Denise (![[staff profile]](https://www.dreamwidth.org/img/silk/identity/user_staff.png)
denise) wrote in ![[site community profile]](https://www.dreamwidth.org/img/comm_staff.png)
dw_maintenance2020-07-26 06:43 pm
denise) wrote in dw_maintenance2020-07-26 06:43 pm
Update on entry/comment page changes
We are aware of the issues that people are having with the entry/comment page changes and we are working, and will continue to work, to fix them over the next few days. This entry covers what we know, what you need to know, and what we're doing to fix things.
* If your font size is unspeakably large, please check both your browser default font size and check the current magnification level you've set for Dreamwidth and Dreamwidth journals. Most browsers will remember your last-used magnification setting on a per-website basis, so if you changed your magnification on DW in the past because the site was setting a smaller-than-browser-default font size, it is probably now too large now that we are using the browser default font size. This has been the root cause for lot of the people reporting extremely large font sizes, so take a moment to check.
* A number of page elements in all of the available site skins are either missing background colors or are having the wrong background colors applied to them (highlighted backgrounds too bright, wrong color, bad contrast, etc). This is a bug and will be fixed as soon as possible.
* A number of page elements have the wrong padding, margins, or line-spacing. This is a bug and will be fixed as soon as possible.
* User-supplied CSS in entries (character sheets, tables of contents, icon tables, etc) are inheriting wrong padding, margins, or line-spacing. This is a side effect of some of the other CSS changes we made: we try to exempt user-provided CSS from sitewide CSS changes, but sometimes when we change the order in which site CSS files load, things behave differently even if we didn't deliberately change any of the CSS that targets those elements. I don't know how much of that we're going to be able to fix, because nobody has had the time to look at it yet, but please give us a week or so while we figure out what's ultimately causing it and see what we can do.
* Third party extensions such as userscripts, Stylish changes, custom site skin CSS, and the like have changed how they behave -- this is due to the changes we've made to the underlying HTML structure of the page, more about which in a moment. The creator of those third-party styles and add-ons will need to figure out the changes that need to be made in order to get it working again; you can let the author of the specific one that you're having the trouble with know that we're happy to help them if they need assistance.
People have reported a photosensitive-vertigo (PSV) reaction (that's the fancy shorthand for "this website makes my eyes/head/etc hurt, gives me migraines, etc") to the new version of the entry/comment pages. PSV has multiple causes, each individual's PSV triggers are different, and scientific research into what the most common causes/triggers are is in its infancy. There's no single checklist or automated test for design elements that are most likely to cause PSV issues. We've done our best to create one, and we check everything against that checklist, but it will never be perfect, because things that fix one person's PSV can trigger another's.
The good news is that all the research points to PSV problems being a sort of collection of cumulative triggers: with the exception of strobe-related epileptic issues (which are actually a completely different section of the brain and different underlying cause), most of the very small number of people who are likely to experience photosensitive vertigo issues can tolerate one or two things that might be a PSV trigger on any given page and only begin to experience issues when things go past a certain threshold, which is why small changes can cause or fix problems.
We believe the vast majority of the PSV-related issues people have been experiencing are due to the above display bugs -- the three most common causes of PSV issues are certain intervals of line spacing (bug, line spacing changed for some elements due to misinherited CSS), missing or wrong contrast colors in specific spots (bug, a lot of things got the wrong colors due to misinherited CSS), and tall blocks of text or page elements that are consistently wider than about 80 characters' worth of size[1] (bug, some things are the wrong width due to misinherited CSS).
We are reasonably confident that fixing those display bugs will fix the problem for about 90% of you who are reporting issues, at which point we'll be able to get better data on how many people continue to have problems after the known-to-be-likely causes are handled. If your photosensitive reaction is severe enough that it would harm you to keep using the new version of the entry/comment page over the next few days while we push fixes as we can fix them, you can go to the beta features page and enable the button under the second heading, "Temporarily revert updated journal page components", for a few days. Please then go back to that page in a few days, after we're able to fix the display bugs, and turn it back off again so that you're able to evaluate whether those fixes are enough to fix the issues you're having; if they don't, we will want to hear about it then.
We are not ignoring the fact that people are reporting photosensitive vertigo issues, but you don't need to let us know until we fix the display bugs I've listed above. We're very sorry for the problem.
If you've had particularly severe photosensitive reactions to any website, not just DW, in the past, I'd like to urge you to activate the beta testing flags whenever we mention the release of a new one indw_maintenance, because that will let you let us know when our methods of avoiding the problems aren't sufficient for a particular change or redesign as early as possible in the process. We use both the ongoing beta feature sets and one-off new beta feature sets as a method of trialing new designs for many more uses and use cases than we can catch in our extensive pre-beta testing, and the more people who activate each beta feature set, the better data we get.
