Most things worked far, far better 10 years ago than today. Apparently this is progress.
Social everything means we can't visit a website today without calls to a billion external domains. Phones are doing the same as they're mini-tablets that suck at actual phone calls. Operating systems have taken on mobile centric (and the hideous flatten everything) design that is broken paradigm for a 15" laptop or 24" desktop screen. Firefox was a stripped down, light browser so it stayed fast - now they include more useless garbage than Netscape Communicator ever did. All that crap should be in official plugins so you add pocket only if you want it etc.
Nothing at all is user centric any more, it's all about how much lovely data can be sent back to hundreds of places.
The web is a lot prettier these days to be fair, but I don't call that progress if it's at a cost of a 1TB page load.
> Most things worked far, far better 10 years ago than today. Apparently this is progress.
Sigh. If you wanted to look up a business on your phone you'd have to call directory enquiries - did that "work better" than your smartphone? Did you prefer RealPlayer over HTML video? Was WAP better than LTE? Was Windows XP the pinnacle of OS design?
Yeah, sometimes progress is a double-edged sword that outdates some things we used to do. Yes, it's worth the price.
Having so many apps constantly spitting unnecessary data out to trackers and heaven knows what isn't better. Neither is simple apps that can't cope when you go out of data coverage, which seems to be becoming more common.
I'd rather have gnome 2 than 3, I'd rather have Win 7 than 10 or XP. I don't consider any of them the pinnacle of design, but I don't think they're currently moving in a helpful direction either. There's been precious little progress or innovation in the OS space for years.
I'd rather have google talk, which was small, light, minimal and crucially reliable over hangouts or current skype.
Just because software isn't moving in a way that you personally want doesn't mean it's wrong or that it was objectively "better" before.
We don't have apps that work offline? Maybe look around more. FFS half of the web apps i use work fine offline.
And claiming that there has been no innovation in the OS space for years is silly. There is plenty of innovation, it's just that people don't like change and will instantly reject something that isn't the same as what they are already using. FFS just look at stuff like the ip command or systemd in linux. People lament that it's the end of days because a command is different exactly because they want to innovate past the constraints of the old systems.
And if we are playing the "What i want is what is best" game...
I much prefer Windows 10 over linux now. It's gotten that nice and i'm tired of constantly fixing stuff on my workstation.
And I much prefer current hangouts to the old google talk plugin. "small", "light"? Did you ever use it. You needed to download and install 2 separate programs to get Google Talk video chat working. Sometimes you'd need to do some port forwarding bullshit just to get it to work.
Now i install chrome and sign in, and within a minute hangouts is installed and working.
> And I much prefer current hangouts to the old google talk plugin. "small", "light"? Did you ever use it. You needed to download and install 2 separate programs to get Google Talk video chat working. Sometimes you'd need to do some port forwarding bullshit just to get it to work.
I'm pretty sure the GP is referring to the native google talk client that gtalk was originally launched with. It was indeed a very simple jabber+sip client (by far the simplest of the major available clients) and had a tiny memory footprint. It did nothing but chat and voice calls.
And you needed to install that client, then install the video and voice chat plugin if you wanted that, then sign in to your account (and if you had 2 factor auth you needed to jump through some other hoops).
It just seems like people will rewrite history. It was cumbersome, had strange scaling issues on many of my PCs at the time, didn't work on macs, you needed to just do it yourself on linux (and i'm not even sure if videochat worked on linux...).
I still think you're talking about a different, later thing. The client I'm talking about had no plugins and didn't support video, but did support voip directly. It was actually a selling point at the time over MSNM/AIM, iirc. You're right it didn't work on macs, or linux for that matter, because it was a native windows program and they never released a client for any other platform and there was no web interface to it at the time (but you could use any xmpp client with it, as you still can).
I'm pretty sure it was also before google even had TFA.
