Topic on Talk:Reading/Web/Desktop Improvements/Features/Limiting content width/Flow

Tl;dr - please ditch the margins and bring the auto-scaling, because the proposal makes things worse on a modern desktop.

Dear All,

MediaWiki has a very archaic UX based on hardcoded numbers of pixels, small font, clutter and laughably small thumbnails with the 220px width coming from an era of 15", 800x600 monitors in the year of 2004. It is high time to improve.

Unfortunately, this very proposal goes in a very '''bad direction''' and makes life on a modern desktop only worse. Instead of improving the paradigm and giving a website adaptive to the users' devices, it is trying to recreate and already old tablet look of 2010 - with hardcoded pixels, unadequate for modern desktops. It is a lazy design, copying the worst errors of today websites.

Firstly, it ignores the nature of the website and the latest design trends. The fad of tablet-friendly, wide-margin design has been very noticable in the "content consumption" pages, especially when the high hopes were placed with a new form factor (tablet) and cost-cutting with the "mobile-first" design. But it never made it right.

It has never landed so well in the "workspace", where every bit of space is being apreciated; and nowadays you can see some calls for a retreat to different paradigms in design. Moreover, we can afford to not follow the mobile-centered design, and provide the UX adaptive to the users' device.

Secondly, it is not bold enough. The proposal disappoints me personally, because it keeps the MediaWiki archaic where it is really, REALLY long in tooth (non-responsive design based on fixed pixel numbers, UX clutter, too small thumbnails on modern desktop), while it is blindly following a yesteryear's trend, adding superficial hurdles to read and write the content because of IMVHO misguided pursuit for legibility. It is like taking the worst from the two worlds.

The most glaring problem is the usage of fixed px and wide margins instead of auto-fitting on a computer screen. This issue will be only growing with the growing pixel width and physical sizes of the desktop screens. Even already on a WQGA or 4K monitor the proposed page '''looks ugly''' with nonsensical wide margins. And it will get only worse.

The pursuit is IMHO misguided, as it is focused on "limiting the text width", rather than on the overall experience (especially font SIZE, spacing) and the general UX of a person using a modern, large desktop screen.

The proposal quotes some random text of an IBM researcher. It is not even a scientific paper, it does not seem to be even reviewed or reproduced; we don't know which year it is from and what actual environment does it represent. Are these thoughts based on a 14" VGA screen, 14" 1368x768 laptop, or a 28" WQGA setup? Was font scaled as well? What was the font, actually? We don't know.

I was honestly trying to find a source of these margins and "up to 75 characters in line" space, but all I got was an echo chamber with no contemporary, real research to back it. But I failed, I encourage everyone to provide me some papers.

What I got was the insight that actually:

* it's the size, font, spacing which really matter

* long lines seem to be pretty fine on monitors

* some claims are rooted in the old guidelines from the times of printing press and paper - not really relevant for e.g. contemporary use cases, sharpness and e.g. distance from the screen.

https://www.viget.com/arti.../the-line-length-misconception/

https://pielot.org/pubs/Rello2016-Fontsize.pdf

https://uxdesign.cc/legibility-how-to-make-text...

As you can see, the 18pt font was recommended for Wikipedia. No margins or artificial limits. Just size.

Any contemporary research on actual, contemporary desktops (Full HD, WQGA and 4K screens of 24"-29" size) would be appreciated.

On top of that, please test on some modern desktop: 27", 28" WQGA and 4K monitors, as well as on modern 14" laptops with the same WQGA or 4K. These are the standards in the beginning of the decade; not complying with them makes any proposal irrelevant.

When you will have tested, you will know that the proposed page is actually uglier. :/

Thus, as it has been mentioned before - we desktop users have our high pixel density, large screens for a reason. And we will be having them more and more often, because these are current IT trends. Instead of imposing silly margins, please work on proper auto-scaling and an adaptive UX design.

@Aegis Maelstrom it would be great if you could provide examples of the user experience you envision. I understand that you disagree with the approach we're currently taking but I am unable to fully understand your proposal. It would be awesome if you could reference other sites that, in your opinion, properly balance readability (across a variety of screen sizes) with density/utility. Or better yet provide some mockups or a prototype. We have done an extensive review of popular information/content sites and have found nothing along the lines of what you describe, unfortunately.