[1] This would take a whole lot of digression to explain: the short version is that human brain expects certain visual ratios in content blocks based on how large the text we're reading is, and for some of us, when we don't get them, our eyes or brains complain. This is about half the reason why every site skin without a sidebar restricts text to a certain length across even very wide screens, the other half being that long unbroken lines of text are bad for triggering dyslexia issues. Side note based on this principle: if you're having photosensitive vertigo issues and have installed an extension or user style to increase the width of text or of site elements because you're using a higher resolution monitor and have the site window fullscreened, you may be better off disabling the extension, and instead reducing the margins by tiling your windows. It's not a guarantee, because again everyone's triggers are idiosyncratic to an extent, but it is one of the very few triggers that's widely identified.
An extremely common question people are asking is why we can't just let people continue to use "the old version". This particular change isn't a new site skin or a purely-visual redesign: it's part of the ongoing-since-2012 process of converting the underlying code that generates the site to use industry-standard, widely-available frameworks instead of trying to manually reproduce 20 years' worth of advances in frontend web development by hand with a team of only a few people. Every page on the site needs to be converted. We've been doing it bit by bit, and this particular change has been the conversion of the entry/comment page -- probably the second-most-complicated conversion we have still outstanding.
We've explained this ongoing project before, but since it's been long enough that people either don't remember the explanation or weren't DW users yet: These changes are necessary, because without them, we can't make any of the changes that people ask us for without something that should be a few days' work turning into a nightmare of months upon months of debugging, edge cases, and programmer tears. It isn't possible to keep "the old version" of the underlying code, because the entire purpose of the changes are so that we can finally, finally stop maintaining a custom HTML-like markup "language" that was created in 1996 and that we, LJ, and other sites using either the LJ or DW code are the only sites on the internet to ever use. It's not only a massive barrier to entry for people who want to start contributing to the project, it's frozen with the features that were available on the web in 1996 and every single modern web feature we or our predecessors have been able to bolt on top of it since then, including such basic things like "leave a comment without having to load a separate page" and "make it possible to choose a user icon and see which icon you chose", is one more layer of things that have the potential to break in modern browsers, needing us to spend hours or days diving into bugs and issues that should be a few moments to fix. We cannot keep the old version, because the old version is a twenty-year-old bolted-together mess that has become impossible to maintain.
This is pretty apocalyptic language, I know. We've all resisted using that sort of language before, because when we do it sounds like we're saying that the site is made of silly string and tinfoil and may explode at any second or that we don't know what we're doing. (The site is not made of silly string and tinfoil, it is not likely to explode at any second, and we know what we're doing.) However, it's become clear that by not saying, very bluntly, "if we do not modernize certain parts of the codebase, we will eventually never be able to do anything to the site other than chase various display glitches ever again, and that day is a lot closer than it should be", we've failed to convey the urgency motivating those modernization projects.
It's virtually impossible to keep a page's styling looking similar when you're going from one underlying modern framework to a different underlying modern framework, much less when you're converting the eldritch probably-haunted katamari of handwritten-in-a-text-editor-in-1999 HTML and CSS that lies at the heart of every page on the site we haven't converted yet. Part of the reason the conversion and modernization is going so slowly is that we're trying very, very hard to keep as much of the converted pages looking like the old code-katamari version as much as is possible within the limits of what technology is possible; when it's not possible to match the old visuals (such as when the old visuals rely on HTML tricks that are old, outdated, or deprecated), we've been trying to modernize smartly and in ways that get us useful benefits.
I won't lie, another part of the reason this conversion project is going so slowly is because it is exhausting, demoralizing, and discouraging to receive a flood of abuse after every page or page-element conversion when people assume the worst possible motives or accuse us of making changes in bad faith. We are human beings who work on this project (mostly in our spare time) because we believe in what we do here, but it's incredibly hard to motivate yourself to do something that you know has a 100% chance will get you cursed at, called a moron, and accused of not caring about people.
We've left the comments to the previousdw_maintenance post screened, and I have set all comments on this post to automatically screen, because we can either moderate the public discussion or we can fix the problems people are cursing at us over, and right now we are choosing to put our attention to fixing the problems. Yes, we want to hear when there's a problem, and we want to hear people's feedback about changes we make and how we can make them work better for you, but there's a difference between "the display of this thing changed, was that intentional?" and "this is fucking garbage and you should be ashamed of yourself". (I am only slightly paraphrasing.)
At this point, please wait for us to fix the already-reported issues before you report any additional visual, display, or design problems: the vast majority of issues that people have reported in the last six hours are duplicates of things that have already been reported, and the fixes for the first round will affect the fixes for subsequent rounds.