Windows 10 is a pretty good OS but has serious issues with drivers (it tries to be clever and breaks stuff). And the startup is sooooo slow and the UI crashes twice before I ever get the chance to give it input.
The Mail, Calendar and OneNote apps kept me hooked. But the lack of a decent terminal was killing me. Also, Skype didn't work properly half the time and the new "Video" and "Messaging" apps built on top of Skype were crashing 90% of the time.
Windows 8 came preinstalled with the computer and then just updated to Windows 10 from it. If the main upgrade method is broken, I'm not sure what to believe anymore...
If Windows 8 came preinstalled, and you didn't wipe it and install from a clean ISO yourself, then you've got more crapware installed than you can shake a stick at, with hooks riddled all through your system. Doubtless, some of that vendor-supplied junkware is incompatible with Windows 10.
There's the occasional odd package under Linux which remains installed and you need to manually remove it, but nothing unfixable. But I guess that depends more on the underlying OS structure. I'm looking forward to see more widespread usage of Snappy Ubuntu.
Since that is the main way people migrate to Windows 10, I would've expected them to be more careful with it.
Many linux systems don't even support in-place upgrades between major versions. And I don't think i've ever had one work where nothing broke.
RHEL only started supporting it for the transition between 6 and 7 on a like 3 architectures on one edition.
And OSX has it's share of issues. For me personally, updating to El Capitan was the worst upgrade processes i've done in recent memory. From not running the installer but giving no issues, to not finding the harddrive during the install process, to corrupting the current install of yosemite, to bluetooth being broken after the upgrade, and i can't get wifi speeds over 1mbps.
A clean install solved all of that, but that's pretty normal with every single OS i've ever encountered.
It's not a question of being "more careful", it's that writing software in a way that it can be in-place upgraded to something you don't know will exist at the time is EXTREMELY HARD! I really feel it's one of the big "unsolved problems" in computer science and i don't see it being solved any time soon.
I agree with untog, you need to get your install fixed. I'm running on hardware from 2011(except the SSD) since launch and Windows 10 is extremely fast to boot and the UI hasn't crashed once. It even runs my security camera software that was last updated in 2008. I only recently re-installed Skype so I can't speak to that but it has problems all of it's own.
> There is plenty of innovation, it's just that people don't like change and will instantly reject something that isn't the same as what they are already using.
It's not change that people don't like, it's the mountains of randomly missing features, settings and control remotely subverted away from the system owner, tons of software no longer working, a terrible UI nobody asked for. Who wants 10 steps of innovation forward if it comes bundled with 1,000 undocumented steps backward?
The ip command is a monstrosity. It is basically trying to emulate Cisco's IOS in a command to placate those that want to use Linux as their router firmware.
compare something simple as enabling a network card using ifconfig vs using ip, as you can easily tell the difference. You end up 3 layers deep before you can even enter the card id you want to do anything with!
When seasoned kernel devs don't want to touch a tool, warning lights should come on.
When it's not something you want, it's to "Placate people". But if it was something you wanted, i'm sure it would be a welcome feature!
And it's not that complicated, just different.
Want a "traditional" list of interfaces? `ip addr`
Want to list running interfaces? `ip link ls up`
Want to set a device up/down? `ip link set dev {DEVICE} {up|down}`
Quick question! How do you set the MTU length in a legacy REHL system? Because with ip it's just `ip link set mtu 9000 dev eth0`!
Yeah, its new and will require some learning, but it's not inherently bad...
And talking about seasoned kernel devs, name one time that there wasn't a minority of kernel devs bitching about a new feature! Every change is going to make some people worse off.
"just". Ifconfig gets the job done, unless RH has been odd selves again...
As for devs, Torvalds didn't have much love for the ip command last time i looked into things. And a year or so back i ran into a Ts'o posting about his dislike for polkit (a close cousin of systemd).
> If you wanted to look up a business on your phone you'd have to call directory enquiries - did that "work better" than your smartphone?
