It makes sense; you want to make your app so that even the tiny spinning circles aren't even visible. I added a feature to a webapp in a previous project that would only start showing a spinner if loading a page or section took >1 second. If it's any slower you'd see a flashing spinner, which is not very nice.
The Netscape loading logo is the one I feel very nostalgic about, if we can have animated gravestones when I die, that is the one I want to have on mine https://imgur.com/a/G6quKu8
Connections were much slower back then. Sometimes you'd go a few seconds or more without anything happening on the screen. The loading animation was nice so you knew your browser wasn't frozen.
5.5 was the last version that MS actually cared about IIRC. 6 had a few meaningful additions and tweaks but it had become noticeable that MS wasn't prioritizing development. And of course after 6 MS didn't care much at all until competition came along...
No Mac versions, pocket versions, console versions yet. I loved the Mac version back in the days before Safari, it was actually a pretty fantastic browser for the time.
For some reason, that domain name arouses my suspicion... if I saw that in the search results I'd probably think "SEO spam site" and skip over it.
One of the obvious things you can see between the versions is the gradual trend of dumbing-down and flattening the UI, most apparent between 8 and 9. The older versions' UI give the impression that they are very customisable and powerful, while the newer ones seem to be implicitly saying "you have no control (anymore)." I'd love to use a browser with a newer rendering engine, but retain the old UI.
Their sizes seem to be off (or including something extra). But there definitely is a huge change in IE5 requirements. I'm guessing that's related to the US v Microsoft antitrust bundling lawsuit.
I wonder if IE 6 could lean more on existing libraries (renderers etc) in the OS itself; maybe the IE 6 version for older systems than Windows XP were comparable in size to IE 5?
That seems likely. What's funny is how Edge (and eventually IE) became even more entrenched in the system than before, yet nobody cares anymore. Almost like the whole thing was a bunch of arm-waving by nerds and regulators that wanted to pretend they knew what was going on.
One of the attractions of Chrome when it first came out was that it had less chrome, that most of the precious screen real estate could be devoted to showing the content of the web page, or at least to ads. I'd guess IE9 was a reaction to that.
When we only had 640 vertical pixels, it was a huge change not to have 150 of them taken up with rows of icons and bookmarks by default. Now we have the space to get those controls back, but we're used to not having them.
I had roughly the same amount of rows of pixels on my desktop computer in 2008 as I have now (1280x1024 to 1920x1080). I had even more for a few years prior to that, when I had a CRT monitor that I'd gotten for free (1600x1200).
It's AFAIK pretty much only with 4k monitors that we have caught up with CRTs resolution-wise, and not a lot of people have either of those. w3schools lists a bunch of resoltion statistics. In 2008, the most common resolution was 1024x768. 1366x768 in 2018. Same vertical resolution. Since I thought those statistics may account for phone users as well, I looked further. Globalstats lists 1366x768 as the most common desktop resolution in 2019. Desktop in this case probably accounts for laptops as well, as I believe there is no way to tell the two apart with User-Agent.
1366x768 has been the default resolution for cheap laptops for over a decade now. It's only very slowly being replaced with 1920x1080, but even when this transition is complete, effective information density is unlikely to change much. There are only so many pixels you can squeeze onto a 13" screen before most users need to scale up the UI.
For someone with bad eyesight like me, the sweet spot for a 13" laptop would be 4K at somewhere between 250% and 300% scaling. That would give me just about the same effective information density as good ol' 1366x768. I guess there's a reason why this awkward resolution still refuses to die.
Seeing the size of the chrome on the old IE versions, particularly version 3 where they upgraded to the large buttons with help text, made me think about how nice it was to have the relatively plain content on the web at that time. The information density was high enough that the small space for the actual document was still enough for a decent experience. Even with those low screen resolutions, the amount of content on the screen was probably similar to that of many websites today with all their bloat.
> the newer ones seem to be implicitly saying "you have no control (anymore)."
Since all those toolbars can still be made visible in newer versions, I think MS's choice to hide them is more about a minimalist aesthetic than actually removing control.
> One of the obvious things you can see between the versions is the gradual trend of dumbing-down and flattening the UI,
This has been my experience for the majority of software, from native to web and mobile to desktop. Less options, less tools, less customization.