Regarding your statement: "The proposal quotes some random text of an IBM researcher. It is not even a scientific paper" — you can find details about the research we've quoted here: https://link.springer.com/chapter/10.1007/11555261_59. It was a conference paper, presented at the IFIP Conference on Human-Computer Interaction in 2005.

Regarding not being "bold enough" — we do not have a goal about being bold. Nor is it a project goal to ensure that 4K monitors are filled with content. We are trying to make small, considerate improvements to the reading experience. Limiting line-length is a critical aspect and therefore we have decided to tackle it up-front. There are many more changes to come.

@AHollender (WMF) hi there, and thanks for the detail of the research you are using. Let me focus on this one now. I am afraid you it is proving my point, and I encourage everyone to revisit the paper you quote, as it 1) is quite anachronic 2) actually says it is far from being conclusive.

Firstly, the experiment setup is quite distant from the use case of 2021, or 2030. The paper from 2005 describes an experiment performed with two IBM T41 laptops (model from the year of 2003), with most probably standard 14" screens with an aspect ratio of 4:3 (nothing like contemporary 16:9 or 16:10) and XGA screen resolution (1024x768). What is more, a somewhat limited part of the screen was used (80%/9" vs. 40%/4.5") with a too small, 10 pt Verdana font. This is a setup nowhere near a modern desktop, or even a modern laptop. It is several generations behind the screens we need to address in the 2021-2030 decade.

Secondly, even under these circumstances the experiment was not really conclusive. "Making the paragraph width wider reduces the number of return sweeps required for reading [and increases the probability to read to the end!], but increases regressions and decreases [information] retention." The reason of the higher retention despite the shorter reading was not entirely clear, and "It would be interesting to repeat the histogram analysis...", as the text says. I can only repeat this call for more research, while recommending using contemporary and future screen sizes, also looking for a proper font, font size and spacing.

Besides these two points, I think the research itself is in some strong need of a replication, as e.g. results in the retention tests (average 1.25 and 1.75 good answers out of 3 questions!) say that something went REALLY wrong with the legibility of the text or the quiz itself - regardless of the line width).

Summing up: we need a contemporary, adequate test, with contemporary people, contemporary computers and some proper typography. 10 pt Verdana on some 17 years old XGA laptop is FAR from being a "modern desktop" or anything we should be aiming for.

@Aegis Maelstrom I agree entirely. I setup this phab task for that very reason: https://phabricator.wikimedia.org/T261174. I wish that we had greater resources at our disposal, however I am optimistic that we can find some clever way to do more research in an inexpensive way. I'm curious what gave you the impression that we are satisfied with the current research? The tone of your message, to me, feels like you are accusing us of not doing our research properly. I have tried to be specific about the limitations of the research we have found thus far, and try to always mention that we want to do more research.

Secondly I am curious what you think about this — do you think these are valid assumptions?

I agree that more research should be done and/or found. We are indeed making some assumptions however I think it is worth being specific about the assumptions we're making. The first assumption is that if longer line lengths were beneficial to reading we would see them out in the world more often. I am hard-pressed to find a popular website, or any printed material, that exceeds the suggested line length range. I am not saying that everyone else is automatically right about line-length, but I would expect to at least find ''some'' examples of line lengths in the 200+ cpl range if that was indeed better for reading (if you can find some please post them here). The second assumption is that the extensive amount of research on this topic has intentionally been focused on the optimal range. Without speaking to the folks who conducted the research (and I have started to reach out to them) there is no way of knowing why they didn't study longer line-lengths, but I do feel like this is a relatively safe assumption again considering how many studies are conducted within the same range. Lastly I think it would be valuable for you (and others) to further investigate your own personal aversion to the max-width. We have had some great conversations with those willing to dive a bit deeper on this topic and a few folks have found that there are other factors (mainly frustration about blank space) that are informing their opinions, rather than an actual dissatisfaction with shorter line-lengths in and of themselves.

(originally posted here)

Thanks for your thought partnership on this matter