Bad example. I didn't need a smartphone for this task because there was a real live human being, who brought to the task all of the adaptability and intuition that modern systems lack. I didn't need Yelp, or Google Maps, or even the entire name of the business I was looking for. "I need the number to the printing company on the east side" was enough to get me a name, an address, and the call routed, without paying monthly bandwidth fees or spending seven hundred dollars on a pocket supercomputer.
Realplayer sucked but it was about a thousand times more reliable than the current compatibility crapshoot (sans bandwidth issues).
Worth the price? Sure, but to quote Pulp Fictoin, let's not start beating each other off just yet. There is plenty of room for improvement, and we've lost a lot of ground.
I used to be able to take a supersonic flight across the country, without having my body inspected via millimeter-wave radar.
> I didn't need a smartphone for this task because there was a real live human being, who brought to the task all of the adaptability and intuition that modern systems lack.
OK, my example was simplistic. I want to find the best chicken tacos in San Francisco. I can't call directory enquiries for that - I have go find reviews. Then call to make a reservation at the place, then find driving directions of how to get there, somehow. In a paper road atlas? Am I having to advocate for the benefits of the internet on Hacker News?
> Realplayer sucked but it was about a thousand times more reliable than the current compatibility crapshoot
Well of course it was, it only had to run on one platform. Compatibility is a lot easier when that's the case. Today we have HD video being played on mobile devices in your pocket. I can't remember the last time I ran into a compatibility problem with online video.
> I used to be able to take a supersonic flight across the country, without having my body inspected via millimeter-wave radar.
That's a political decision, not a technical regression.
> Has anyone set foot on the moon in your lifetime?
No. Can I load, on demand, stunning satellite photography of planets orders of magnitude further away than the moon, on my phone, on the bus to work? You bet.
> That's a political decision, not a technical regression.
It's both. Private supersonic flights are no longer available, and it is more of a pain in the ass to get on the inferior flights that are available.
I could go on, but it's moot -- the only point I'm here to make is that progress always been paired with regression, and it's always going to be a personal decision as to whether the tradeoffs were good ones.
I call BS on the review side. Where is the 'best chicken tacos in San Francisco'?
I have no idea and Yelp is not going to help. So, the fallback is I can ask some friends for advice and directions or use some companies best of list. But, again the internet did almost nothing in this space.
Sorry, but I'd be surprised if Google or Siri don't answer your question to "I need the number to the printing company on the east side", it's a pretty trivial thing to lookup. Also, Google and Siri bring you cost efficiency and scale, something you can't easily do with directory enquiries.
> Realplayer sucked but it was about a thousand times more reliable than the current compatibility crapshoot.
So Vorbis/MP3 and WebM/H.264 is not good enough for compatibility over all the browsers?
They might have been better at the small subset of things Windows and Ubuntu can do that the Amiga and ST were also capable of, but they are not "better".
A single screwdriver is better than a toolbox at being light and easy to carry. It is not better overall.
And I'll look you right back and call bullshit. I used an Amiga in the late eighties. It was a fantastic machine for its time, certainly more advanced than the contemporary 286 PCs, but Windows 95 was markedly superior (as it should have been, a decade later), let alone modern Windows or Linux.
Yes, sometimes things get screwed up and we get understandably frustrated, but that does not license revisionist history.
> we can't visit a website today without calls to a billion external domains
NoScript really brings this into visibility. I don't mind the domain running scripts. But the 10+ other domains that needs to run a script - sigh. My wife keeps asking my why I put up with it - it's quite simple; I decide if the page I'm about to visit is worth all the extra crap running and tracking me.
Personally I prefer uMatrix - it lets you enable specific things across domains instead of a binary allow/deny, and allows global versus per-domain or per-subdomain rules; e.g. I might be fine with one site collecting metrics but not globally.
In my observation, this has certainly been the trend.
In my experience, in this climate you have to look harder to find it user-focused software, and to some extent you have to write some of it yourself.
