Not supporting websites from a fairly big country with a strong economy is a failure on their part, either way. I bet they support TLDs like TV (Tuvalu), or IO (Indian Ocean Territory).
What's the reasoning behind blocking, apparently, all but 5 TLDs? I could see sketchy ones like XYZ being blocked, but the default whitelist is a bit... Bare.
No idea, but my point (and I think the point of OP) was that people shouldn't treat some rando's side project with the same amount of scrutiny as they might a new release of Firefox.
This is obviously some person's side project to test a bunch of ideas. One of which was to add in TLD whitelisting, so they came up with a couple of TLDs to whitelist with the intention of expanding it once it went viral on hackernews and they were bought out for a billion dollars by Walmart's venture capital vehicle.
That'd be a lot of work (clean up the code, look through everything to see whether things can actually be released and not covered under licenses disallowing that) for .. what benefit to Microsoft exactly? Trident and EdgeHTML are not really used anywhere anymore, or would that just be out of curiosity how things look inside?
"As you may have heard Microsoft Edge no longer uses Chakra. Microsoft will continue to provide security updates for Chakracore 1.11 until 9th March 2021 but do not intend to support it after that."
IE implemented DOM without using tree. No, really. Up to 2017. After reading the article, all my questions why IE was so buggy evaporated and replaced by how could IE work at all.
After learning about IE and MSVC, I gained respect for Microsoft developers and disdain for Microsoft. Tightrope walking is amazing display of skill. Letting your employees to commute by tightrope walking for years is amazing display of stupidity.
Excellent article on Microsoft's journey of refactoring IE's DOM model. I actually felt sympathy reading how it all started with a linear text-based data structure, which must have seemed reasonable at the time. (In hindsight, a tree should have been the natural shape, and they should have been aware of parsers and ASTs. But can we blame them?) And how the layers of complexity were added, using parallel data structures that needed to be synced on every DOM operation.
By the time the article reached "Modernizing the DOM tree", I could just imagine the years of devs suffering that the whole architecture must have caused.
> According to an internal investigation, from IE7 to IE11, approximately 28% of all IE reliability bugs originated from code in core DOM components.
The second article is quite relevant too, how the C/C++ compiler used token streams instead of a real AST, until 2015! There are parallels to the above issue with the DOM model, how the original design of the data structure was inadequate, so they had to keep adding more layers of complexity. It's understandable how it happened, but makes me wonder if they should have had more experts in parsers/compilers during the design phase.
To give credit where due, both articles describe major successful refactorings of core design. I'm with you, that after learning about their struggles, I have more respect for developers at Microsoft.
Isn't IE on track to be phased out? I guess that also means the Trident engine component will also likely not be available on future versions of Windows?
I cannot imagine Windows removing the trident engine, they have a strong focus on backward compatibility and it's guaranteed a lot of businesses depend on trident being available.
The recent move to make Edge something you can't uninstall is a likely first step in changing that. You can't remove the old "can't remove" browser engine without first designating a replacement.
Eh.... it used to be true that Microsoft cared deeply about backwards compatibility, but it seems to me that's becoming less and less the case. I wouldn't be very surprised if Microsoft just ripped it out one day and damn the consequences. I guess it's just another thing that has fallen out of fashion there, like testing, a working control panel, and updates that aren't a trainwreck.
Pasting the whole article here since it requires registration.
------
"Microsoft last week began a process that will, eventually if not immediately, result in it canning Internet Explorer (IE), the browser it's built since 1994.
Although IE survived this initial round of deprecations – or depredations, depending on your browser viewpoint – there's no guarantee that the final edition of IE will die a natural death in its bed.
Microsoft could kill it at almost any time.
Sinking fast
IE's share of global browser activity has been shrinking, and sinking, fast. According to U.S.-based analytics company Net Applications, IE's share fell by nearly 40% between June 30, 2019 and June 30, 2020. If it continues that rate of decline, IE will account for just 2.8% of all browsing worldwide a year from now.
The browser's fall has been long, slow and with one exception, regular as clockwork. Computerworld's records of Net Applications' browser share data go back as far as January 2005, when IE controlled almost 90% of all share. Since then, it's been downhill for the browser, with only a recovery that started at the top of 2012 and ended two years later, to interrupt the trend line.
