7 ms·

I am glad they are looking into this: Sorry for the following rant, but I really don't think of Firefox browser when I think of code quality. (Still don't even

I am glad they are looking into this: Sorry for the following rant, but I really don't think of Firefox browser when I think of code quality. (Still don't even though it has improved a lot in some areas). When it comes to browsers, I still feel the old Opera browser (presto), that had more features, was blazing fast and small in size (in memory) is a very good example of a finely engineered software. In the early versions they really cared about optimizing for performance, using less memory and less power (on mobile devices) - I am amazed that years after, modern browsers are bloatier now than ever and just don't seem to care as much about these aspect as much as Opera did. Questions like how modular is Firefox, how easy is it to customize Firefox (add or remove feature from the code), how easy is it for someone to use the Gecko rendering engine in their own project etc. all kind of indicate the bloatier, messier nature of the code within Firefox (in my opinion). It's like browser developers now a days only care about adding more features (read bloat) without considering any constraints ...

gsnedders 6y ago
Presto wouldn't be so memory efficient on today's web; web applications typically ran into plenty of places where memory consumption was worse in Presto versus WebKit, for example. (A simple example here is that each DOM Node was larger in Presto than any other browser, if I'm not forgetting.) But ultimately a lot of this comes down to different considerations: when the majority of your revenue comes from companies who care about running on limited devices, of _course_ the company invested more in running well on them. And I wouldn't hold up Presto as a great example of modularity or the ability to embed the rendering engine in other projects.
- webmobdev 6y ago
  > Presto wouldn't be so memory efficient on today's web; I disagree. (I continued to use the last version of Opera Presto for a few year even after it was abandoned, and it continued to outperform other browsers on both platforms - Windows and macOS, though the macOS version was a bit more buggy). Let's not forget that Presto already supported a lot of HTML5 features(1) before it was abandoned. (Even today, it works surprisingly well on most websites, though it does show its age a bit). And while the last few versions did seem to be a bit more resource heavy, I attribute that to development suffering because of the upheaval in the company - morale was low as many developers were laid-off while the management was also negotiating a sell out with investors. This naturally stunted the development of the core browsers parts in the end. (Also, it was a suite and had an integrated email client, IRC, Torrent download manager, RSS reader, apart from the browser, all of which was packaged in a smaller size than Firefox or any other browser of its time!) > when the majority of your revenue comes from companies who care about running on limited devices, of _course_ the company invested more in running well on them. True. But let's not forget that Opera browser earned its reputation for being blazing fast, its tiny size and its low memory consumption on the desktop platforms long before it became a popular browser on the mobile platform. If I remember right, the first few versions of the Opera browser (then a non-free product) was less than 3-4 MB(2)! > And I wouldn't hold up Presto as a great example of modularity or the ability to embed the rendering engine in other projects. I can't comment much about this - The source code of Opera Presto is out there on the net, and it is definitely much easier to understand than the Firefox codebase. I also remember that the popular Macromedia (now Adobe) Dreamweaver used the presto engine at one point. It says a lot about Mozilla that despite being older than both Chromium / Blink and Webkit, the latter are more popular in many open source and commercial applications because of the ease of use in coding and integrating it in their applications. A good example of this is QT5 which offers the QTWebView and QtWebEngine module (3) that allows you to embed webkit or chromium in your application, out of the box. Whereas qtmozembed is a barely maintained unpopular option for this same use (probably only surviving because Jolla develops their mobile browser for Sailfish OS using it). (And I suspect this was a deliberate decision made by Mozilla after Gecko / Firefox became a cash cow for them - they just didn't want the code to be modular and easy to understand as it would hinder the creation of competing browsers, which would otherwise potentially cannibalise their own product. Ofcourse, with GeckoView this slowly seems to be changing perhaps on the Android platform, but I cynically suspect this has more to do with the fact that Firefox Mobile is not very popular on the mobile platform, and Mozilla now is looking to