Hacker News new | past | comments | ask | show | jobs | submit login
The new Disqus 2012: Elevating the Discussion (disqus.com)
60 points by dctrwatson on June 13, 2012 | hide | past | favorite | 55 comments



I block third-party cookies, and as a result I never see a Disqus comment. I've found that if I don't allow Disqus to set a cookie their JavaScript widget just spins ineffectively.

I wonder if the new Disqus will stop requiring me to send them a tracking cookie in order to read the comments. (I accept the need to identify myself if and when I want to post. But it would be nice to be able to visit a web site with Disqus installed without that visit going into my permanent record next to my name.)


Not only do we not require you to set third-party cookies to read comments (we never did afaik), the new version also allows all users - regardless of cookie settings - to comment. It just won't remember your session between page refreshes.


Very good, thanks. I was certainly hoping that I had misdiagnosed this problem, and apparently I did.

I'll follow the suggestions of the other commenters and suspect the Do Not Track plugin. I guess the fact that my problems go away when I whitelist Disqus cookies is some sort of crazy side effect of my crazy toolchain.


Would you do me a favor, go check out the volokh conspiracy. In Chrome with adblock plus off, but disconnect on, no comments can be seen. Disconnect seems to find a google tracker that it blocks (but provides no other information on.)

Thanks.


I block third party cookies and can read Disqus comments. It seems to be working exactly as you suggest because it warns me "Warning: A browser setting is preventing you from logging in. Fix this setting to log in". Unfortunately, adding an exception for disqus.com doesn't fix that in Firefox 13.

Disqus can also be blocked by NoScript, Ghostery and similar/equivalent addons or plugins so maybe that's the reason, not third party cookies.


Heh. I deliberately block all disqus resources with Adblock because they are frequently the slowest to load. My first question reading this was to ask if they made it faster, but the sub second updates mean its going to stay blocked. I want my browser sending less traffic, not more. (I also block gravatar and a bunch of other website "plugins")


It would be nice if disqus would detect these things and send a static HTML thing with "allow us in noscript to join the discussion!" and maybe a static view of the thread so far.


Slowest to load? That sucks. It would be great if you could provide details. We serve all our static assets using Akamai, and they're aggressively cached client-side.


It may be a matter of perception, because you're firing off ajax requests after loading. I visit a page, stuff happens, and then there's "waiting for disqus.com..." in my status bar. The page could be ready to view, but out of habit I generally wait until everything is done. Whatever the last thing a site loads is, if it doesn't seem necessary for the site to function, it goes in adblock.


I use Flashblock and Do Not Track, one of those messes with Disqus and I never see discussions either.


The new disqus feels very much like the users are actually disqus users that happen to be commenting on your site. Not your users talking on your site. I'd put this on a fun site or a site I did not care very much about, but on a business related site, I'd want to control my own users and not hand them all over to disqus.


This is a huge advantage. I want to be a disqus user, not a user of 4000 tiny blogs. When I make comments on x, y and z random blogs, I want to be able to follow replies to them in one central place.

I especially don't want to give a username and password to all these little poorly-secured blogs.

If you want to "control" "your" users, build your own comments system. And turn off RSS.


  I want to be a disqus user, not a user of 4000 tiny blogs.
Yes! And in addition to the above arguments, I would say, having a single identity across a variety of sites may be convenient.


> I especially don't want to give a username and password to all these little poorly-secured blogs.

How many blogs require you to create a blog to post?

> And turn off RSS.

What's wrong with RSS?


Nearly all blogs require you to authenticate with some kind of account. The alternative is spam city.

RSS: There's nothing wrong with it. But if you object to Disqus centralising "your" users, you should equally object to eg Google Reader centralising them. It's exactly the same principle. You lose "control", to the benefit of the user.


Agreed.

They should get rid of that huge navigation bar. The last thing I want on my site/blog/whatever is additional, official-looking navigation buttons driving the user away from my site.

There's no reason for such a clumsy, intrusive widget. Disqus has plenty opportunity to brand the user within the registration popup already.

A tiny button with a popup-menu would serve their goals better. The current design will definitely turn off precisely the webmasters that would have been the most likely to consider a paid plan.


The demo and screenshots on disqus.com are a little behind our latest design - but the nav bar has been toned down a lot. Take a look on avc.com or slashfilm.com.


Frankly, not impressed at all.