Net Applications' numbers don't count every instance of IE. In fact, because of the way the metrics vendor measures share, it's almost certain that it undercounts enterprise usage, perhaps substantially so. In other words, IE may be much more widely used than the numbers indicate.
Absent numbers from Microsoft itself, however, Net Applications' are as good as any. And those show IE running toward not just obsolesce – it's already there, been there for years – but to extermination.
It's no surprise that Microsoft has started to withdraw support.
The end begins
Last week's support changes were relatively minor, but still telling.
On Nov. 30, the Teams web app will stop support IE11. Then, as of Aug. 17, 2021, all remaining Microsoft (and Office) 365 apps and services will no longer support IE11.
While that was a clear shot across IE11's bow and meant to wean enterprise users – the customers who run Microsoft (and Office) 365) – off the browser, Microsoft made clear that, for now, IE11 would remain on the support list. "We want to be clear that IE11 isn't going away,"
Microsoft said in the unsigned post that announced the support changes. "Our customers' own legacy IE11 apps and investments will continue to work. Customers have made business-critical investments in IE11 legacy apps and we respect that those apps are still functioning."
This promise of IE11's longevity was reminiscent of Microsoft's long-standing position that "Internet Explorer is a component of the Windows operating system and follows the Lifecycle Policy for the product on which it is installed." By that definition, IE11 will be supported on Windows 10 for the same length of time as the operating system itself.
The Microsoft (and Office) 365 support stick has been wielded by Microsoft before. Editions of the perpetual-licensed Office – Office 2010, 2103, 2016 and 2019 – drop connectivity support with Office 365's services starting in October 2020. Microsoft's goal? Shove customers to subscriptions, since Office 365's applications will never lose that support.
The Redmond, Wash. giant is throwing its weight around in a similar fashion as it denounces IE11. This time, though, the goal is to prod enterprises still using IE to abandon the browser, if not immediately then as soon as possible.
(Something important to remember: Microsoft's said that the lapsing support for Teams (November) and the rest of Microsoft 365 (August 2021) applies to Edge's IE m...
You can enable experimental support for TLS 1.3 on Internet Explorer for Windows 10 (not for older Windows) in the advanced options tab. You should also disable third party add-ons and enable enhanced protected mode. If you also disable jscript.dll[1] you will get all the benefits of the IE shell that was posted here without having to download anything.
No, IE11 (including Trident, JScript, etc) will remain supported as long as Windows 10 is supported, as it's a built-in Windows component.
You're probably asking "Isn't IE on track to be phased out" because Microsoft recently announced that the web versions of Office, OneDrive etc will stop supporting IE11 in mid-2021.
IE11 will continue to be supported to make it easier for enterprises to upgrade their intranets over time. The new Chromium-based Edge actually has a deployment mode where you can use Group Policy to load some sites using the IE engine[1]. This requires the IE DLLs to be present on the system.
I agree with you on the latter two points, but for the first one, it sounds like the application is using IE's rendering engine built into all versions of Windows. So the executable is super small given that all the heavy lifting is in libraries already on your system.
I don't know how the author of Nessie specifically did it, but generally on the .NET platform, you have access to the Trident rendering engine by adding the WebBrowser control to one of the forms of your new application. From there, you gain all the event based functions and callable methods you'd expect a browser to use. I've written several of these type of 'toy' browsers over the years, and they work just like IE does. The only caveat is that you have to patch the registry first to allow the control to report (and render as) a useragent above IE 8. You can set it to report/render as IE 11 as the maximum.
Via NuGet, you can download the WebBrowser2 control which is based on the new Edge Chromium based renderer. I haven't seen any new browsers using this properly yet (and I've not played with it either yet) - it will be good when people do, because although the renderer will still be Chromium, the browser UI can then be anything you want. I'm looking forward to people challenging the accepted browser UIs that we currently have.
Of course, if you want to avoid the WebBrowser2 control, there's webkit.NET and Embedded Chromium which also can be built to bind with .NET apps.
In case Jaruzel's too shy to mention it, he wrote a Gopher browser for Windows. If you're reading this thread because you're interested in small web topics (Gopher, Gemini, Tilde.Town, NetSurf, alt browsers) then you might like his browser too. It was my gateway into Gopher.