I think it's partly because software has become more mainstream over time, and instead of trying to teach people to use more complex things, developers have chosen or been forced into appealing to the lowest common denominator. Make it as simple as possible so people who aren't interested in learning how to use the tools can still get by.
I don't necessarily think this is a problem on its own, but combined with the business needs enforced by people looking to maximize profits it becomes less palatable. If we've built something that works for everyone, it's a hard sell to create more powerful options for the 10% (pulled that number out of my ass, I have no idea what the actual amount would be). Then we're all stuck with worse tools.
I disagree -- the difference between IE8 and IE9 is that IE8 expects the user to interact with the browser's interface, while IE9 attempts to get the browser out of the way of the user and instead puts the browser's content front-and-center.
Personally, I like this thought-line a little better. Users ultimately want to interact with the application only with regards to their content, not the applications that house that content.
Difference being it's served as 200, while same command for example.com with no host gives 404.
I think the web is just broken without host headers, as too much is running on the same ips / through routers before touching the real underlying page.
Host headers are explicitly required by the spec, and there are no clients left that leave them out (including non-browser clients like curl and wget). There is simply no reason to facilitate requests that have no Host header, so servers receiving them quite rightly serve some kind of error response.
Yeah they should be; both Microsoft and Intel have a big reputation when it comes to backwards compatibility in their operating systems and instruction sets. It's only been fairly recently that they're starting to break with that, notably with Edge. Not sure about Windows yet, but, wouldn't be surprised if eventually they just put an older windows VM in there - if they haven't already.
Nah, you can't run 16-bit executables or NE executables in modern 64-bit Windows any more. The virtual 8086 mode that was being used is basically unavailable in long mode.
I felt a bit of nostalgia too, but I think mainly for Windows XP UI. Maybe it wasn't very polished but it looked pretty funny and friendly; it made you feel like at home -- especially Windows XP Home Edition (ok ok I'm sorry...)
Yes. September was the month that many new users discovered the Internet for the first time, because they were given access at University and other higher education establishments, which had many and varied implications.
For instance there was an influx of beginner questions in Usenet groups, questions to which the answer was "please read the FAQ before asking more questions".
There was also just a general influx of activity: groups would see a bump in user numbers and activity that didn't fall back off over time.
New users arrived all through the year of course, but in smaller numbers and tended to be more technically proficient (and aware with network etiquette) to start with because they had sought (and paid for) access for a reason, rather than being given it and being told to go explore.
From 1993 onwards the public more generally were becoming a lot more aware of the Internet and being told about the things it could potentially do for them, so there was a more constant flow of new (often relatively non-technical) users & activity rather than just a big glut each September, which is why that year was referred to as "the September that never ended".
Same. I remember arguing with a friend in late 2000 who still insisted on using Netscape 4. I had given up on Netscape years before and stuck with IE from around 3 through 5.5. I actually don't remember using 6 much on purpose. I know I switched to Phoenix at v0.4 and never touched IE again, which was apparently October 2002; IE6 had been at the end of 2001. It's possible I was using IE6 for that yearish, though I didn't start using XP for awhile, preferring to stick with Windows 2000.
In 2012 I worked with a Windows 2000 Server machine that had IE5.5, and had to make a UI for a simple task on it. I chose to go with HTA. I tried to make my life easier by installing IE6, but failed to install it. It took me a while to even find an IE6 installer (the original links were all dead), and once I finally did, I couldn’t install it (it asked the internet for something as part of the process, and that part was broken).
If you could find what the installer was actually looking for, you could setup a proxy server to rewrite the URL and point it to someplace that has it e.g. locally.
I have no idea whether what it sought was important or not. Half of the difficulty in doing any debugging of anything was that most tools even in 2012 were no longer supporting 2000. Even Wireshark may not have run there; but I didn’t try it. I just gave up after a couple of hours of trying to get it working, and stayed on IE5.5.
Weeks of my life have been wasted working around the incidental complexity of IE6. Yet, it was a very sad day when I heard they're killing off Edge.
This tweetstorm of mine[0] is relevant:
I'm sad they are killing it off. Like really, really, sad. Edge is a feat of engineering. Like operating systems, and other major browsers. These things are rare. And they're way too advanced to just kill off.