This is because the software world has been flooded with "easy to use" programs that are either located on third party computers or must automatically connect to third party computers in order to be useful.
To be truthful, I'm not sure that many users of today, i.e., younger ones, really understand what we mean by user-centric and user-focused. They may think if something is made "easy" for them, then it's user-focused.
But where computer software is involved, we've seen it is precisely the opposite. If users want more control then they have to endure a little inconvenience, at least in the short term.
The easier and more "frictionless" the software, the more the user should understand that the software has not been written so much for their benefit as for someone else's.
I found out more about "system" addons. There are different types of addons: app-profile, app-system-addons, app-system-defaults, app-global, app-system-local, app-system-share, app-system-user, app-temporary.
The "app-system-defaults" (in my case Pocket & Loop/Hello) don't seem to be shown to the user, but you can find them in the extensions.xpiState config.
>Most things worked far, far better 10 years ago than today. Apparently this is progress.
Not everything; cars today are a big improvement over cars 10 years ago, and especially cars 20 years ago. Just in the last 10 years, cars have made huge improvements in fuel economy alone. Cars have gotten really good for the most part. Some article I recently read on an auto blog, about the "10 worst" cars, even complained that modern cars are so great now that their picks were really just about which ones were the worst values relative to their competition, and that, unlike in past ages, they couldn't point to any true stinkers.
Also, I think mobile phone software is generally better now than 10 years ago. 10 years ago the iPhone was brand new, and most people still used crappy feature phones running things like BREW. They just didn't give you the capabilities that modern phones do. Some people complain that newer phones have poorer call quality than old flip-phones, but personally talking on the phone is one of the least-used features. I spend a lot more time using it for texting, photos, voice mail, and various apps (taking notes, calculator, playing games, Tinder, etc.).
However, in the software world, I will readily agree that most PC-based software was far, far better 10 years ago than today. PC-based software in the last 5 years has really, really gone down the toilet. Windows 8/10 are a prime example here, but even Linux distros really aren't doing that great either. The web is a horror show, with every site full of dozens of JavaScript scripts running, mainly for spying/tracking purposes. I don't even agree that the web is prettier; too many sites are mobile-centric, and look stupid on a regular PC screen.
Are you kidding? Automatic seatbelts are incredibly stupid, and are provably bad. The only reason they were invented was because the federal government required cars to have "passive safety devices" installed, and they were aiming for airbags. But lots of shitty car companies like Saturn were too cheap to install airbags so they installed mousebelts instead.
The whole problem with them is that they're dangerous as hell, and offer zero benefit, and only drawbacks.
They're dangerous because they encourage people, by supposedly being "automatic", to not buckle up. They're not effective if you don't buckle the lap belt, but during that time lots of people didn't, because they thought it wasn't necessary. After all, what's the point of an "automatic seatbelt" if you still have to manually buckle part of it?
So, if you're going to go to the trouble of buckling the lap belt to get full effectiveness, what exactly is the benefit over having a normal 3-point belt? There is none.
Finally, the belts were a total hindrance if you got in your car with anything in your hands, such as a briefcase or purse or bag. Anyone driving alone (as most of the population does most of the time) and carrying something frequently just jumps in the front seat holding the bag, and puts it in the passenger seat or floor. With the mousebelt in the way, this became a good way to get tangled up.
If you're still driving that piece of shit, do yourself a favor and get a newer car. That thing has no airbags and has terrible crashworthiness compared to anything newer than 25 years old. You're very likely to die in a crash in that car, which in a new car you would walk away from.
>Firefox was a stripped down, light browser so it stayed fast - now they include more useless garbage than Netscape Communicator ever did.
Sorry, I don't think Pocket and Hello surpasses the junk bundled with Communicator, which came with a custom mail client. Mozilla's mail client is not only not installed with Firefox, but it's also pretty much dead now, in fact.
Email was the only other useful part. Thunderbird is not exactly dead, just forgotten. I still run it, since I de-googled, though sadly it's also turning into bloatware.
