Reminder: Beta features are available to test!

[denise]

It's been a while since we've reminded you of the existence of the Beta Features collection. This is where we make new features and major page redesigns available to people in advance of changing them for the whole site, so people can help find the weird edge case bugs, let us know about usability and accessibility problems, and generally bang on certain changes before we make them the default behavior so the deploy runs more smoothly.

I'm posting about beta test opportunities because 2023 has been the Year Of Finishing That Last 10% On So Many Things, and part of that is us making a real push to get some more things out of beta. In particular, the redesigned Create Entries page has been in beta for an absolutely mortifying length of time by now, and we are absolutely determined to finish up the last few things that are blocking us from releasing it and get it out the door relatively soon.

This is the last call for "I'm having a problem with this page that will keep me from being able to use the site if you make this the only way to update". If you haven't enabled the New Create Entries Page beta to test it, please consider doing so soon! It's a full, ground-up rewrite and redesign of the update page in order to modernize it, make it easier to maintain, significantly improve usability and accessibility, and generally give us a lot more opportunities to improve the posting process. In particular, the page is fully customizable: if you hit the "Settings" link in the upper right corner of the page, just above the "Subject" field, it will let you customize which panels show when you load the page, change some visual settings about how the page behaves, and (if you scroll down) move around the individual panels so they're displayed in a place you find logical for your workflow. We think it's a massive improvement from the old update page, and we're ready to do the last bits of work we need to do in order to switch everyone over.

We will soon be removing the New Create Entries Page beta and making the new, improved version of the page the only option for everyone who uses the site. This is your last chance to let us know if something is very broken for you: checking now, and reporting any problems you might have, will prevent future disruption to your use of the site. Please take a moment and enable the beta so you can help us find any last lingering issues.

The other major beta we have running right now is the New Inbox Page conversion. This is part of our ongoing modernization efforts: Dreamwidth itself was launched in 2009, but because it's based on LiveJournal's open source code, parts of the site date as far back as February of 1999. (If you're as bad at math as I am: that means three months from now will be the 25th anniversary of the code that runs the site existing in some fashion. Please don't tell me if my website's source code is older than you are; I already feel old as hell just thinking about it.) As you can imagine, the technology that runs the internet has changed significantly in the last 25 years, and we've been engaged in a long, slow-motion project of updating the code that generates all of the pages on the site to use more modern technology. Like with the new Create Entries page, this makes it seriously easier to maintain and allows us to benefit from the accessibility, usability, and bugfixing work that other people using the same technologies have done: the old way of doing things was unique to DW, while the new way of doing it uses more widely-used frameworks and modules so we can benefit from many other people's work.

Converting site pages to the new way of doing things involves a complete rewrite from the ground up, though, and while we try to keep the two versions looking as similar as we can, it's not always possible: some changes to how the pages look and feel are inevitable. There's also, of course, the opportunity for new and exciting bugs. So, when we've been converting major pages, we run it through a few rounds of beta first. The inbox is one of the last remaining big user-facing conversion projects, and we wanna get it done and dusted so we can pick off the last outstanding pages to convert and finally be done with this massive albatross of a project we've been working on for-freaking-ever. (We're about 80% done, and this fact is very exciting after having been working on it for so long.)

If you haven't enabled the New Inbox Page beta to test it, please consider doing so soon! (And if you had previously enabled it and found a bug that made you turn it off, please try turning it back on: we've done a lot of bugfixing work.) Because of how old the old Inbox page was, this is one of the more dramatic "things will look different" conversions we've done and the new version may take a little while to get used to, but all the major functionality should match and it includes a number of bugfixes that the old version of the inbox doesn't have.

We will soon be removing the New Inbox Page beta and making the new, improved version of the page the only option for everyone who uses the site. This is your last chance to let us know if something is very broken for you: checking now, and reporting any problems you might have, will prevent future disruption to your use of the site. Please take a moment and enable the beta so you can help us find any last lingering issues.

Finally, a few years ago we converted the journal entry view from the old code to the new code. Despite us aggressively running it through the beta system before making it the default for everyone, a few people experienced some dramatic bugs (including some wetware bugs: it triggered photosensitive migraine in a small number of people). While we worked to fix the problems people reported to us, we enabled a beta flag to let people turn the new code off (as opposed to our usual method of using the beta flag to turn the new code on).

