4 ms·

I do not think people remember how bad software was over a decade ago or even more. Web browsers for example. I remember the days before Firefox got that quant

I do not think people remember how bad software was over a decade ago or even more. Web browsers for example. I remember the days before Firefox got that quantum update and the crashes went down. Chrome tabs crashing. Adobe Flash. Windows Vista came out in 2007. Windows 7: Windows Vista with proper QA and drivers. Windows 8: the user interface. Yes, some of you liked it, most didn't. It was very bad. Internet Explorer... Windows 98... Windows ME... Windows 10 was so much better. Windows 11 is the next Vista phase. Mesa on Linux before Valve started hiring contractors to work on it. Remember the days when a major release of Mesa was based on if it implemented a new version of OpenGL? AMD graphics before AMD properly supported it on Linux. AMD drivers on Windows before Youtubers started actually using AMD cards and complained about the drivers. Games For Windows Live was closed on August 15 2013. This was before fun things like mitigations for hardware vulnerabilities that software has to deal with. Or the growing scale problems AWS has to deal with it controlling 90% of a growing market. Even Bing has gotten... a little better... I guess Mac OS X has objectively gotten worse, but it has been that way since Snow Leopard (half joking). Google search has objectively gotten worse.

Aardwolf 3y ago
> Adobe Flash. The fact that it got killed is literally an example of software being worse now than 10 years ago, thousands of fantastic little games were implemented in Flash, this was an easy way to create and distribute games causing many little labors of love being made, and then the deliberate choice was made to kill it. Fortunately the ruffle project revived many of those games now.
- dvngnt_ 3y ago
  websites used it too.
- csdreamer7 3y ago
  > thousands of fantastic little games were implemented in Flash True... but also when your right clicked in a Flash widget on Firefox on Linux in would cause the browser to hang. This problem existed for years. That is why I listed it.
  - Aardwolf 3y ago
    > when your right clicked in a Flash widget on Firefox on Linux in would cause the browser to hang. I used flash in firefox in linux all the time and used right click often to pause flash animations and skip frames forward/backward and this never crashed/hung for me, so at least it wasn't a general linux issue... The only annoying thing was that there was a time period of multiple years where the linux flashplayer was not being updated anymore and more and more Newgrounds games started not working for Linux, then at some point adobe finally turned around and made linux versions again... (maybe it was between 2012 and 2016 with versions 11 to 24 but somehow my memory was that this was longer ago)
hinkley 3y ago
About half the processes I was driving myself crazy trying to get adopted fifteen, twenty years ago are more or less standard now.
nonrandomstring 3y ago
> I do not think people remember how bad software was over a decade ago Or how good it was. Pro24 and then Cubase on the Atari ST ran for months without crashing. Desktop publishing on the Amiga (PageFactory? - I forget now) producing LaTeX/Postscript like output, faultlessly. Those were in 1993. Before that I used VLSI software I ran on a Sun 380 workstation and designed whole microprocessors on software that simply never crashed - even when it ran out of memory or was given 'impossible' routing tasks it auto-saved in a recoverable state - in 1989! There's always been the same spread of software quality. Some amazing stuff written by diligent, competent engineers who actually use and test their wares. Some absolute rubbish written by fly-by-night spivs and opportunists who want to make money fast. Sadly the latter gained the upper hand after about 2000. Mostly led by companies like Microsoft whose products are notoriously shite - a total disgrace to computing. They got away with it simply by pumping vast amounts of money in to marketing and perception management. They simply drowned out the competition, creating a normalised very low standard of software by sheer force and dominance. Two problems lie beneath this. Firstly, software quality metrics have never really matured. It's too complex a product a gauge. The author mentions functionality, usability, reliability and security as if they were orthogonal. But in reality they are not. For instance, security is mainly a quality issue relating to visibility and control and is inversely proportional to complexity. And aren't ease of installation, accessibility and ease of development also quality factors? Secondly, I don't think the average persons has much rational expectation around software quality. Again Microsoft and then Google proved that. Many people are pleased by eye-candy and UX fluff, or by the social aspects of software, but they disregard actual functionality. Ask random people whether software got better or worse this last decade and I think you'll get a lot of confusing, scattered responses,
II2II 3y ago
You forgot the Windows 3.x era, where up-time was often measured in hours. From my recollections, early Windows 95 was not much better. Of course, Windows 9x and earlier were single user operating systems with no security to speak of. Organizations frequently put on a layer of access control, and that layer was frequently completely ineffective. Speaking of ineffective security, I recall sessions being identified by IP address. Have more than one computer logged in behind a NAT router, well, you were all accessing the same account. Passing the password as part of the URL? Why not! Many protocols even sent passwords as clear text, never mind encrypting the actual data being shared. And while a lot of that stuff happened in the 1990's, some of it was happening well into the 2000's. That being said, I think we could be doing better than we currently are if we had simpler and better understood systems. It is almost as though we insist upon making new mistakes after learning from the old ones.
- MichaelZuo 3y ago
  As a real world example my ipad pro on the latest iPadOS 17 can't even maintain 2 weeks of uptime without crashing or getting so slow it becomes unusable without a restart. And I've tested it multiple times on normal workloads using just the stock apps without ever connecting it to external devices I wasn't expecting 20 years of uptime like an AS/400 but 2 weeks is absurdly short reliability for under the most ideal circumstances.
xnx 3y ago
It's been at least 10 years since I've gotten a blue screen of death (or equivalent)

