flak rss random

edge vs chrome

Some observations regarding edge vs chrome. Not a complete investigation, some aspects not mentioned may have been outside the scope of inquiry.

battery

The topic of browser effect on battery life is a popular one. Microsoft made a big fuss about edge being better, then maybe chrome fixed some things, and some people ran some other tests and observed different results. The problem seems to be that we have data from bizarre fake tests and anecdotes from real users, and the two differ in results.

On the sites I like to spend most of my time reading, there’s no difference. I get maybe ten hours time with this browser, I get maybe ten hours time with that browser. But it’s not hard to find pages where I can measure a difference.

Conveniently, Microsoft themselves provide a nice test page. Consider this blog post about Edge UI responsiveness. Scrolling down until there’s a few gifs on the screen results in increased power consumption by both browsers, but in differing amounts. According to HWMonitor, I see about 1W when idle. When displaying the linked page, edge uses about 2W. When using chrome, that climbs closer to 3W.

One might suppose Microsoft booby trapped the page to cripple chrome, but the gifs in question are hosted by gfycat. And are actually mp4 videos in an iframe. I haven’t verified everything is bit for bit identical, but doesn’t seem to be any foul play. Though it’s possibly worth considering that if chrome is the browser of choice for developers, they might start using the newest whizbang shiny feature for chrome, but take a more conservative (low power) approach for other browsers. The more code you let someone run on your computer, the more code they’re going to run.

It’s difficult to get a super precise reading because the CPU is bouncing around between speeds quite rapidly, but it’s something to work with and we can cross check results. A 45 WHr battery with a load of 5W (1W CPU + 4W screen etc.) gives 9 hours run time. Increase the load to 6W and that’s closer to 8 hours. Increase the load to 7W and that’s 7 hours. Which roughly seems to match reality. I see nine or ten hours browsing very light sites. Other people using edge on heavier sites only see eight hours, and chrome users only get seven hours. The work that went into timer coalescing and whatever else in edge still seems to give an advantage, though this depends on how inefficient and poorly designed the site itself is.

Battery life is as much a function of site selection as browser selection. Leaving a tab open to the Bloomberg home page with its scrolling news ticker is going to eat a lot of power regardless of browser. Edge may be better at wasting less power, but closing the tab will provide substantially better results.

touch

Since I’m conducting this experiment on my Surface, I’m using it in tablet mode a lot. Chrome is still pretty rough in this department. A short list of things that edge does better.

Edge supports double tap to zoom. I want a small skinny column of text to fill the window, I double tap it, just like on my phone. Works great. Chrome can’t do this, forcing me to pinch and zoom by hand. It’s a small thing overall, but it’s a pretty big part of expected tablet behavior. Not sure if the iPhone invented this trick, but it’s been ten years.

The onscreen keyboard provides suggestions in edge, but not chrome. I guess you need to provide some hint back to the system that input is focused on a text input? Makes writing longer comments much more convenient instead of trying to get every letter correct.

Speaking of, sometimes chrome misinterprets inputs. I hit ctrl-L to jump to the location bar, and that happens, but then it overwrites the text with the letter l. Occurred repeatedly, even being very careful to only tap the key once. Didn’t happen in edge, switch back to chrome, still happens. Underlying cause unknown.

Chrome does this thing where every link has a hit box that’s bigger than the link. So if you tap near a link, it activates the link. What happens when two links are close to each other? The one hit box covers the other, and it becomes impossible, not difficult, impossible to tap the second link. This is not annoying. It’s appalling.

General weirdness in chrome. Strange things happen, like what you might see if an invisible extra finger were on the screen. Drags turn into pinches, selections are randomly unselected, etc. I don’t see anything like this in other apps, so I don’t suspect the screen. It seems chrome specific. Restarting chrome often helps.

other

Edge has a nice dark theme. I prefer white on black text at night. There’s like 99 extensions that do something similar for chrome, but it’s nice that it’s included, and none of the chrome extensions seem able to darken the location bar.

Edge includes a reader mode, again without an extension, but alas it’s somewhat broken. Lots of articles are truncated before the end, which is less than great. Also on medium, unless you scroll all the way to the end before hitting reader mode, it won’t have any images because they’re lazy loaded. Blame medium for that I guess, but definitely limits the utility of the feature.

In chrome, zoom settings are per site, which makes a lot of sense. Different sites scale differently. Edge seems to only have a global setting, which requires more constant adjustment.

PDF viewers are a very personal choice, but if we’re going to use a browser, I’d say edge wins easily. It scrolls without lag and feels very smooth. When reading a paper like this one on CFI, it reacts as if there were a real sheet of paper under the glass. Chrome is kinda janky, and responded to a pinch to zoom with some horrific oscillating flashing behavior. (Damnable invisible finger!) On a reload, it behaved better, but always with some lag. Something weird going on there.

The uBlock Origin logger UI seems broken in edge. Alas. Maybe a one time bug? I think it works now.

security

Also something to consider.

Posted 28 Jun 2017 23:17 by tedu Updated: 30 Jun 2017 19:01
Tagged: software web windows