This website's performance is shameful.I would spend easily twice as much time on Reddit as I do right now if it was snappy
Another issue is pretending comment threads are modal?? If I click to read comments, it presents a sheet of comments that if you accidentally click outside them, they just disappear.
And they broke the back button, so when you go back to read those comments you accidentally dismissed, you don't land at your previous scroll position.
It took 15 seconds to load the profile, but I couldn't view any tweets until I'd pressed the "Not Now" button to the pop-up asking me to install the app. Five seconds later, I was able to view my most recent / pinned tweet.
This is all for a service that delivers a timeline of (primarily) text, with some images and video. It should not take 20 seconds to load!
Nitter.net[0] solves a lot of these problems and paired with Redirector[1], it's a transparent switch out.
There's an expired cert issue 3 days ago, but there are other instances[2]
https://github.com/zedeus/nitter
When/if the tweet finally loads, it shows totally unrelated messages at the bottom. I still have to click on it and load it again if I want to see the comments thread.
The upside is that the redesign was the most effective way to keep me completely off twitter.
I don't remember it being like that a decade ago. Maybe there was a big rewrite or two that bloated things up?
In 13 years it went from something I pushed and installed at every client I helped with business processes, help desk, Dev… to an overbloated frustrating mess to operate and manage.
Just changing a dropdown field’s content now is hidden in sublayers that requires a bookmark if you are not toying with that on a daily basis.
(not everybody is an actual full time Jira admin, and a Jira admin should’nt have the mental load of an Exchange admin for such an application).
Happy to hear about alternatives experiences as well.
Couldn't agree more. I've got admin rights to quite a few systems, including Exchange, which is a pretty big ship,so to speak. However, having tried Jira it's hands down the most convoluted piece of software I've ever seen. It took me less time to spin up an instance on AWS+ configure its access to the internet than it took to make a few changes on Jira. God knows how they managed to do It.
There are quite a few competitors popping up, the main issue is that they all seem to be bootstrapped and don't have the funding to make a disruptive marketing splash.
The thing that's driving all of the bad behavior you're talking about is the advertising business model. The sooner we can destroy web advertising as a viable business model, the better.
- 105 requests
- 3.9MB transferred
- 13.9MB resources
- 32.57s load time
Here's a nice repo of other examples: https://github.com/dominictarr/your-web-app-is-bloated
For personal stuff, bloated websites just get replaced with less bloated alternatives. At work, I don't have that luxury...
Slack and GMail I have open all the time. Slack is by far the worst of the two but they both consume a ton of resources. I'd probably be more annoyed by GMail if the UX was as bad as Slack though.
Then there's any atlassian product. Slow, confusing, constantly changing for the worse. There was a time in my life when I used both daily and they weren't that bad. Now I use a slack bot to interact with jira and and make one edit in confluence a day (for a standup and it takes about three minutes to edit, update and save).
And reddit... If I had to rely on that daily I'd probably rage quit the internet. I still use the old interface for most things. It's ugly but it's not a dumpster fire of slow loading API calls to render bloat.
I kind of understand their desire to separate these divisions and focus on insurance and (I suppose still) core banking, but I really miss having a singular place to log into and access the bulk of my financial activity and data.
EDIT: Seems they farmed their 529 plans out to Victory Capital. Going directly to their site they only seem to support: financial professionals, institutional investors, and USAA individual investors. Which suggests, to me, the problem. They aren't really set up to deal with individuals. At least Schwab already dealt with individual investors so my brokerage account and IRAs weren't screwed with too much.
Jira and Confluence are excruciatingly slow, too, but at least the information isn't hidden behind useless feeds.
The dual motivations of SEO and making money from ads have just saturated those sites.
<ad>
Make Cookies Like Grandma Used To!
<ad>
[Jump to recipe] <- only works half the time
Let me tell you about my time with my grandmother,
making cookies with her on Sunday afternoons in the summer.
<ad>
[READ MORE]
So there I was, making a mess like usual.
<ad>
<20 more sentence or paragraph and ad pairings>
Recipe:
[READ MORE]
Ingredients:
- 2 1/2 cups of flour
- 1/2 cup of sugar
- Eggs (wait, how many?)
...
1. Beat sugar and butter in a mixer.
[READ MORE]When I'm on a repository, the sidebar has 5 items that I don't use and cannot remove (Github in comparison lets me hide almost anything I don't use: discussions, issues, wiki...).