Netscape had a chat thing, html publisher, and I think something else.
WebRTC, pocket, the webIDE, even things like the dom inspector, style editor etc and pdf are bloat for most mortals. Perfect for some "official" plugins. They managed it for the FF OS simulator.
It's all optional crap that increases bloat and attack surface, except it's not optional.
Why is WebRTC bloat? A lot of people us it for video calls nowadays. Pocket is a neat solution for avoiding having lots of tabs eating your memory, which in turn ends up causing page swapping and messing with your experience.
"pdf are bloat for most mortals"
That's alright, it means in 99% cases it's just a very small javascript file on your disk. No harm done.
Bloat is relative and subjective. I've never used WebRTC, Pocket, or Loop/Hello, so I'd consider them bloat. On the other hand, I need to deal with PDFs fairly frequently, so it's convenient to have a browser that supports them. Someone else is going to have different requirements. I don't have a problem with that, but I'd prefer to have an option to disable or remove the features that I don't need or want.
I'd say the situation now is about the same as it was then. Only the context has changed. Communicator had a mail client and a news client. I used them both over dial-up, back in the day. If enough people still wanted newsreaders, you can bet Mozilla would write one.
I used Thunderbird daily for the better part of a decade until about 3 months ago. It's nothing personal against the project, which is currently the best serious desktop mail client available afaik. It's simply that Mozilla doesn't care about it anymore since everyone uses webmail now. Thunderbird is doing feature releases about once a year now [0], and Mozilla has pulled most (all?) of its full-time developers [1]. A few months ago, Mitchell Baker sent a message that said they were looking to detach Mozilla from the project completely, partially because it didn't have enough industry-wide impact to be worth Mozilla's resources [2].
The sad fact is that desktop mail clients are going away, and even the big projects like Thunderbird are barely limping along, which I guess is slightly nicer terminology than "pretty much dead"?
What makes an email client so difficult to write? I don't mean this to sound insulting, I just really don't understand where the complexity is. It seems like you should be able to get to a point where it is feature complete and all development efforts shift to maintenance mode. Or maybe I don't understand what's been going on with Thunderbird and perhaps it's been "done" for years now.
It is basically done. There could be UI refinements and updates to the occasional protocol/security change, but yeah, there's not been much innovation in email standards for a while.
Although, to be honest, email standards are a complete mess in some regards and I suppose this make such an endeavour not as trivial as it looks on first sight.
Just an example: try to create (and use) an IDN-based email address with non-ASCII characters in the local part. You will be amazed how incompatible on so many layers this actually is. I can hardly imagine there weren't dozens of similarly tricky situations where the developers of this, or any other email client for that matter, were forced to make quite serious design choices because of either ambiguity in standards, or widespread lack of support thereof.
I don't like the fact that it uses an external API. I don't like a third party (other than my mail host) to have access to my email password and messages.
Thanks. It wasn't meant to be sarcastic although it surely looks like it when I read it now. I was genuinely interested to find out about it since it's a program I rely on and if it won't be developed further, I'll have to look for alternatives.
As the responses in this thread demonstrate, there's plenty of cause for both pessimism and optimism. So let's all choose to be optimists, then get on with improving our own little piece of the world.
Social everything means we can't visit a website today without calls to a billion external domains. Phones are doing the same as they're mini-tablets that suck at actual phone calls. Operating systems have taken on mobile centric (and the hideous flatten everything) design that is broken paradigm for a 15" laptop or 24" desktop screen. Firefox was a stripped down, light browser so it stayed fast - now they include more useless garbage than Netscape Communicator ever did. All that crap should be in official plugins so you add pocket only if you want it etc.
Nothing at all is user centric any more, it's all about how much lovely data can be sent back to hundreds of places.
The web is a lot prettier these days to be fair, but I don't call that progress if it's at a cost of a 1TB page load.
I could go on, but you get the idea.. :)