encourage developers to create even alternate browsers using their technology hoping it'll provide a filip for their own Firefox Mobile browser. That's certainly going to be interesting to see ... Another intresting tidbit I recall reading somewhere is that Mozilla also makes some developers sign non-disclosure agreements (4) which seems very odd and out of place to me, for an open source project.) (1): https://en.wikipedia.org/wiki/Presto_(browser_engine) https://en.wikipedia.org/wiki/Presto_(browser_engine) (2): https://filehippo.com/download_opera/8.02/ https://filehippo.com/download_opera/8.02/ (3): https://stackoverflow.com/questions/29055475/qwebview-or-qwebengineview https://stackoverflow.com/questions/29055475/qwebview-or-qwe... (4): https://medium.com/@smsnobin77/firefox-build-macos-f1f53d643e81 https://medium.com/@smsnobin77/firefox-build-macos-f1f53d643...
  - GeorgeWBasic 6y ago
    > The source code of Opera Presto is out there on the net I don't suppose it's been open-sourced, has it? It would be great to have an alternative browser engine out there, instead of just three (or two and a half, depending on how you count Blink and Webkit).
    - webmobdev 6y ago
      > I don't suppose it's been open-sourced, has it? No, it was leaked illegally. Opera filed a lot of DMCA complaints with many online Git repository services to get the code removed from them. But you can still find it if you Google hard. I also recall reading that some Russian developers were working on the leaked code and had also released a "newer" version with some patches (Opera presto was really popular in Russia) but I couldn't find it. There is another little known open source browser project independent of Gecko, webkit and Blink currently being developed but I just can't remember its name, even though I tried compiling it on my system ... damn ... I'll post an update when its name comes to mind.
      - Arelius 6y ago
        Servo maybe?
        webmobdev 6y ago
        No, it was NetSurf - https://www.netsurf-browser.org/ https://www.netsurf-browser.org/ ...
    - lsh 6y ago
      No, but there is an effort to recreate it with the Otter Browser: https://otter-browser.org/ https://otter-browser.org/
      - Ndymium 6y ago
        Otter browser uses WebKit, so it's not a new engine.
acdha 6y ago
Opera felt small and fast because it did less and ran websites doing less. If you look at an old site in a modern browser, it's much faster; if you even could run a modern site in an old browser, you'd see just how many optimizations you're taking for granted — not to mention things like high-quality text rendering which fully supports Unicode and complex scripts.
- webmobdev 6y ago
  > Opera felt small and fast because it did less and ran websites doing less. Let us assume that is true (even though it isn't) - what's Firefox and other browsers excuse for running slower than Opera on these same sites, during the same period? Opera presto supported a lot of the features of HTML5(1). So it didn't necessarily "do less" than the others. And second, if you really want to compare Opera presto with other browsers, you should do it with the versions of the competing browsers that were available then. (I still remember how people complained about Firefox consuming a a lot of RAM when you opened more than 10 tabs, where as Opera could easily handle 100+ tabs(2) without any complaints.) Even otherwise, if you test it out yourself today by downloading an Opera presto version (version 12.15 / 12.16) you will find how usable it still is on 70-80% of the websites today, even after not being developed for years now). (1) https://en.wikipedia.org/wiki/Presto_(browser_engine) https://en.wikipedia.org/wiki/Presto_(browser_engine) (2) https://forums.opera.com/topic/19313/how-to-retrieve-accidentally-closed-100-tabs-with-stupid-close-all-tabs-option https://forums.opera.com/topic/19313/how-to-retrieve-acciden... ... and ... https://portableapps.com/node/59581 https://portableapps.com/node/59581
  - acdha 6y ago
    Nobody is arguing that Opera wasn’t fast, only that there’s been more than a decade of massive improvements which is not reflected in your comments. I used, and paid for, Opera but its golden age is past.
    - webmobdev 6y ago
      > ... only that there’s been more than a decade of massive improvements which is not reflected in your comments ... Oh ok. I see the confusion - I was speaking specifically about Opera (Presto engine) versions. Not the current Opera Chromium clone. And the point I was trying to make is that it was a much, much better product than its competitors. And if it had survived, with the original team still working on it, it still would be better than Firefox / Gecko and Chrome / Blink of today.
Causality1 6y ago
I second your opinion. I used to give Firefox a tremendous amount of leeway because its support for extensions and customization set it head and shoulders above competing browsers. Now that they've thrown away that advantage I find myself using alternatives more and more.