Issues are bloated and nothing can be hidden: designs, linked issues, votes, comments, right sidebar with todo, assignees, time tracking, due date, labels, weight (even though the feature is "locked"), confidentiality, status and references.
That's true on every single page on Gitlab: tons of great stuff hidden between things that I don't use.
I love that Gitlab has so many features, but I hate that I cannot cherry pick what I actually use to cleanup the UI.
You bring up quite a few great points. We have worked hard on bringing every tool from the DevOps lifecycle into our platform, and I agree with you that quite a few areas have become overwhelming, especially for new users or people who don't use our entire set of features.
We are actively working on refining that, but as you can imagine, that's no easy feat, as we have a vast variety of project types, workflows, use cases, user types, etc...
One of the most problematic areas, as you perfectly mentioned, is the left sidebar. We already enable project administrators to remove items from there by going to Settings -> General -> Visibility, project features, permissions, and then toggling off the areas you would like to remove. However, this is a project setting, not per user.
At the same time, we also work on making that left sidebar better for everyone, in multiple smaller iterations. As first step, we are making some small changes to the visual design (https://gitlab.com/gitlab-org/gitlab/-/issues/322680) and make the navigation easier to understand (https://gitlab.com/gitlab-org/gitlab/-/issues/322687). We are also aware that our left sidebar has become too crowded with too many top-level/sub-level items, and we are validating an idea that would be the first step for us to get out of this (https://gitlab.com/gitlab-org/ux-research/-/issues/1421). There are already a few follow-up ideas (e.g. grouping items into larger groups or pinning most important items per user), but they are a bit further away.
I hope that gives you a better understanding how our UX team at GitLab thinks about this. Let me know if you have any other feedback or ideas
> One of the most problematic areas, as you perfectly mentioned, is the left sidebar. We already enable project administrators to remove items from there by going to Settings -> General -> Visibility, project features, permissions, and then toggling off the areas you would like to remove. However, this is a project setting, not per user.
I can't believe I missed that. Is it a somewhat recent feature? I could swear I could not disable everything a few months ago when I tried it (I remember not being able to hide Operations, Analytics, etc.). Maybe I just went too quickly through the visibility options.
Well I apologize, it is indeed feasible to hide the features I don't use. Thanks for the heads up.
> I hope that gives you a better understanding how our UX team at GitLab thinks about this
Thank you for these explanations. Although I think some pages are too bloated, I like the fact that you act on it. And to be honest, I've always found each iteration of the UI better than the previous one, so there's that.
Yes and no We did have this section, but only about half of the items in the left sidebar were initially included. This was an oversight from our teams, and we managed to get most of that fixed over the last 2 milestones (more refinements coming up as well, so if you have any suggestions, feel free to create an issue).
> And to be honest, I've always found each iteration of the UI better than the previous one, so there's that.
That's amazing to hear and proves our high focus on iterating fast and improving each time. Thanks so much for your feedback!
[1] https://civilization.fandom.com/wiki/Civilizations_(Civ6)
Whenever I use the drive file browser I miss Apache directory listings.
That new giant F button in the tab bar always has a number on it that disappears when I tap, but there's no obvious content for me to look at there.
Desktop FB is pleasant. No fake notification badges.
I avoid the app because I think it's data harvesting, but the mobile site is such a pain. Sometimes it's the little things like writing "the..." in the text box, and accidentally tagging "Theodore" because of a generic word. I have to type all my Facebook comments in a separate text editor.
Data transfer isn't a major factor, it's all delay in their servers.
I usually have time for everything and sometimes even to research and do my option trades in the time it takes to check and pay 3 cards with chase. Cancelling chase sapphire soon in part because of that (I have another similar fee card with a normal site), and stopping the use of another no-fee card purely because of the site.
If we ever interview someone who worked on this atrocity I'd have to recuse myself :)
- JIRA (Need I say more)
- AWS new UI: For some reason, recent changes by AWS in their UI/Console is terrible. When I open S3, the buckets now take few seconds to appear compared to before when it was instant. The number of buckets are similar. WTF AWS.
What does bloated website actually mean? What is characteristics of a bloated website?
I could imagine something that bloated for you, might not be bloated for me? Are there any metrics that quantifies this?