I remember IE9 being the very first browser to bet on the web being a gaming platform (not counting Flash here). Chrome was the fastest for documents/GDocs-style apps, but nothing beat IE9 in multimedia performance.
I remember learning about layout thrashing and how it causes re-paint and jank. There was an infinite zoom demo somewhere on the internet that worked by resetting width/height on wheel events. Webkit/Gecko/Presto would render it at like 2 fps. Trident? 60 fps with lots to spare.
Interactive audio, pre Audio API was impossible in all other engines. I remember they had rewritten Cut The Rope in JS, but had to use Flash for audio. Reason was, calling el.play () on a fully loaded <audio> tag was laggy on all browsers. All except IE9 of course.
I remember they did not support 3D transforms. But the 2D stuff would run the smoothest on IE9. No scaling artifacts; no aliased edges; IOW, actually usable in a multimedia setting.
You cannot have interactive content without smooth animation and responsive audio. This vision, of seeing the web as a multimedia platform, was unique to IE. As in, they were the first one actually delivering on it.
In fact, IE9 was so good, the dev team behind it so talented, it only served to show how all that talent had gone to waste all those years between IE6 and 8. (One out of many examples of financial incentive stifling talent and innovation.)
And it wasn't only technical talent either. Their developer outreach program was special. They sponsored a whole bunch of projects, from rewriting popular mobile games in JS, to original web-based experiences, showcasing what had only just then become possible to do on the web.
That outreach program showed that the open web can be more than a static platform of paper-esque documents. Rather it can be a safe medium of interactive content, allowing ever more complex ideas to be communicated. This was powerful, and inspiring.
Back to the technical side, the vast majority of innovations in EdgeHTML will never be incorporated into Blink. The architectures surely are too different. Edge's innovations will simply bitrot in a close-sourced codebase waiting to be EOLed.
I can't imagine how disappointed the team must be. Years of their best work—truly great work—thrown out. Just like that. To my fellow engineers at Edge: You all did an amazing job. And you deserved better.
Last thing I'd say even if obvious: Use Firefox. Contribute to Firefox. Tell your friends to use Firefox. You'd be fighting the monoculture, AND getting the best browsing experience out there.
One of my favourite bits of IE was its high contrast support, which no other browsers had. It actually modified the styles of individual elements (rather than using a global color filter overlay like Chrome), so it performed really well, matched the OS high contrast theme, and could also be adjusted by developers with prefixed media query.
I was glad to hear that Microsoft will be adding it to the new Edge[0]. You can read about how it works in this excellent explainer[1].
Internet Explorer 2, in 1995, had support for VRML. A full 20 years before the Oculus Rift officially launched. I had never heard of that markup language and had to Google it. Talk about being ahead of its time...
Maybe my filter bubble, but VRML was kind of talked about in the late 90s (when we got on the internet) but the implementation, features, and general application state was more of a running joke. Also that's probably why it faded so relatively quickly, nobody used it.
I remember the most famous techno club from my nearest city used that to provide a virtual room tour on their website. I was too young then to get past the bouncers and liked that feature :)
The VR hypetrain left the station in the early-mid 90s. There were movies and TV shows about it. VR was going to be the revolution that changed everything -- William Gibson's cyberspace in your living room. But it was all experimental -- big clunky headgear and CPUs and GPUs that were too weak to produce any but a clunky, janky, laggy, polygon-y experience. And most people couldn't afford the fancy gear; for the most part VRML happened on people's monitors.
The recent VR hype happened because with modern PC hardware, the tech finally caught up to sustain an immersive, low-latency experience -- and some very smart dudes had the ambition to make 90s retrofuturism real.
I noticed the Edge page does not make any distinction between Spartan (current EdgeHTML-based version) and Anaheim (future Chromium-based version). The downloadable rar contains Anaheim though.
Browsing through the screenshots of each version on this site, I was struck by how good the IE logo looked from the version 4 installer. I realized why when I got to version 10, released some 15 years later. The flat "modern" update to the logo is identical to the one from IE4, just with a lighter font weight.