We've done a significant amount of work on the problems people reported with the new journal entry view, including working with a number of experts on photosensitive migraine and a panel of the users who reported that the new code triggered photosensitive migraines for them. At this point, we're reasonably sure that we've fixed every issue that was reported to us, and we're making preparations to remove the beta flag that allows people to turn the new code off. (Keeping two versions of the code for this long is a lot of maintenance overhead, and we can't keep the old code around forever.)

If your account has the "Temporarily revert updated journal page components" beta flag on -- that is, if you previously experienced a site-breaking bug with the new code and had to disable it while we fixed it -- please go to the beta page and turn that beta flag off. This will switch your account to using the new code that every account created since the conversion has been running and let you verify that the particular issue you were having has been fixed. This is probably the most critical action for people who had problems with the conversion to take, because we really need to close out that beta and stop having to maintain two versions of the code, but we don't want to break things for you again!

We will soon be removing the "Temporarily revert updated journal page components" beta and making the new, improved version of the journal entry view pages the only option for everyone who uses the site. If you previously enabled the beta flag in order to temporarily use the old code while we worked to fix it, it is very important that you turn off that flag now and verify that your issues with the new code have been fixed. Please take a moment and disable that beta if you have it enabled so you can help us confirm that the problem you had is fixed.

(The remaining individual beta flag, two-factor authentication, is a very early test that was released as a beta so people can test out the process of configuring an optional two-factor authentication setup. While you can set your 2FA preferences, none of the site login pathways actually use 2FA yet, so enabling it won't do much. The final flag, "site-wide canary", is a generalized "this is the cutting edge code that we've just committed and will be released in the next site update" setting, and you should only enable it if you like living dangerously and occasionally finding that everything breaks horribly.)

For all of these beta flags, if you enable them and have problems, check the Beta Features page listing for that specific beta flag listing: there's a link to the specific entry in dw_beta where you can report problems or bugs.

Comments

[cyberbrain]

Thank you for everything you do!

I was tangentially aware that there were features in beta but didn't have any of them switched on until I saw this post. So, thank you, also, for bringing attention to that!

The new create a post page was a little shocking at first but now I'm used to it. FWIW, I think the ticky boxes for the display date, comment settings and age restriction should be ticked by default (and crossposting too, I guess, I just don't use that feature) because their sudden disappearance looks precisely like that—a disappearance. I wouldn't have thought to look for a ticky box to return them to me if I hadn't seen a comment mentioning it. I saw the Settings button but ignored it because I didn't expect it to allow me to change anything about the page's layout; most sites don't let users customize their UI, so this was unintuitive. (Generally, I feel like there's too much padding around everything, nothing fits fully on my screens like the current live pages. The current new post page is very compact and I quite like that, same goes for the inbox pages but more on that in a bit.)

The idea for moving panels is lovely! I found I couldn't move them on my tablet (Galaxy Tab S8+ with Opera) but they move fine on my PC (Windows 10, Firefox). I like the addition of the entry link panel, a custom link sounds nice. And the ticky box for a sticky entry is nice too. What I dislike is that the column of panels is on the right and gets in the way of resizing the entry text box. I can pull to lengthen it vertically but not horizontally. I do this occasionally when I'm dealing with a lot of text on the site, like right now. I know I can set the entry box to take up the whole width but that look also isn't comfortable (everything is shoved under the text box then).

Regarding the new inbox, the font is rather big (I think). Paired with the seemingly increased spacing, the page just feels bloated on my screen and on my phone screen it's needlessly long and needing to click to open the folders is an added barrier that slows down navigation (this is a common problem I have with mobile sites/responsive design). I had a moment of panic that the folders in the inbox were gone too before I realized I had a little triangle to tap. I can deal with the tap to see the folders but the font is way too big for me. I don't want every rando on the subway next to me to be able to read my screen so easily.

Maybe something like a font-size field can be added in the site settings so we can all choose what we're comfortable with? I see that there are people praising the larger font and I don't want to get in their way. (On mobile devices, we can set the font to be larger from the device settings if we want but we can't really force it to be smaller.)

Is there any way we could opt out of the responsive design bits through a (new) setting?

