics and patterns you can derive.
IMO you guys are not seeing the forest from the trees. Efficient bug tracking is important, but not be-all end-all for a business like Discourse.
-
-
Dogfooding
Q : Do you know why dogs spend so much time licking their bollocks and sniffing each others' arses?
A : If you'd tasted dogfood, you'd know..If they had a separate bug tracker, many of their users would go only there, where communication is superficial and highly formalized.
Bug discussion should be superficial and formalised.Information controlDishonesty
FTFY.If they had a decent bug tracker, they might be able to prioritise work to fix important bugs rather than whatever's itching Jeff's arse, show what they are currently fixing, what their ETA on specific fixes is, have a global view of not only what bugs are outstanding, but what bugs have been fixed (and thus an idea of where there might be underlying problems1)
As it is, they're working like the amateurs they are, not like professionals. Hell, there's still been no reply as to whether there is a design document for this fucking mess beyond "Awesome, dude!"
1. Here's 3 obvious ones - Implementation of infinite scrolling, parsing via regexp, bad practices.
-
- Dogfooding
There're only so many working hours in a day. Discodevs need to ration development, community interaction, bugtracking and dogfooding using this limited budget.
Which is why they shouldn't have to waste time using a forum software as a bugtracker.Regarding dogfooding: using your own product is far more important in the early development stages, when the amount of serious malfunctions you can detect that way is very high.
- Community interaction
Discourse needs a place to track bugs and a place to discuss features.
Yes. "One" place to track bugs, and "one" place for discussion. Not "one" place for tracking bugs and discussion.That being said, I wouldn't mind if they pimped Discourse so that it could be used as a full-grown bug tracker. The infrastructure is there, all you need would be a couple of additional database fields for severity and importance and a special category that treats things slightly differently. And, voila!, you have a bug tracker where you can order topics by severity, or creation date, or reporter...
-
Q : Do you know why dogs spend so much time licking their bollocks and sniffing each others' arses?A : If you'd tasted dogfood, you'd know..
Hmm... mental note taken.
Regarding dogfooding: using your own product is far more important in the early development stages, when the amount of serious malfunctions you can detect that way is very high.
Disagree. The more you are forced the use your software, the better you understand its weaknesses and strengths and better are you able to plan future development. It's not just about finding bugs. It's understanding the difference between how things are in theory, and how they work in practice.
For example, discodevs had implemented a bunch of features to effectively spin off many smaller threads, yet in reality, both them and us end up participating in these long-ass megathreads. There's something wrong there. And you don't see that from design documents and bug reports. You need hands on experience with your product, if at all possible.
Yes. "One" place to track bugs, and "one" place for discussion. Not "one" place for tracking bugs and discussion.
You can do it like that, but as pointed out, there are valid business reasons to do it all in the software you're making.
-
The more you are forced the use your software, the better you understand its weaknesses and strengths
And the more likely you are to say things like "Oh but I don't have that problem." or "You are doing it wrong, you should do it like this." Because other people are not following your workflow and stuff like that.
-
Maybe that makes me a bad person.
Correlation is not causation. Especially in this case, I think.
-
That being said, I wouldn't mind if they pimped Discourse so that it could be used as a full-grown bug tracker. The infrastructure is there, all you need would be a couple of additional database fields for severity and importance and a special category that treats things slightly differently. And, voila!, you have a bug tracker where you can order topics by severity, or creation date, or reporter...
I could see plugging discourse in as the comments system of a real bug tracking system. That would make sense. At minimum, like you say, a few extra fields and even some external plug-in to give you the sorts of views that a bug tracker gives you.
Something.
-
And the more likely you are to say things like "Oh but I don't have that problem." or "You are doing it wrong, you should do it like this." Because other people are not following your workflow and stuff like that.
Ok that's fair. The other side of the coin.
-
And the more likely you are to say things like "Oh but I don't have that problem." or "You are doing it wrong,
Which is exactly what is happening and has happened. You get stuck in your way of doing things, fix those bugs, but never venture beyond your work flow and see how varied people might use it and the way they might break it.
Jeff is building a product that is perfect for him, but not necessarily thinking about what society might want.
This software is also only open source in that we can see the code. It is not open source in spirit though. I can guarantee that if we had a PR that included pagination as an option, it would never see the upstream. I can also guarantee that if we forked it and included pagination, it would not be long before upstream broke how we accomplished it, leaving us on our own branch.
-
@Intercourse said:
Jeff is building a product that is perfect for him, but not necessarily thinking about what society might want.
This is somewhat true, but not entirely. There are plenty of people over there who agree with him. And not just discodevs. But they just provide the confirmation bias he needs.
-
Might be this:
Sam also mentioned that in the last update message:
http://what.thedailywtf.com/t/docker-upgrades/1929/38?u=boomzilla
-
Where do I go to report issues in the issue tracker?