I took the liberty to fix it for you; http://cl.ly/1d2T3S0B0G3F3j343q2g

The user just wants to see the comments and possibly add his own.

The user doesn't want to learn why there is a "Comments"- and a "Reactions"-tab and what the difference is. The user doesn't want to be bothered with yet another meaningless star-gauge. The user doesn't care about your "community".

None of that clutter provides any value. It's just noise.


Are you sure you're not just substituting "I" with "the user" in there? I'm going to give Disqus the benefit of the doubt and say that they have user tested their new design. By what criteria do you say that "the user" (which is a neat getaround for saying the unprovable "all users") doesn't care for these things?


By the common-sense criteria.

We've seen the cycle countless times. $Company damages their product in an attempt to create "social lockin", without realizing the futility of trying to compete with the established social networks. This is then inevitably followed by either a quiet rollback, or a less cluttered competitor eating their lunch.

I like disqus and sympathize with their efforts to find a business model, but they're not going to be an exception here.


> By the common-sense criteria.

Stopped reading right here. Either provide UX research to back up your claims, or stop pretending that they apply to your mythical invention of "the user."


UX Research, eh?

Let's just say your downvote on the original comment didn't make a dent, so much for a quick, unrepresentative poll. What I call Common Sense is ultimately no more than the essence of past UX research and can be obtained by a mere glance at what prevails in the grand petri dish.

Cf. Altavista/Google, Myspace/Facebook, and the rise of Apple for just a few major case studies in the area of UI clutter. With regard to the siren-song of social-lockin I'll recommend to study the fate of Ping, which exemplifies what happens when even the big guys try to go against the grain.

If you find a piece of formal "UX Research" claiming anything else then I'd be curious to see it.


I really hope you see the irony at some point that you're complaining about social elements in an inherently social page element: the comments section.

And I didn't downvote your original comment, so I have no idea what you're talking about. Then again, I don't understand why people are sensitive about fake internet points to begin with.


And suddenly the formal "UX research" is out of the picture - now that was quick.

irony ... complaining about social elements in an inherently social page element: the comments section.

That statement makes no sense. So you wouldn't mind HN adding all sorts of widgets, because it was social to begin with?

Further if you read the thread again you'll notice that nowhere did I complain about social elements per se.


How is formal UX research out of the picture? I asked you for some, you refused. If you're not going to even bother backing up your claim I'm not going to bother discussing it any more.


I applied the gist of what I gathered from UX studies that I have seen, made plenty of claims based on that, and backed it with observations from the real world.

Since you started out so keen on science I would have expected you to effortlessly debunk my claims with hard facts or at least a coherent counter-argument. Too bad that fizzled out so quickly...


Now you're just jumping through hoops to avoid linking to anything concrete. It's a shame, I was looking forward to hearing real arguments from you, instead of a fabricated "common sense."


Sorry, I'm still not clear what exactly you want from me. A paper explaining why clutter is universally bad? Seriously?

You try so hard to come across as if you had a remote clue about the UX field, yet you want to debate one of the most basic premises?

May I suggest to review the standard works (you have them on your shelf, I know you do);

http://www.amazon.com/Dont-Make-Me-Think-%20Usability/dp/032...

http://www.amazon.com/dp/156205810X/

http://www.amazon.com/dp/0385267746/

Somehow you must have forgotten, well, everything, since you last read them.

That's okay, happens to the best of us. However, you might come across a little less unarmed if you could at least memorize the title of Krug's book.

Thanks for playing anyways.


And again, you're completely missing the point. One user's UI clutter is another user's essential UI component. Doing actual UX studies on your interface is essential (which I assume Disqus has done thoroughly).

Your original comments claimed that a mythical "the user" doesn't care about the Reactions, Community, or star buttons, citing "common sense." That's not a UX study. That's you trying to spin your personal opinion as fact. Which is cute, but is not useful criticism in the slightest.


which I assume Disqus has done thoroughly

And your assumption is based on what?

Just by looking at the result it's obvious they haven't.

That's not a UX study

Exactly. It's the common sense that you would possess if you had been involved with any similar project of significant size. And I'll proceed to criticize obvious violations of well established common sense without performing studies on other peoples sites, thank you very much.