It's Windows-only because it's using the IE rendering engine made by Microsoft that comes built-in to Windows itself.
The difference with a new lightweight browser like NetSurf is that they actually have their own rendering engine crammed into their 12MB EXE (and it's cross-platform).
That means the slowness comes uniquely from IE, not from the engine? I find that surprising, for some reason. I would have imagined the browser UI would be easier to optimize.
Maxthon switched dynamically from trident to webkit.
The best browser performance wise is the one that switch it's engine on demand depending of a known list of websites being faster on engine X or Y
My needing to rely on one modern browser engine is already too much software complexity for my tastes.
In other words, I would have preferred to be able to live a typical day of my life without relying on any software as complicated (and consequently difficult for me to predict and control) as a modern browser engine.
So, arranging for that typical day to be reliant on two browser engines is something I will strenuously try to avoid.
Indeed. I didn't stop reading when I hit "Trident layout engine". I stopped reading at "Requirements: Windows 7 or later."
Though I should have assumed as much given "trident layout engine", but here I was hoping this had been open sourced or something along those lines.
I absolutely do not want to speak ill of anyone's project. But let's judge it for what it is... it's a windows only, MSHTML based, proprietary software.
At that point, it honestly doesn't matter what it can do, or how it does it. It's already off to the worst possible start for any hope of adoption. And it will likely never see the light of day on either Linux or MacOS. Porting it would tantamount to writing it again from scratch.
That said, if it works for windows users. It's not like it is a bad contribution. So kudos for that.
On a less polite opinion: I honestly think a core component such as a web rendering engine that is locked to a OS (pretty much exclusively Trident/EdgeHTML) does the world way more harm than good. Any project with some kind of future prospects of not being windows-only, does themselves disservice for picking Edge. Valve did this with steam. I don't know how painless the transition was when moving to WebKit and then to Chromium, but I'm sure it wasn't free.
HN is pretty average. Windows is less than half for devs (but is still the largest OS share).
However, as is expected, those chosing to go out of their way to change their machines would naturally be more vocal than those that are indifferent to what it runs.
Well, someone in this comment tree linked stats, but I think those stats and the stats most people that visit this site come up with via observation understate how many boring jobs there are in e.g. fortune 5000 companies doing Windows-centric work.
There will always be bias in stats, but the numbers are from SO surveys - I don't think SO is going to have "boring" Windows-centric development underrepresented. I'd assume the contrary.
Yes please, have not used windows for 15+ years here, I heard it is embracing linux nowadays, a typical microsoft move, the traditional trojan horse approach.
277KB is still a bit large for something that just embeds the IE webview... it reminds me of this one (original site is now gone, had to look in the archive) which is only 2.7KB(!):
>Its codebase is very small (about 2500 LoC currently) and could be open sourced in the next few months if the project gains enough traction and support.
Yeah, at this point I view “install closed source software from a not-well-known source” as a very high risk choice. I was thinking of adding this to my set of tertiary browsers, until I couldn’t find the license/source. I wish them well but closed source is a no-go for me.
There's a new term for this type of software distribution. Sponsorware. Wherein, the maintainer only open sources the project if they get enough people to pay them a monthly sponsor fee.
I think I would actually take dillo with an updated chrome over this: if I am to use two web browsers, I don't necessarily need javascript on both.
dillo lacks a good UI for bookmarks, a browsing history (by choice, but I like those), touch support (for postmarketos/mobile), wayland support, as well as some https quirks. Otherwise, I like it a lot.
92 comments
[ 3.6 ms ] story [ 175 ms ] thread.com .net .org .co .io
Just add a TLD, restart and voila.
Wtf?
AFAIK, IE/Edge was developed only by Microsoft, so there wouldn't be much if any thirdparty code?
- https://devblogs.microsoft.com/oldnewthing/20180515-00/?p=98...
I thought edge was reskinned Chrome, why don't they use V8?
"As you may have heard Microsoft Edge no longer uses Chakra. Microsoft will continue to provide security updates for Chakracore 1.11 until 9th March 2021 but do not intend to support it after that."
The new Edge uses Blink for both.