*Facebook has placed DW on its blocklist of prohibited domains to link to. This means nobody will be able to link to any Dreamwidth page from your Facebook account. We don't know why they made that decision, and they deliberately make it difficult to appeal or contact them about it. We'll do what we can to try to get answers from them later in the week once we've handled the above collection of issues, but if you're a FB user affected by this, you may want to contact them as well.
EDIT: someone knew someone who knew someone, which is the only way this apparently ever gets fixed, so linking from FB is working again!
DISPLAY BUGS
* If your font size is unspeakably large, please check both your browser default font size and check the current magnification level you've set for Dreamwidth and Dreamwidth journals. Most browsers will remember your last-used magnification setting on a per-website basis, so if you changed your magnification on DW in the past because the site was setting a smaller-than-browser-default font size, it is probably now too large now that we are using the browser default font size. This has been the root cause for lot of the people reporting extremely large font sizes, so take a moment to check.
* A number of page elements in all of the available site skins are either missing background colors or are having the wrong background colors applied to them (highlighted backgrounds too bright, wrong color, bad contrast, etc). This is a bug and will be fixed as soon as possible.
* A number of page elements have the wrong padding, margins, or line-spacing. This is a bug and will be fixed as soon as possible.
* User-supplied CSS in entries (character sheets, tables of contents, icon tables, etc) are inheriting wrong padding, margins, or line-spacing. This is a side effect of some of the other CSS changes we made: we try to exempt user-provided CSS from sitewide CSS changes, but sometimes when we change the order in which site CSS files load, things behave differently even if we didn't deliberately change any of the CSS that targets those elements. I don't know how much of that we're going to be able to fix, because nobody has had the time to look at it yet, but please give us a week or so while we figure out what's ultimately causing it and see what we can do.
* Third party extensions such as userscripts, Stylish changes, custom site skin CSS, and the like have changed how they behave -- this is due to the changes we've made to the underlying HTML structure of the page, more about which in a moment. The creator of those third-party styles and add-ons will need to figure out the changes that need to be made in order to get it working again; you can let the author of the specific one that you're having the trouble with know that we're happy to help them if they need assistance.
VISUAL/ACCESS ISSUES
People have reported a photosensitive-vertigo (PSV) reaction (that's the fancy shorthand for "this website makes my eyes/head/etc hurt, gives me migraines, etc") to the new version of the entry/comment pages. PSV has multiple causes, each individual's PSV triggers are different, and scientific research into what the most common causes/triggers are is in its infancy. There's no single checklist or automated test for design elements that are most likely to cause PSV issues. We've done our best to create one, and we check everything against that checklist, but it will never be perfect, because things that fix one person's PSV can trigger another's.
The good news is that all the research points to PSV problems being a sort of collection of cumulative triggers: with the exception of strobe-related epileptic issues (which are actually a completely different section of the brain and different underlying cause), most of the very small number of people who are likely to experience photosensitive vertigo issues can tolerate one or two things that might be a PSV trigger on any given page and only begin to experience issues when things go past a certain threshold, which is why small changes can cause or fix problems.
We believe the vast majority of the PSV-related issues people have been experiencing are due to the above display bugs -- the three most common causes of PSV issues are certain intervals of line spacing (bug, line spacing changed for some elements due to misinherited CSS), missing or wrong contrast colors in specific spots (bug, a lot of things got the wrong colors due to misinherited CSS), and tall blocks of text or page elements that are consistently wider than about 80 characters' worth of size[1] (bug, some things are the wrong width due to misinherited CSS).
We are reasonably confident that fixing those display bugs will fix the problem for about 90% of you who are reporting issues, at which point we'll be able to get better data on how many people continue to have problems after the known-to-be-likely causes are handled. If your photosensitive reaction is severe enough that it would harm you to keep using the new version of the entry/comment page over the next few days while we push fixes as we can fix them, you can go to the beta features page and enable the button under the second heading, "Temporarily revert updated journal page components", for a few days. Please then go back to that page in a few days, after we're able to fix the display bugs, and turn it back off again so that you're able to evaluate whether those fixes are enough to fix the issues you're having; if they don't, we will want to hear about it then.
We are not ignoring the fact that people are reporting photosensitive vertigo issues, but you don't need to let us know until we fix the display bugs I've listed above. We're very sorry for the problem.