It is weird that IE 5 listed Mac OS 5.2.3 in the operating system. That's totally wrong for several reasons. First Mac OS was a name introduced since version 8 or so, and before that it was called System. Second I don't really think System 5.2.3 actually exists. Third and perhaps more importantly the Mac versions of IE were done by a completely different team within Microsoft, using a different rendering engine and having a different set of quirks. IIRC the Mac version gained full PNG support many years before the Windows version did; it was in fact the first browser to do so.
I wouldn't really consider IE for Mac together with IE for Windows. They are more like two different products with a same name.
Actually Internet Explorer 1.0 was a rebranded Spyglass browser . After failing to get an exclusive license from Mosaic Netscape and failing to get an exclusive license from the NCSA, Microsoft does get a non-exclusive license for a browser from Spyglass, that was 'given away' with Windows as Internet Explorer 1.0 around 1994. As such Microsoft deemed it unnecessary to actually pay Spyglass. Mosaic Netscape was released in 1994, despite this factoid, in 1996 we have Bill Gates welcoming Netscape into the Industry.
69 comments
[ 2.7 ms ] story [ 124 ms ] threadOne of the obvious things you can see between the versions is the gradual trend of dumbing-down and flattening the UI, most apparent between 8 and 9. The older versions' UI give the impression that they are very customisable and powerful, while the newer ones seem to be implicitly saying "you have no control (anymore)." I'd love to use a browser with a newer rendering engine, but retain the old UI.
I also noticed the sizes are a little weird:
I remember using those older versions of IE on systems with hard drives smaller than that.Their sizes seem to be off (or including something extra). But there definitely is a huge change in IE5 requirements. I'm guessing that's related to the US v Microsoft antitrust bundling lawsuit.
https://en.wikipedia.org/wiki/Internet_Explorer_4#System_req...
https://en.wikipedia.org/wiki/Internet_Explorer_5#System_and...
https://en.wikipedia.org/wiki/Internet_Explorer_6#System_req...
Instead Google got a huge fine in Europe now for being the default search engine on Android.
When we only had 640 vertical pixels, it was a huge change not to have 150 of them taken up with rows of icons and bookmarks by default. Now we have the space to get those controls back, but we're used to not having them.
It's AFAIK pretty much only with 4k monitors that we have caught up with CRTs resolution-wise, and not a lot of people have either of those. w3schools lists a bunch of resoltion statistics. In 2008, the most common resolution was 1024x768. 1366x768 in 2018. Same vertical resolution. Since I thought those statistics may account for phone users as well, I looked further. Globalstats lists 1366x768 as the most common desktop resolution in 2019. Desktop in this case probably accounts for laptops as well, as I believe there is no way to tell the two apart with User-Agent.
https://www.w3schools.com/browsers/browsers_display.asp
http://gs.statcounter.com/screen-resolution-stats/desktop/wo...
I also remember Firefox 2 being configurable enough that you could easily get less chrome than Chrome. The big win for me was process-per-tab.
For someone with bad eyesight like me, the sweet spot for a 13" laptop would be 4K at somewhere between 250% and 300% scaling. That would give me just about the same effective information density as good ol' 1366x768. I guess there's a reason why this awkward resolution still refuses to die.
Since all those toolbars can still be made visible in newer versions, I think MS's choice to hide them is more about a minimalist aesthetic than actually removing control.
This has been my experience for the majority of software, from native to web and mobile to desktop. Less options, less tools, less customization.
I think it's partly because software has become more mainstream over time, and instead of trying to teach people to use more complex things, developers have chosen or been forced into appealing to the lowest common denominator. Make it as simple as possible so people who aren't interested in learning how to use the tools can still get by.
I don't necessarily think this is a problem on its own, but combined with the business needs enforced by people looking to maximize profits it becomes less palatable. If we've built something that works for everyone, it's a hard sell to create more powerful options for the 10% (pulled that number out of my ass, I have no idea what the actual amount would be). Then we're all stuck with worse tools.
Personally, I like this thought-line a little better. Users ultimately want to interact with the application only with regards to their content, not the applications that house that content.
(Can't find any pages to open, though, guess it's too old http and possible tls stuff in the way?) Edit: Found some urls that worked
Edit: Not quite correct that it works. My domain responds with a different dummy-page in IE 1.5. I get the same dummy-page with curl
Difference being it's served as 200, while same command for example.com with no host gives 404.I think the web is just broken without host headers, as too much is running on the same ips / through routers before touching the real underlying page.
