It looks like a threaded view, but it ain't

I’ve tried to get it to crash as before, and it no longer does so! Nice one, thank you.

However it is still horribly slow to load posts at the end of a page.

And of course the UI massively fails POLA (eg the replies dropdown), the tons of white space mean you are scrolling the whole time, and the whole thing is painful and appears to be predicated on design decisions that are bound to drive me (and others, probably) away. But none of those are programming prioblems per se.

On our near future roadmap is to start unloading stuff way as you load up tons of topics. At the moment we are keeping every post you have loaded into a topic you are browsing, in memory. We are going to start unloading posts as you scroll through large topics which will lead to a very big perf improvement for Firefox.

Can you explain what POLA is? Keep in mind, the styling is totally controlled by @beschizza and team, if they wish to tighten up whitespace, remove controls, disable animations and transparency effects and so on, it is totally up to them.

1 Like


The Principal of Least Astonishment. I seem to remember Wikipedia has a page on it.

Briefly, cos I’m at work: if it looks like a duck, and quacks like a duck, it had better bloody well swim and lay duck eggs, otherwise duck users are going to get confused as hell.

Egg-xample: The “replies” dropdown has confused a bunch of people into thinking they were looking at a threaded reader that didn’t work right. Lots of people have posted here saying so. It got me too, and I’ve been designing UIs for twenty years.

It looks like a threaded view, but it aint. Breaks POLA. There are a few other examples on the bbs, but that’s the biggest one AFAICS.

1 Like

Have you used Twitter? Perhaps you’d be less “astonished” if you had :smile: – it works in the same way. You expand and get just the parent and 1 level of children.

Although looking at this more closely, I now notice that’s not exactly correct. You see the parent-of-parent because I was not directly addressed there:

  • Resig said something
  • McGraw replied to Resig and mentioned Atwood (and Stack Exchange)
  • Atwood replied to McGraw and Resig ← this is the thing clicked to expand
  • Resig replied to Atwood, McGraw
  • Sonntag replied to Atwood, McGraw, Resig
  • Mouton-Dubost replied to Atwood, McGraw, Resig

We could do something like that to walk the chain up, but it’d be flat, like you see here.

The analogs are closer to Twitter because:

  • like Twitter, you can reply to 3-10+ people in one Discourse post if you quote them.
  • like Twitter, there is “reply” metadata that technically makes a tweet a “reply” to someone, but it can easily be a reply to multiple people depending on who is mentioned.

Another live example, actually pulled from a quick edit fix I did for a post by @daneel.

You can view the example in-situ at

1 Like

I honestly do not get your attitude, man.


I find that “astonishment”, like beauty, is frequently in the eye of the beholder.

Seriously. That reads as pretty hostile. Take 5 man.


I am astonished at these replies!

Point taken. I will edit to add a smiley so the appropriate level of playful reparteé is conveyed.

1 Like

I always found that adding “no offence” to the end of whatever I was saying worked wonders for this.



Allow me to elaborate. Those who’ve contributed criticism of the new system are not doing so because they, personally, are too dumb to understand how it works.

Rather, they are trying to help you improve usability via user feedback. If such feedback is unwanted - as seems to be the case - would you be so kind as to make that explicit so that we can all stop wasting each other’s time?


There is feedback, and then there is “y’know, I could maybe see myself driving this car, if only it had a truck bed. Why don’t you just bolt one on?”

Which kind of feedback are we talking about, again?

I’d like to at least try not to end up with a Subaru Brat here. Or god forbid a Pontiac Aztek.

I find this comment kind of odd. @codinghorror is coming around to adding a system that allows you to untangle conversations solely based on user feedback. He may have used a joke that was out-of-place, but the substance of his reply is about improving Discourse to help suite your needs based on your feedback.

@codinghorror and I are all over this forum reading every little bit of feedback we are getting and constantly trying to untangle what it is we must do to make you folk enjoy this little corner of the web better.

Our primary goal is to give you folk the software needed to have kick ass conversations.

[quote=“sam, post:14, topic:3090”]
He may have used a joke that was out-of-place […][/quote]

More than one.

1 Like

See, you’re still mocking people… Anyhow, wishing y’all good luck with the project, but I’m out of this thread.


He’s probably not going to see that, ever. Just FYI.

sensitive artist

I love community feedback but the reality is that a lot of it just isn’t actionable for a variety of reasons. I believe open and direct honesty on community feedback leads to better long term relationships, versus generic feel good platitudes. You can read more about my position on this here:

Working on large public projects you get a lot of feedback. Some of it is even good! I just don’t do the “wonderful idea, we will certainly consider that in the future” brush off thing. (Or, ignore altogether, with no response.) Does not feel right to me, I hate it when that is done to me, and I won’t do it to others.


Yeah, but it still ain’t thread in a way discussions can be followed.

There are tons of flat forums out there where people successfully have discussions.

However, flat discussions do require a bit more context in each post – if you’re looking for a place to post context-free one line (or one word) zingers, they will not go over well in a flat format because replies are not docked under the post.

Notice how short everything is here, for a reason:

1 Like