If you've had particularly severe photosensitive reactions to any website, not just DW, in the past, I'd like to urge you to activate the beta testing flags whenever we mention the release of a new one in
dw_maintenance, because that will let you let us know when our methods of avoiding the problems aren't sufficient for a particular change or redesign as early as possible in the process. We use both the ongoing beta feature sets and one-off new beta feature sets as a method of trialing new designs for many more uses and use cases than we can catch in our extensive pre-beta testing, and the more people who activate each beta feature set, the better data we get.[1] This would take a whole lot of digression to explain: the short version is that human brain expects certain visual ratios in content blocks based on how large the text we're reading is, and for some of us, when we don't get them, our eyes or brains complain. This is about half the reason why every site skin without a sidebar restricts text to a certain length across even very wide screens, the other half being that long unbroken lines of text are bad for triggering dyslexia issues. Side note based on this principle: if you're having photosensitive vertigo issues and have installed an extension or user style to increase the width of text or of site elements because you're using a higher resolution monitor and have the site window fullscreened, you may be better off disabling the extension, and instead reducing the margins by tiling your windows. It's not a guarantee, because again everyone's triggers are idiosyncratic to an extent, but it is one of the very few triggers that's widely identified.
"THE OLD VERSION"
An extremely common question people are asking is why we can't just let people continue to use "the old version". This particular change isn't a new site skin or a purely-visual redesign: it's part of the ongoing-since-2012 process of converting the underlying code that generates the site to use industry-standard, widely-available frameworks instead of trying to manually reproduce 20 years' worth of advances in frontend web development by hand with a team of only a few people. Every page on the site needs to be converted. We've been doing it bit by bit, and this particular change has been the conversion of the entry/comment page -- probably the second-most-complicated conversion we have still outstanding.
We've explained this ongoing project before, but since it's been long enough that people either don't remember the explanation or weren't DW users yet: These changes are necessary, because without them, we can't make any of the changes that people ask us for without something that should be a few days' work turning into a nightmare of months upon months of debugging, edge cases, and programmer tears. It isn't possible to keep "the old version" of the underlying code, because the entire purpose of the changes are so that we can finally, finally stop maintaining a custom HTML-like markup "language" that was created in 1996 and that we, LJ, and other sites using either the LJ or DW code are the only sites on the internet to ever use. It's not only a massive barrier to entry for people who want to start contributing to the project, it's frozen with the features that were available on the web in 1996 and every single modern web feature we or our predecessors have been able to bolt on top of it since then, including such basic things like "leave a comment without having to load a separate page" and "make it possible to choose a user icon and see which icon you chose", is one more layer of things that have the potential to break in modern browsers, needing us to spend hours or days diving into bugs and issues that should be a few moments to fix. We cannot keep the old version, because the old version is a twenty-year-old bolted-together mess that has become impossible to maintain.
This is pretty apocalyptic language, I know. We've all resisted using that sort of language before, because when we do it sounds like we're saying that the site is made of silly string and tinfoil and may explode at any second or that we don't know what we're doing. (The site is not made of silly string and tinfoil, it is not likely to explode at any second, and we know what we're doing.) However, it's become clear that by not saying, very bluntly, "if we do not modernize certain parts of the codebase, we will eventually never be able to do anything to the site other than chase various display glitches ever again, and that day is a lot closer than it should be", we've failed to convey the urgency motivating those modernization projects.
It's virtually impossible to keep a page's styling looking similar when you're going from one underlying modern framework to a different underlying modern framework, much less when you're converting the eldritch probably-haunted katamari of handwritten-in-a-text-editor-in-1999 HTML and CSS that lies at the heart of every page on the site we haven't converted yet. Part of the reason the conversion and modernization is going so slowly is that we're trying very, very hard to keep as much of the converted pages looking like the old code-katamari version as much as is possible within the limits of what technology is possible; when it's not possible to match the old visuals (such as when the old visuals rely on HTML tricks that are old, outdated, or deprecated), we've been trying to modernize smartly and in ways that get us useful benefits.
CIVILITY
I won't lie, another part of the reason this conversion project is going so slowly is because it is exhausting, demoralizing, and discouraging to receive a flood of abuse after every page or page-element conversion when people assume the worst possible motives or accuse us of making changes in bad faith. We are human beings who work on this project (mostly in our spare time) because we believe in what we do here, but it's incredibly hard to motivate yourself to do something that you know has a 100% chance will get you cursed at, called a moron, and accused of not caring about people.
We've left the comments to the previous
dw_maintenance post screened, and I have set all comments on this post to automatically screen, because we can either moderate the public discussion or we can fix the problems people are cursing at us over, and right now we are choosing to put our attention to fixing the problems. Yes, we want to hear when there's a problem, and we want to hear people's feedback about changes we make and how we can make them work better for you, but there's a difference between "the display of this thing changed, was that intentional?" and "this is fucking garbage and you should be ashamed of yourself". (I am only slightly paraphrasing.)At this point, please wait for us to fix the already-reported issues before you report any additional visual, display, or design problems: the vast majority of issues that people have reported in the last six hours are duplicates of things that have already been reported, and the fixes for the first round will affect the fixes for subsequent rounds.
UNRELATED BUT WHILE I'M HERE
*
EDIT: someone knew someone who knew someone, which is the only way this apparently ever gets fixed, so linking from FB is working again!
no subject