IE implemented DOM without using tree. No, really. Up to 2017. After reading the article, all my questions why IE was so buggy evaporated and replaced by how could IE work at all.
MSVC up to 2015 also didn't create AST. Unbelievable, I know. The question is not why MSVC's C++ support was so buggy; the real question is how could MSVC implement C++ at all. https://devblogs.microsoft.com/cppblog/rejuvenating-the-micr...
After learning about IE and MSVC, I gained respect for Microsoft developers and disdain for Microsoft. Tightrope walking is amazing display of skill. Letting your employees to commute by tightrope walking for years is amazing display of stupidity.
By the time the article reached "Modernizing the DOM tree", I could just imagine the years of devs suffering that the whole architecture must have caused.
> According to an internal investigation, from IE7 to IE11, approximately 28% of all IE reliability bugs originated from code in core DOM components.
The second article is quite relevant too, how the C/C++ compiler used token streams instead of a real AST, until 2015! There are parallels to the above issue with the DOM model, how the original design of the data structure was inadequate, so they had to keep adding more layers of complexity. It's understandable how it happened, but makes me wonder if they should have had more experts in parsers/compilers during the design phase.
To give credit where due, both articles describe major successful refactorings of core design. I'm with you, that after learning about their struggles, I have more respect for developers at Microsoft.
I guess it's not wrong.
https://www.virustotal.com/gui/file/ec9ad4a4d6d1ddcb406d96ff...
https://www.computerworld.com/article/3572554/is-it-eol-for-...
Pasting the whole article here since it requires registration.
------
"Microsoft last week began a process that will, eventually if not immediately, result in it canning Internet Explorer (IE), the browser it's built since 1994.
Although IE survived this initial round of deprecations – or depredations, depending on your browser viewpoint – there's no guarantee that the final edition of IE will die a natural death in its bed.
Microsoft could kill it at almost any time.
Sinking fast IE's share of global browser activity has been shrinking, and sinking, fast. According to U.S.-based analytics company Net Applications, IE's share fell by nearly 40% between June 30, 2019 and June 30, 2020. If it continues that rate of decline, IE will account for just 2.8% of all browsing worldwide a year from now.
The browser's fall has been long, slow and with one exception, regular as clockwork. Computerworld's records of Net Applications' browser share data go back as far as January 2005, when IE controlled almost 90% of all share. Since then, it's been downhill for the browser, with only a recovery that started at the top of 2012 and ended two years later, to interrupt the trend line.
Net Applications' numbers don't count every instance of IE. In fact, because of the way the metrics vendor measures share, it's almost certain that it undercounts enterprise usage, perhaps substantially so. In other words, IE may be much more widely used than the numbers indicate.
Absent numbers from Microsoft itself, however, Net Applications' are as good as any. And those show IE running toward not just obsolesce – it's already there, been there for years – but to extermination.
It's no surprise that Microsoft has started to withdraw support.
The end begins Last week's support changes were relatively minor, but still telling.
On Nov. 30, the Teams web app will stop support IE11. Then, as of Aug. 17, 2021, all remaining Microsoft (and Office) 365 apps and services will no longer support IE11.
While that was a clear shot across IE11's bow and meant to wean enterprise users – the customers who run Microsoft (and Office) 365) – off the browser, Microsoft made clear that, for now, IE11 would remain on the support list. "We want to be clear that IE11 isn't going away,"
Microsoft said in the unsigned post that announced the support changes. "Our customers' own legacy IE11 apps and investments will continue to work. Customers have made business-critical investments in IE11 legacy apps and we respect that those apps are still functioning."
This promise of IE11's longevity was reminiscent of Microsoft's long-standing position that "Internet Explorer is a component of the Windows operating system and follows the Lifecycle Policy for the product on which it is installed." By that definition, IE11 will be supported on Windows 10 for the same length of time as the operating system itself.
The Microsoft (and Office) 365 support stick has been wielded by Microsoft before. Editions of the perpetual-licensed Office – Office 2010, 2103, 2016 and 2019 – drop connectivity support with Office 365's services starting in October 2020. Microsoft's goal? Shove customers to subscriptions, since Office 365's applications will never lose that support.