gsnedders 6y ago

Presto wouldn't be so memory efficient on today's web; web applications typically ran into plenty of places where memory consumption was worse in Presto versus WebKit, for example. (A simple example here is that each DOM Node was larger in Presto than any other browser, if I'm not forgetting.) But ultimately a lot of this comes down to different considerations: when the majority of your revenue comes from companies who care about running on limited devices, of _course_ the company invested more in running well on them. And I wouldn't hold up Presto as a great example of modularity or the ability to embed the rendering engine in other projects.

webmobdev 6y ago
> Presto wouldn't be so memory efficient on today's web; I disagree. (I continued to use the last version of Opera Presto for a few year even after it was abandoned, and it continued to outperform other browsers on both platforms - Windows and macOS, though the macOS version was a bit more buggy). Let's not forget that Presto already supported a lot of HTML5 features(1) before it was abandoned. (Even today, it works surprisingly well on most websites, though it does show its age a bit). And while the last few versions did seem to be a bit more resource heavy, I attribute that to development suffering because of the upheaval in the company - morale was low as many developers were laid-off while the management was also negotiating a sell out with investors. This naturally stunted the development of the core browsers parts in the end. (Also, it was a suite and had an integrated email client, IRC, Torrent download manager, RSS reader, apart from the browser, all of which was packaged in a smaller size than Firefox or any other browser of its time!) > when the majority of your revenue comes from companies who care about running on limited devices, of _course_ the company invested more in running well on them. True. But let's not forget that Opera browser earned its reputation for being blazing fast, its tiny size and its low memory consumption on the desktop platforms long before it became a popular browser on the mobile platform. If I remember right, the first few versions of the Opera browser (then a non-free product) was less than 3-4 MB(2)! > And I wouldn't hold up Presto as a great example of modularity or the ability to embed the rendering engine in other projects. I can't comment much about this - The source code of Opera Presto is out there on the net, and it is definitely much easier to understand than the Firefox codebase. I also remember that the popular Macromedia (now Adobe) Dreamweaver used the presto engine at one point. It says a lot about Mozilla that despite being older than both Chromium / Blink and Webkit, the latter are more popular in many open source and commercial applications because of the ease of use in coding and integrating it in their applications. A good example of this is QT5 which offers the QTWebView and QtWebEngine module (3) that allows you to embed webkit or chromium in your application, out of the box. Whereas qtmozembed is a barely maintained unpopular option for this same use (probably only surviving because Jolla develops their mobile browser for Sailfish OS using it). (And I suspect this was a deliberate decision made by Mozilla after Gecko / Firefox became a cash cow for them - they just didn't want the code to be modular and easy to understand as it would hinder the creation of competing browsers, which would otherwise potentially cannibalise their own product. Ofcourse, with GeckoView this slowly seems to be changing perhaps on the Android platform, but I cynically suspect this has more to do with the fact that Firefox Mobile is not very popular on the mobile platform, and Mozilla now is looking to encourage developers to create even alternate browsers using their technology hoping it'll provide a filip for their own Firefox Mobile browser. That's certainly going to be interesting to see ... Another intresting tidbit I recall reading somewhere is that Mozilla also makes some developers sign non-disclosure agreements (4) which seems very odd and out of place to me, for an open source project.) (1): https://en.wikipedia.org/wiki/Presto_(browser_engine) https://en.wikipedia.org/wiki/Presto_(browser_engine) (2): https://filehippo.com/download_opera/8.02/ https://filehippo.com/download_opera/8.02/ (3): https://stackoverflow.com/questions/29055475/qwebview-or-qwebengineview https://stackoverflow.com/questions/29055475/qwebview-or-qwe... (4): https://medium.com/@smsnobin77/firefox-build-macos-f1f53d643e81 https://medium.com/@smsnobin77/firefox-build-macos-f1f53d643...
- GeorgeWBasic 6y ago
  > The source code of Opera Presto is out there on the net I don't suppose it's been open-sourced, has it? It would be great to have an alternative browser engine out there, instead of just three (or two and a half, depending on how you count Blink and Webkit).
  - webmobdev 6y ago
    > I don't suppose it's been open-sourced, has it? No, it was leaked illegally. Opera filed a lot of DMCA complaints with many online Git repository services to get the code removed from them. But you can still find it if you Google hard. I also recall reading that some Russian developers were working on the leaked code and had also released a "newer" version with some patches (Opera presto was really popular in Russia) but I couldn't find it. There is another little known open source browser project independent of Gecko, webkit and Blink currently being developed but I just can't remember its name, even though I tried compiling it on my system ... damn ... I'll post an update when its name comes to mind.
    - Arelius 6y ago
      Servo maybe?
      - webmobdev 6y ago
        No, it was NetSurf - https://www.netsurf-browser.org/ https://www.netsurf-browser.org/ ...
  - lsh 6y ago
    No, but there is an effort to recreate it with the Otter Browser: https://otter-browser.org/ https://otter-browser.org/
    - Ndymium 6y ago
      Otter browser uses WebKit, so it's not a new engine.