-
-
-
The usage of "It's turtles all the way down" never fails to make me chuckle.
-
Good choice sam! Have fun with the family.
-
-
The Way I Assume Vampires Breed Pigs
-
It's a turtle. It doesn't have to stand on anything. It swims.
-
This only applies if the original report is reported on discourse bugs, then moved to a real bug Tracker for long term reference and advisement of status.
Discourse is rarely used to get more info on bugs and it's closed so no real discussion can take place, and old items are deleted.
Those secondary features are the reason a bug Tracker is required. Not because of any particular bad issue to get initial reports from the forum or have follow up discussions about it.
-
Nice. I'll fix it.
-
Disagree. The more you are forced the use your software, the better you understand its weaknesses and strengths and better are you able to plan future development. It's not just about finding bugs. It's understanding the difference between how things are in theory, and how they work in practice.
And that's exactly what's not happening.There's no acceptance that anything about Discourse (apart from a few unfortunate bugs) is anything other than unicorn
and 
s. There's no willingness to accept that anything about Discourse is anything other than unicorn and s.For example, discodevs had implemented a bunch of features to effectively spin off many smaller threads, yet in reality, both them and us end up participating in these long-ass megathreads. There's something wrong there. And you don't see that from design documents and bug reports. You need hands on experience with your product, if at all possible.
Exactly. This was pointed out, that people weren't going to use Discourse in the way Jeff thinks it should be used, and that its fundamental design philosophy of infinite scrolling wasn't going to work in many cases. Jeff's response?I can't find the actual post (after all it's buried in a megathread1 but it went along the lines of
- Big threads don't exist
- Even if they do exist, they shouldn't exist
- We only load in a few posts at a time so it's no problem, performance-wise2
i.e. the actual issue went straight over his head. Why? Because Jeff has no concept of how actual people use software, and how he expects software to be used is the one true way™. Hell, as I recall it, in the same thread he said that infinite scrolling was his religion, and there was no room for other religions in his worldview.
The view from Discoland is that the design of Discourse and the Discourse development "process" could literally not be any more perfect. If you disagree with that, you're considered a troll.
1 Only 400+ posts, actually, but enough to confound discosearch and too hard to scroll through to find it.
2 Said claim since proven to be a "Jeffact"
-
The World is a Very Big Place
-
Nice. Merged your pull request.
-
Nice. I'll fix it.
I sent you a pull request, so that it could take you slightly longer to deal with it than just changing the text directly.
EDIT:
Nice. Merged your pull request.
I probably should have read the whole topic before replying...
-
It did take a slightly longer time because mobile didn't indicate a pull request was pending, then when I tried to delete the extra http it wouldn't delete. Then I loaded desktop and saw there was a pull request for your edit from the text editor, and I had to figure out how the merge feature worked on bit bucket because it's usually just me doing commits via push. . Merging isn't a thing.
Filed under: less usability than a turtle.
-
There's no willingness to accept that anything about Discourse is anything other than unicorn and s.
Maybe you're just buying their marketing pitch too hard. They ARE trying to sell the product, after all.
IMO Jeff was pretty disappointed by the mobile performance. He was betting the farm on making a single-page js app instead of the usual expensive combo of web + native mobile. But that's not panning out, at least for now. Also notice that in his recent pitches, the "FREE open source" part is much more prominent than it used to be.
Not that discourse is bad, I think they are just realizing they are not gonna conquer the world as quickly as they did with stack overflow.
-
IMO Jeff was pretty disappointed by the mobile performance.
Funny enough, I have much better performance when I use this site on mobile instead of on my quad-core Xeon workstation PC.
-
So do I. It won't load at all, so I have to go and do real work® instead
-
For example, discodevs had implemented a bunch of features to effectively spin off many smaller threads, yet in reality, both them and us end up participating in these long-ass megathreads.
That's because this isn't forum software, it's webIRC. New topics are like new Channels, and you participate in whatever channels you feel like.
Some channels die and some live on forever depending on user participation/interest.
-
-
I still say StackOverflow is successful for the same reason Steam is: it's never had any effective competition. Defeating ExpertsExchange wasn't a "challenge", it was fucking trivial-- nobody had done anything to improve that site in like 10 years.
It's not impressive to be on top of the pile if nobody else is trying to climb the pile. Also the pile is dirty diapers. And the dirty diapers are the kind of Internetpointzzz-obsessed Wiki editor types that make every gamified site awful.
-
I think that getting a kick out of watching them fail back in May would have made us bad people. But now, we've tried to help them. In response, we got spit in our faces and shit on our doorstep. Not to mention our beloved, rickety shack was replaced with an even more questionable hut. And when we try to get the construction team to shore up the foundation and fix the leaks in the roof and walls, 80% of the time we get ignored. I think at this point, it's ok to enjoy watching the Spiral of Doom™.
The badge disaster is my doing
Discworld (world) - Wikipedia