The Redmond, Wash. giant is throwing its weight around in a similar fashion as it denounces IE11. This time, though, the goal is to prod enterprises still using IE to abandon the browser, if not immediately then as soon as possible.
(Something important to remember: Microsoft's said that the lapsing support for Teams (November) and the rest of Microsoft 365 (August 2021) applies to Edge's IE m...
These browsers fail connect when TLS 1.3 is mandatory, warning the user that "cannot connect with the site that might be using insecure protocols"
[1] https://borncity.com/win/2020/01/18/warning-0-day-vulnerabil...
You're probably asking "Isn't IE on track to be phased out" because Microsoft recently announced that the web versions of Office, OneDrive etc will stop supporting IE11 in mid-2021.
IE11 will continue to be supported to make it easier for enterprises to upgrade their intranets over time. The new Chromium-based Edge actually has a deployment mode where you can use Group Policy to load some sites using the IE engine[1]. This requires the IE DLLs to be present on the system.
[1] https://docs.microsoft.com/en-us/deployedge/edge-ie-mode
Via NuGet, you can download the WebBrowser2 control which is based on the new Edge Chromium based renderer. I haven't seen any new browsers using this properly yet (and I've not played with it either yet) - it will be good when people do, because although the renderer will still be Chromium, the browser UI can then be anything you want. I'm looking forward to people challenging the accepted browser UIs that we currently have.
Of course, if you want to avoid the WebBrowser2 control, there's webkit.NET and Embedded Chromium which also can be built to bind with .NET apps.
In case Jaruzel's too shy to mention it, he wrote a Gopher browser for Windows. If you're reading this thread because you're interested in small web topics (Gopher, Gemini, Tilde.Town, NetSurf, alt browsers) then you might like his browser too. It was my gateway into Gopher.
http://www.jaruzel.com/gopher/gopher-client-browser-for-wind...
The difference with a new lightweight browser like NetSurf is that they actually have their own rendering engine crammed into their 12MB EXE (and it's cross-platform).
[Alt]+ [<-] or [->] works for going Back / Fwd
Interestingly Alt + N for new window and then Ctrl+Tab cycles the windows.
I like this old school workflow for some browsing tasks rather than tabs (but not all the time obviously)
The Edge browser was pretty decent, awful tracking aside. Arguably better than Chrome in many aspects.
In other words, I would have preferred to be able to live a typical day of my life without relying on any software as complicated (and consequently difficult for me to predict and control) as a modern browser engine.
So, arranging for that typical day to be reliant on two browser engines is something I will strenuously try to avoid.
Though I should have assumed as much given "trident layout engine", but here I was hoping this had been open sourced or something along those lines.
I absolutely do not want to speak ill of anyone's project. But let's judge it for what it is... it's a windows only, MSHTML based, proprietary software.
At that point, it honestly doesn't matter what it can do, or how it does it. It's already off to the worst possible start for any hope of adoption. And it will likely never see the light of day on either Linux or MacOS. Porting it would tantamount to writing it again from scratch.
That said, if it works for windows users. It's not like it is a bad contribution. So kudos for that.
On a less polite opinion: I honestly think a core component such as a web rendering engine that is locked to a OS (pretty much exclusively Trident/EdgeHTML) does the world way more harm than good. Any project with some kind of future prospects of not being windows-only, does themselves disservice for picking Edge. Valve did this with steam. I don't know how painless the transition was when moving to WebKit and then to Chromium, but I'm sure it wasn't free.
> ● Privacy-oriented: zero behavior tracking, zero telemetry > ● Lightning-fast startup time > ● Tiny binary size > ● Low RAM usage
why would I use an OS that’s the opposite of all of these? In particular, why would I trust a proprietary application to be “privacy-oriented”?
However, as is expected, those chosing to go out of their way to change their machines would naturally be more vocal than those that are indifferent to what it runs.
Maybe on SF.
http://web.archive.org/web/20081226192818/http://www.kakeewa...
Why not open source it from the start?
I'd be more impressed if the webview was introduced by not using a OS component.
Fix't it for ya.
dillo lacks a good UI for bookmarks, a browsing history (by choice, but I like those), touch support (for postmarketos/mobile), wayland support, as well as some https quirks. Otherwise, I like it a lot.