In the background, it can be that your browser is making requests to heaps of tracking services and CDNs when loading the page.
I’ve basically just stopped browsing the web at this point. It’s nigh on unusable sometimes. I buy recipe books, read the NYTime’s daily newsletter, and browse hackernews. Reddit’s unbearable these days, Twitter’s a bit of a dumpster fire even if it wasn’t bloated (I don’t have social media accounts to curate the experience either), and god forbid you click a link on a news aggregator site like Fark (where the headlines are as misleading as something your Great Aunt shared from some long chain of disinformation anyways), even with ad blockers half the pages are just absolutely bonkers.
It’s just not worth it anymore. My print copy of the New Yorker and occasionally some blog posts from here.
On the other hand, JIRA displays lists of tickets. It should be a bit snappier probably. There’s gotta be a way to display 200 tickets without it slowing my computer down. It’s certainly not the text that’s the issue, HN displays that many comments fairly easily and without hiccup.
If you, as a user, feel that the site is sluggish, unresponsive or difficult to use because it feels too crowded or has too many features, you might judge that site to feel bloated.
In a more technical sense, bloated might refer to very large (or poorly optimized) files, either javascript or static assets, required to run the thing, causing it to load slowly (i.e. at a speed that the user feels is "slowly").
So maybe a good summary might be: "does too much stuff and/or poorly optimized/coded"
Again, not exactly a tight definition, but UIs are inherently characterized by the perception of their users.
I know it is subjective thing that is the reason why I asked this question. For me all UIs are bloat the CLI experience is the best, for someone else, it might not be.
The bloating problem from my perspective might just be a UX problem that the framework you use or the library you use.
It is difficult. It's probably the central difficulty of UIs on the web. However what I said is still true: The quantification inherently rests on users' perception of their experience. This is more a question of psychology than of computer science.
What percentage of your users feel that the app is too slow or confusing? What percentage would you be okay with? If you're selling some kind of enterprise software whose users will not be the buyers but rather the buyers' employees, who cares if it takes 30 seconds to load every page and another 30 seconds to find the widget you need once it's loaded?
The definition of "too bloated" is a function of both how users feel and also how much value you get from reducing it, and the measurement of "bloated-ness" (as well as many other parameters of UX) can't happen without getting feedback from users, which means there's a pretty strict limit to how much we can understand just from discussing theory.
And just to further complicate things: If I'm making a richly featured web app (something like Slack for example) I have a predefined set of features that will need to be implemented using javascript. Obviously there are many parameters that determine how well these are designed and implemented and optimized, all of which will affect how the users perceive the app as a whole. However, what about users (say, some plurality of HN readers) who intrinsically value websites which function without javascript, and will always consider those that don't to be bloated; What use is it to quantify that if it's always going to run up against my stated goal of building a chat app in the browser?
All the stuff that is extraneous to the reason you visited the webpage would be bloat.
Think about a news article on [insert your favorite news website]. Often before you can see the article, you have to process a bunch of JS (SPA framework, tracker code, etc) before the text loads and you can get the info you were looking for.
Sure, adding tracker and other stuff might be a jarring experience. But I think you can have all the stuff in your app and make it async so that it does not impact the core functionality of the app.
A page of text should just be a page of text. That's the beauty in sites like hackernews and craigslist. They are down-and-dirty, just-the-facts, we'll-get-out-of-your-way websites.
You see bloat when you start adding stuff that obscures that core functionality and/or slows down/degrades the experience.
Of course, the core functionality changes depending on who you are. The customer might just want to read the article, but the sales team wants to get you to see/click on ads and the marketing team wants you to signup for their newsletter. So, one person's bloat is another person's core functionality.
For example:
* Page should load to a useful state (could still be background tasks) in under 2s
* Clicking a link should take me to the next page and it should be useful in under 2s
* Making a change to a config item and saving the update should complete in under 2s
If your service can do that, you customers will feel like your system is fast. On the other hand, this is generally considered a technical crowd and we probably look at resource consumption. A site that eats 25% of one of my CPUs and requires 128M of RAM just to load is going to check my box for bloated.
Every time I go back to check my old Gmail (I use it for signing up to products I know will spam me) I get very irritated at everything.
Dell/HP driver download sites
- JIRA
- AWS
- GCP
- Slack
- Gmail
- News
- Recipes