Aardwolf 3y ago

> Adobe Flash. The fact that it got killed is literally an example of software being worse now than 10 years ago, thousands of fantastic little games were implemented in Flash, this was an easy way to create and distribute games causing many little labors of love being made, and then the deliberate choice was made to kill it. Fortunately the ruffle project revived many of those games now.

dvngnt_ 3y ago
websites used it too.
csdreamer7 3y ago
> thousands of fantastic little games were implemented in Flash True... but also when your right clicked in a Flash widget on Firefox on Linux in would cause the browser to hang. This problem existed for years. That is why I listed it.
- Aardwolf 3y ago
  > when your right clicked in a Flash widget on Firefox on Linux in would cause the browser to hang. I used flash in firefox in linux all the time and used right click often to pause flash animations and skip frames forward/backward and this never crashed/hung for me, so at least it wasn't a general linux issue... The only annoying thing was that there was a time period of multiple years where the linux flashplayer was not being updated anymore and more and more Newgrounds games started not working for Linux, then at some point adobe finally turned around and made linux versions again... (maybe it was between 2012 and 2016 with versions 11 to 24 but somehow my memory was that this was longer ago)

hinkley 3y ago

About half the processes I was driving myself crazy trying to get adopted fifteen, twenty years ago are more or less standard now.

nonrandomstring 3y ago

> I do not think people remember how bad software was over a decade ago Or how good it was. Pro24 and then Cubase on the Atari ST ran for months without crashing. Desktop publishing on the Amiga (PageFactory? - I forget now) producing LaTeX/Postscript like output, faultlessly. Those were in 1993. Before that I used VLSI software I ran on a Sun 380 workstation and designed whole microprocessors on software that simply never crashed - even when it ran out of memory or was given 'impossible' routing tasks it auto-saved in a recoverable state - in 1989! There's always been the same spread of software quality. Some amazing stuff written by diligent, competent engineers who actually use and test their wares. Some absolute rubbish written by fly-by-night spivs and opportunists who want to make money fast. Sadly the latter gained the upper hand after about 2000. Mostly led by companies like Microsoft whose products are notoriously shite - a total disgrace to computing. They got away with it simply by pumping vast amounts of money in to marketing and perception management. They simply drowned out the competition, creating a normalised very low standard of software by sheer force and dominance. Two problems lie beneath this. Firstly, software quality metrics have never really matured. It's too complex a product a gauge. The author mentions functionality, usability, reliability and security as if they were orthogonal. But in reality they are not. For instance, security is mainly a quality issue relating to visibility and control and is inversely proportional to complexity. And aren't ease of installation, accessibility and ease of development also quality factors? Secondly, I don't think the average persons has much rational expectation around software quality. Again Microsoft and then Google proved that. Many people are pleased by eye-candy and UX fluff, or by the social aspects of software, but they disregard actual functionality. Ask random people whether software got better or worse this last decade and I think you'll get a lot of confusing, scattered responses,

II2II 3y ago

You forgot the Windows 3.x era, where up-time was often measured in hours. From my recollections, early Windows 95 was not much better. Of course, Windows 9x and earlier were single user operating systems with no security to speak of. Organizations frequently put on a layer of access control, and that layer was frequently completely ineffective. Speaking of ineffective security, I recall sessions being identified by IP address. Have more than one computer logged in behind a NAT router, well, you were all accessing the same account. Passing the password as part of the URL? Why not! Many protocols even sent passwords as clear text, never mind encrypting the actual data being shared. And while a lot of that stuff happened in the 1990's, some of it was happening well into the 2000's. That being said, I think we could be doing better than we currently are if we had simpler and better understood systems. It is almost as though we insist upon making new mistakes after learning from the old ones.

MichaelZuo 3y ago
As a real world example my ipad pro on the latest iPadOS 17 can't even maintain 2 weeks of uptime without crashing or getting so slow it becomes unusable without a restart. And I've tested it multiple times on normal workloads using just the stock apps without ever connecting it to external devices I wasn't expecting 20 years of uptime like an AS/400 but 2 weeks is absurdly short reliability for under the most ideal circumstances.

xnx 3y ago

It's been at least 10 years since I've gotten a blue screen of death (or equivalent)