I am, apparently, a giant weirdo when it comes to browsing the internet on mobile devices and especially phones. I always want to look at desktop pages, even on my phone (a now four and half year old Huawei P20). I use Opera on my phone for that reason too. This is the only mobile browser that allows me to set it to desktop view and then also has word wrap to make this view navigable (and also has force enable zoom).

[regshoe]

I am, apparently, a giant weirdo when it comes to browsing the internet on mobile devices and especially phones. I always want to look at desktop pages, even on my phone

I was just (in the context of reacting to these changes) thinking I felt like the only one who found that! You're not alone :D A non-mobile-optimised page that requires a bit of zooming in/out, horizontal scrolling and resizing is still easier for me to use than a crowded, confusingly minimalist mobile-'friendly' page.

[cyberbrain]

Ahh, it's so nice to see a fellow desktop page user! *g*

A non-mobile-optimised page that requires a bit of zooming in/out, horizontal scrolling and resizing is still easier for me to use than a crowded, confusingly minimalist mobile-'friendly' page.
Exactly this, mobile pages are so weirdly designed and often they simply don't have all options/features on them at all (or they do but hidden behind several menus).

[denise]

To save myself some of the typing about the font issues (I'm kind of hitting my physical capacity for sitting up without agony after the last two days): lemme refer you to the second part of my answer here where I explained the underlying issue!

There is, currently, no possible technical way to add an opt out for the responsive design changes. (It is a very long and boring explanation about the fact that the framework we're using is the "least worst" of all the options but that does not make it exactly good; we've wrangled it into something approaching submission where it does 95% of what we need it to do and we're constantly trying to get that last 5%, but it's not always possible.) Right now, even with how much of the site still behaves very badly on mobile because we're still stuck at "80% complete" hell with the conversion project, over 80% of our traffic comes from mobile devices, and the overwhelmingly most frequent piece of feedback we receive is that people are extremely upset at how hard it is to use the site on mobile devices and in mobile browsers. Solving that problem is the #1 issue we're dealing with right now, and in order to solve it, we also need to finish the code conversion project that's been stuck at 80% forever (and ever and ever and ever). We also need to finish the conversion project because the old code is, at this point, 25 years old, extremely fragile in modern browsers, and keeps breaking all over the place -- we're doing an incredible amount of "desperately treading water because every time one of the major browsers updates or one of the technologies we rely upon has a security fix, there's a very good chance it's going to completely break half the site and we will need to drop everything and fix it immediately".

Because of the urgency of that modernization effort and the general resource constraints we operate under (which can't be fixed just by throwing more people at the problem because part of the reason we're modernizing everything is that the old code all uses nonstandard packages that nobody but us uses, so new people have an incredibly sharp learning curve and it takes a lot of help from the experienced people), we're very limited in how much work we have the capacity to do to allow people to customize individual pages that are being updated, over and above the fact that we already have so many settings and customization options that it's actively offputting to people and causes cognitive accessibility issues so we have to be very judicious in how we add additional settings. (Especially because the existing settings page is a UX disaster that desperately needs rearranging, but again, all of our capacity is going to way more urgent stuff.)

It's one of those situations where the work we're doing now is painful and tedious for us and painful and annoying for y'all, it involves a huge number of compromises and trade-offs, but it desperately needs to get done because we can't do any of the other six dozen urgent projects until it's finished. I'm sorry that some of the exact details aren't working for you, and we will continue to tweak things as we have the capacity, but for getting these specific things out of beta and advancing the modernization project, at this point we really do need to restrict the category of feedback we will act upon to "this feature literally does not function in my browser" and handle "the UX of this is not ideal for me" at some point in the future. It's not only that level of urgent, it was that level of urgent years ago and we've been delayed over and over again in finishing the project because of how much work it requires and how difficult it is to try to balance so many conflicting "the UX of this is not ideal for me" issues.

We will continue to try our absolute best to handle the most glaringly urgent of the UX issues people report, but the new update page has been in beta for twelve years, and we've been through so many rounds of feedback and adjustments on it that are the result of significant compromises, imperfect solutions, conflicting access needs, and other incredibly difficult decisions, that we have reached the point where we desperately need to draw a line and say that it's done.