"the user" doesn't care about the Reactions, Community, or star buttons

Yes, it's still common sense. And quoting what I said does still not make a counter-argument. If the user cares about "Reactions" then where did all those useful pingback-listings go that used to trail every blog post? If he cares about detached community-silos then why can't even Apple make one work? And when was the last time you've seen a useful star-rating on a comment-thread outside an established forum?

That's you trying to spin your personal opinion as fact.

My spin must be pretty good then since you still haven't even tried to dispute any of my claims or references. Oh wait, actually you have once. Turned out not so well...


"hand them all over to disqus"

What possible benefit do you get by having users login on your site as opposed to a disqus frame? Other than the "feel"?


Relationship with my commenters always been an issue for me when I've considered Disqus.

I'm sure I'm not the only site owner who requires real names and cultivates community among his commenters.

Having said that, Disqus looks like an excellent solution for the vast majority of sites, where the quality of the comments and sense of community is very low indeed.


They have VIP features where you can use your own login system.


I'd love to see somebody develop a free clone of this which cuts out the middleman and lets you host it on your own server.

[edit] The videos are pretty good. Their use of "FaceCrack" in the second video is a little cringe-worthy though.


One of the biggest benefits of Disqus is the size of their network. I'm already logged into Disqus and thus can easily reply to someone, comment, upvote and share. When I go to sites that expect me to make a new account, provide my email to a stranger, or give them auth via my Twitter account I likely won't bother at all. Disqus also offers a full account export (or they have in the past, I haven't checked lately) if you ever really feel like you want to go self-host.


The hypothetical self-hosted version would be a perfect candidate for OpenID auth I guess ...


Vanilla has a plugin to do this, but the link seems to be broken at the moment. I'll contact them.

http://vanillaforums.org/addon/embedvanilla-plugin


Me too. Juvia is the closest I've seen: https://github.com/phusion/juvia


I'm interested to know what their new architecture is, I believe that they are a Django shop. Does anyone know if they are still using Django and if so what method they used to do real time?


When I want to know things like this (and I usually do) I take a look at their Jobs listings, this usually reveals the technologies used, in this case I found:

Disqus is the largest Django application and one of the largest pure Python implementations on the web. We love JavaScript and write highly optimized code with performance and compatibility in mind.


From Adam Hitchcock's (Disqus engineer) upcoming EuroPython talk description:

"flask, WebScale!, gevent, redis"

https://ep2012.europython.eu/conference/p/adam-hitchcock


On the client-side, poking around in the console leads me to believe they're using Backbone.js


Django still powers a lot of the DISQUS infrastructure. That said, we're in the process of breaking certain parts of it off into independent services and out of Django. But Django will be the main component for the foreseeable future.


Could you talk a bit about why you're doing that?


Don't misread what Fluxx is saying. The core of disqus and our primary data flows are still powered by Django and that won't change. New servers which are less complex (e.g. the realtime system) are generally written on top of Flask.


The interactive demo on http://disqus.com is pretty amazing.


It's one of the best interactive walkthroughs I've seen. Executed incredibly well.


Top of the layout looks inspired from Rdio.

http://d.pr/i/LsgV


At first I thought "what demo, there's just a screenshot here" but wow, nice demo.


I've never used Disqus, and I guess this might be off topic, but...

Do they let you export the comments people leave on your website so you can easily switch to another commenting platform?


Yes, IIRC it's one of their big selling points. You can easily get your data out and use it for whatever you want. For example I believe many Jekyll blogs use Disqus for comments and periodically grab all the comments then write them into the html so that google will spider the comment section.


I believe in wordpress their plugin automatically stores all comments so that if you turn it off they convert to the wordpress comments.


This is still just threaded comments with some fluff on top for discovery and community. Threaded comments have been around forever, and they don't scale well for big communities.

Someone needs to take the idea of "semantic comments" and run with it to visualize community reactions.


If you use Disqus on your site, I'm not going to comment on it. More than once comments I entered via Disqus have vanished into the ether. Disqus claims that the site owner deleted my comment, but the owner publicly says they have done no such thing. Disqus then refused to do anything further.

Since I don't know whether Disqus comments will actually stick around or not, it seems like a waste of time to participate. The fingerpointing from their support staff didn't fill me with confidence.


We've made a lot of improvements to the system (including the spam filter) since you last contacted us — are you still seeing similar issues? If so, we are always happy to take a look. Feel free to contact Support if you are having trouble: http://disqus.com/support




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: