Notices


Reply
Thread Tools
Posts: 22 | Thanked: 82 times | Joined on Jun 2017
#21
Just go to TJC and take a quick glance at the first page. At this very moment, I see 3 posts regarding Sailfish Browser:


Sure, one of these is an aggregate thread for all the issues, but there is something about "website gazzetta.it on default browser kills Xperia X" and "Complete freeze after FB App and Sailfish -Browser combination" that kinda makes me think that it doesn't actually work that great.

Anyway, as a true "Power User of The Internet" (thanks Zuck) I decided to run some tests, you could probably repeat that too if you somehow consider my numbers to be faked or whatever.

On PC, I'm comparing recent version of Opera (52) and Firefox 38.0.5. I didn't run JetStream on PC since I wasted waaaay to much time on remaining tests to even want to try that.

I'm comparing Edge on Lumia 650 and Firefox 38 (aka Sailfish Browser) on AquaFish. Both phones have the same SoC. I also run it on Firefox 59 through AlienDalvik (I expect overhead from the VM to yield some shitty results).

JetStream:
https://i.imgur.com/MgyUJbF.png - Edge 15 on Lumia 650 (12.861 +/- 0.61225)
https://i.imgur.com/belOD76.png - Sailfish Browser on AquaFish (12.784 +/- 0.56748)
https://i.imgur.com/qmqDY99.png - Firefox 59 on AquaFish (13.284 +/- 2.0083)

Web 3.0 Basemark on PC (Surface Pro 4, Core i5-6300U):
https://web.basemark.com/result/?4KHkpQo5 - Firefox 38 (41.71)
https://web.basemark.com/result/?4KHj27gv - Opera 52 (242.98)

Web 3.0 Basemark on phones
https://web.basemark.com/result/?4KHl0Fa1 - Edge 15 on Lumia 650 (25.58)
https://web.basemark.com/result/?QpNIlIz - Sailfish Browser on AquaFish (22.05)
Unfortunately Firefox 59 on AquaFish crashed

Key takeaways from these tests:
- despite offering a better web browsing experience (Web 3.0 test), Lumia 650 is absolute **** at running asm.js code. These results are just sad
- overhead from AlienDalvik didn't stop Firefox 59 from trashing the competition
- I wasted way too much life on proving people wrong on the Interwebz

Now, can we all stop pretending that SFOS Browser has no issues with either performance or website compatibility, please?
 

The Following 4 Users Say Thank You to pisarz1958 For This Useful Post:
nthn's Avatar
Posts: 621 | Thanked: 2,263 times | Joined on Jun 2014
#22
What do those numbers signify? Are higher numbers better than lower ones?

No one is saying there are no issues with the browser (there most definitely are, the lack of any sort of text selection in text fields is one of the most striking problems as pichlo already mentioned, as well as the problem with crashing the entire device through certain websites which is clearly documented), just that for our purposes, it works just fine and is pretty fast.

I think what pichlo and I are also implicitly trying to say is that most of these problematic websites are those that load dozens of megabytes of images on their front page, along with dozens of massive third party scripts, multiple complete font sets, and so on. The main thing browser developers have to do is not to make their browser faster, but to constantly work around non-standard to completely broken websites. If everyone just chose to stick to the standards and hire competent people for the job, or in the case of companies like Google, stick to the standards they themselves helped create instead of making their websites work only in their own browser, there wouldn't be anywhere near as many problems, and the internet would be much faster.

Of course, a broken website shouldn't make a browser crash, and certainly not the entire system, but I'd take a guess and say 99% of performance problems are to blame on incompetent website developers rather than any real issues with the browsers.

To go off on a tangent, ever since Firefox updated their browser engine to 'Quantum', I've had nothing but trouble with its performance - and now it does the same thing as all the Webkit browsers in that if you scroll too quickly, the page just goes blank for a bit until that piece is loaded. Not looking forward to whenever that 'improvement' lands in the Sailfish browser.
 

The Following 4 Users Say Thank You to nthn For This Useful Post:
Dousan's Avatar
Posts: 1,129 | Thanked: 1,567 times | Joined on Jan 2010 @ Denmark
#23
This has turned all technical o.O

Regardless of poor webpage development, loads of pics or 3. party scripts and whatnot, the sailfish browser is performing poorly. I have no such issues with safari on my iPhone 6.

But the point was it didn’t work for me and my usage of the web.
__________________
We do not see things as they are; we see things as we are...

My Procreate art:
https://procreate.art/dousan

My S°ciety6:
http://society6.com/Dousan
 

The Following 2 Users Say Thank You to Dousan For This Useful Post:
pichlo's Avatar
Posts: 5,326 | Thanked: 16,238 times | Joined on Sep 2012 @ UK
#24
In addition to nthn's summary (I would also like to know WTF the numbers represent ), here is one piece of evidence that I care about:

All the imgur links...
  • On my Jolla 1 - resolve nicely to a picture of some sort with a lot of numbers
  • On my Android tablet - resolve to a black rectangular placeholder, nothing more

All the basemark links...
  • Jola 1 - a page with some text, numbers, buttons
  • Android tablet - page fails to load ("Couldn't establish secure connection ")

In all cases, the default browser was used. It looks like a 2:0 win for the Sailfish browser to me

I would have attached a screenshot to demonstrate if only I knew how to take a screenshot on the damn Android tablet.

EDIT
OK, screenshots added. Had to take them with a camera, the only one I had available (which just happened to be Jolla). The bright dots are reflections of the ceiling lights.
Attached Images
  
__________________
In particle accelerators atoms are indeed not only touching each others. But banging together in a massive explosive orgasm.
-- nieldk in a TMO post

Last edited by pichlo; 2018-04-16 at 12:52. Reason: Added screenshots
 

The Following User Says Thank You to pichlo For This Useful Post:
Posts: 22 | Thanked: 82 times | Joined on Jun 2017
#25
8 year old tablet is not a fair competition, is it? I can dig out my old Windows 98 machine and try loading some websites on Internet Explorer 6. This wouldn't tell me anything else than - well, IE6 sucks.

JetStream benchmarks evaluates how well/quickly the browser performs doing certain tasks, like crypto, loading external libraries, 3D rendering, physics, basic algorithms like Fibonacci numbers. More is better. In-depth analysis of each step, including source codes for the programs that were ran during benchmark here: https://browserbench.org/JetStream/in-depth.html
I know it's a bit specific and doesn't necessarily reflect how you'd use it in real life but it serves as a nice indication of how well JS engine of given browser performs doing various tasks.

That's why I also ran Web 3.0 Basemark, which evaluates WebGL, Canvas, SVG performance, as well as HTML5 support. Again, higher is better.

if you scroll too quickly, the page just goes blank for a bit until that piece is loaded. Not looking forward to whenever that 'improvement' lands in the Sailfish browser.
Too bad, it's already there, the difference is that if you then rapidly scroll up, SFOS browser shows some blurry mess from cache. I have no such issue on my PC though, but then I use one of these webkit browsers :/

Web has changed since 1980s, guys. You might not like these shiny, constantly refreshing, responsive websites, but they're here. According to StatCounter data this release of Firefox is in use by ~3% Firefox users, you can't expect web developers to jump through hoops just to support it, you know. Not being supported by an ancient browser doesn't mean that website is not coded according to standards, it just means that the browser is not supporting them.
 

The Following User Says Thank You to pisarz1958 For This Useful Post:
pichlo's Avatar
Posts: 5,326 | Thanked: 16,238 times | Joined on Sep 2012 @ UK
#26
Originally Posted by pisarz1958 View Post
8 year old tablet is not a fair competition, is it?
Of course not! (I was wrong, BTW, it's only 7 years old.)

Neither is comparing Jolla 1 (or, worse, Aqua Fish) with iPhone 6. Not only is the latter one year more recent, but, crucially, it was the top of the range at the time compared to Jolla which was a mediocre spec even at the time of release.

Which was exactly my point.

Anyway, to each his own. If Sailfish does not cut it for you, fair enough, do not use it.
__________________
In particle accelerators atoms are indeed not only touching each others. But banging together in a massive explosive orgasm.
-- nieldk in a TMO post
 

The Following 2 Users Say Thank You to pichlo For This Useful Post:
nthn's Avatar
Posts: 621 | Thanked: 2,263 times | Joined on Jun 2014
#27
Originally Posted by pisarz1958 View Post
Too bad, it's already there, the difference is that if you then rapidly scroll up, SFOS browser shows some blurry mess from cache. I have no such issue on my PC though, but then I use one of these webkit browsers :/

Web has changed since 1980s, guys. You might not like these shiny, constantly refreshing, responsive websites, but they're here. According to StatCounter data this release of Firefox is in use by ~3% Firefox users, you can't expect web developers to jump through hoops just to support it, you know. Not being supported by an ancient browser doesn't mean that website is not coded according to standards, it just means that the browser is not supporting them.
You're taking the same position of the apologist that's always used in these discussions, which presents both the present and the future as inevitable.

- Standards? Who cares! We know better than to do things in a way that's guaranteed to work! Stop complaining, this is the future!

- Keeping memory usage as low as possible? Who cares! Memory is cheap now, it's only normal for a plaintext chat application to take up 500MB of RAM! Stop complaining, this is the future!

- Cooperation? Who cares! We know better! Stop complaining, this is the future!

- Privacy? Who cares! Companies/governments already know everything about you by now! Stop complaining, this is the future!

It's always the same: if you have any complaints about current goings-on (and future directions) in the technology department, you must be a crazy senile caveman who doesn't want to follow the path of progress, the path of or towards enlightenment. No! We complain precisely because we see that this is not the path of progress. Change will always be fundamentally necessary, but the changes must also be good!

Either way, you're misrepresenting the issue. Save for the occasional actually new technology (a rare occurrence), web developers who want to support older browsers don't need to 'jump through hoops', they just need to adhere to the standards - then they support all browsers (except Internet Explorer, which is genuinely a browser many web developers still jump through hoops for).
 

The Following 3 Users Say Thank You to nthn For This Useful Post:
Posts: 22 | Thanked: 82 times | Joined on Jun 2017
#28
I don't know how suddenly cool things like CSS Grid, Fetch, WebRTC, Service Workers, Web Assembly or WebGL threaten your privacy nor directly lead to badly managing memory, but ok. It's not like anyone cooperated on these features. I know there are polyfills for some of the things I mentioned, some of which have their own weird bugs (CSS grid polyfill). Guess I should just code websites using tables then, since this is kinda guaranteed to work.

I'd like to remind that I use SFOS Browser as my daily driver and I'm concerned about my privacy and I hate that people make **** apps too (Faaceboook). But what I hate more is how toxic these communities, formed around free software, can get.

Last edited by pisarz1958; 2018-04-16 at 16:51.
 

The Following User Says Thank You to pisarz1958 For This Useful Post:
pichlo's Avatar
Posts: 5,326 | Thanked: 16,238 times | Joined on Sep 2012 @ UK
#29
Originally Posted by pisarz1958 View Post
I don't know how suddenly cool things like CSS Grid, Fetch, WebRTC, Service Workers, Web Assembly or WebGL threaten your privacy nor directly lead to badly managing memory, but ok.
You are misinterpreting again. All nthn's examples were examples of other things that are prevalent nowadays, in addition to bad web design.

All could be summarized with simple, "take as much as you can".
As opposed to, "take as little as you have to".

If you are under 35, then you may consider it normal and not even realize it is going on. You need to allocate a memory block for some operation? Old skool geezers like me would calculate exactly how many bytes are needed and allocate exactly that. The current approach is, allocate at least a megabyte, just in case. Memory is cheap, nobody cares. In my previous job, we ran into a bandwidth problem. My preferred approach would be, profile the system, find bottlenecks, optimize. The (much younger) tech lead opted for allocating more resources. I could give millions of other examples.
__________________
In particle accelerators atoms are indeed not only touching each others. But banging together in a massive explosive orgasm.
-- nieldk in a TMO post
 

The Following 2 Users Say Thank You to pichlo For This Useful Post:
Posts: 22 | Thanked: 82 times | Joined on Jun 2017
#30
What if the newer version of the browser engine had some optimizations and fixes? (probably has) What if new APIs not only let you do things you couldn't before, but saved both your time as a developer, and maybe even valuable resources? You can't just dismiss anything that's new because you just assume nobody cares about writing good code anymore.

I don't think it's about being younger or older, it's about being a good coder or a bad coder.
 

The Following 2 Users Say Thank You to pisarz1958 For This Useful Post:
Reply

Thread Tools

 
Forum Jump


All times are GMT. The time now is 04:36.