acdha 6y ago

Opera felt small and fast because it did less and ran websites doing less. If you look at an old site in a modern browser, it's much faster; if you even could run a modern site in an old browser, you'd see just how many optimizations you're taking for granted — not to mention things like high-quality text rendering which fully supports Unicode and complex scripts.

webmobdev 6y ago
> Opera felt small and fast because it did less and ran websites doing less. Let us assume that is true (even though it isn't) - what's Firefox and other browsers excuse for running slower than Opera on these same sites, during the same period? Opera presto supported a lot of the features of HTML5(1). So it didn't necessarily "do less" than the others. And second, if you really want to compare Opera presto with other browsers, you should do it with the versions of the competing browsers that were available then. (I still remember how people complained about Firefox consuming a a lot of RAM when you opened more than 10 tabs, where as Opera could easily handle 100+ tabs(2) without any complaints.) Even otherwise, if you test it out yourself today by downloading an Opera presto version (version 12.15 / 12.16) you will find how usable it still is on 70-80% of the websites today, even after not being developed for years now). (1) https://en.wikipedia.org/wiki/Presto_(browser_engine) https://en.wikipedia.org/wiki/Presto_(browser_engine) (2) https://forums.opera.com/topic/19313/how-to-retrieve-accidentally-closed-100-tabs-with-stupid-close-all-tabs-option https://forums.opera.com/topic/19313/how-to-retrieve-acciden... ... and ... https://portableapps.com/node/59581 https://portableapps.com/node/59581
- acdha 6y ago
  Nobody is arguing that Opera wasn’t fast, only that there’s been more than a decade of massive improvements which is not reflected in your comments. I used, and paid for, Opera but its golden age is past.
  - webmobdev 6y ago
    > ... only that there’s been more than a decade of massive improvements which is not reflected in your comments ... Oh ok. I see the confusion - I was speaking specifically about Opera (Presto engine) versions. Not the current Opera Chromium clone. And the point I was trying to make is that it was a much, much better product than its competitors. And if it had survived, with the original team still working on it, it still would be better than Firefox / Gecko and Chrome / Blink of today.

Causality1 6y ago

I second your opinion. I used to give Firefox a tremendous amount of leeway because its support for extensions and customization set it head and shoulders above competing browsers. Now that they've thrown away that advantage I find myself using alternatives more and more.