Then I felt angry thinking about it remembering all of the hacks to get it to behave as a web developer.
Fun roller coaster.
Yes. September was the month that many new users discovered the Internet for the first time, because they were given access at University and other higher education establishments, which had many and varied implications.
For instance there was an influx of beginner questions in Usenet groups, questions to which the answer was "please read the FAQ before asking more questions".
There was also just a general influx of activity: groups would see a bump in user numbers and activity that didn't fall back off over time.
New users arrived all through the year of course, but in smaller numbers and tended to be more technically proficient (and aware with network etiquette) to start with because they had sought (and paid for) access for a reason, rather than being given it and being told to go explore.
From 1993 onwards the public more generally were becoming a lot more aware of the Internet and being told about the things it could potentially do for them, so there was a more constant flow of new (often relatively non-technical) users & activity rather than just a big glut each September, which is why that year was referred to as "the September that never ended".
Edge just looks like Chrome, which is acceptable to me.
This tweetstorm of mine[0] is relevant:
I'm sad they are killing it off. Like really, really, sad. Edge is a feat of engineering. Like operating systems, and other major browsers. These things are rare. And they're way too advanced to just kill off.
I remember IE9 being the very first browser to bet on the web being a gaming platform (not counting Flash here). Chrome was the fastest for documents/GDocs-style apps, but nothing beat IE9 in multimedia performance.
I remember learning about layout thrashing and how it causes re-paint and jank. There was an infinite zoom demo somewhere on the internet that worked by resetting width/height on wheel events. Webkit/Gecko/Presto would render it at like 2 fps. Trident? 60 fps with lots to spare.
Interactive audio, pre Audio API was impossible in all other engines. I remember they had rewritten Cut The Rope in JS, but had to use Flash for audio. Reason was, calling el.play () on a fully loaded <audio> tag was laggy on all browsers. All except IE9 of course.
I remember they did not support 3D transforms. But the 2D stuff would run the smoothest on IE9. No scaling artifacts; no aliased edges; IOW, actually usable in a multimedia setting.
You cannot have interactive content without smooth animation and responsive audio. This vision, of seeing the web as a multimedia platform, was unique to IE. As in, they were the first one actually delivering on it.
In fact, IE9 was so good, the dev team behind it so talented, it only served to show how all that talent had gone to waste all those years between IE6 and 8. (One out of many examples of financial incentive stifling talent and innovation.)
And it wasn't only technical talent either. Their developer outreach program was special. They sponsored a whole bunch of projects, from rewriting popular mobile games in JS, to original web-based experiences, showcasing what had only just then become possible to do on the web.
That outreach program showed that the open web can be more than a static platform of paper-esque documents. Rather it can be a safe medium of interactive content, allowing ever more complex ideas to be communicated. This was powerful, and inspiring.
Back to the technical side, the vast majority of innovations in EdgeHTML will never be incorporated into Blink. The architectures surely are too different. Edge's innovations will simply bitrot in a close-sourced codebase waiting to be EOLed.
I can't imagine how disappointed the team must be. Years of their best work—truly great work—thrown out. Just like that. To my fellow engineers at Edge: You all did an amazing job. And you deserved better.
Last thing I'd say even if obvious: Use Firefox. Contribute to Firefox. Tell your friends to use Firefox. You'd be fighting the monoculture, AND getting the best browsing experience out there.
[0] https://twitter.com/ariaminaei/status/1072516777937129473
I was glad to hear that Microsoft will be adding it to the new Edge[0]. You can read about how it works in this excellent explainer[1].
[0] https://techdows.com/2019/03/chrome-to-get-native-caret-brow...
[1] https://github.com/MicrosoftEdge/MSEdgeExplainers/blob/maste...
Source: A friend of a friend, of course...
The recent VR hype happened because with modern PC hardware, the tech finally caught up to sustain an immersive, low-latency experience -- and some very smart dudes had the ambition to make 90s retrofuturism real.
I wouldn't really consider IE for Mac together with IE for Windows. They are more like two different products with a same name.
https://web.archive.org/web/20120618054817/http://www.nation...
https://www.cnet.com/news/gates-draws-roadmap-for-intranet/