Comment by arximboldi | Hacker News Reader

Comment by arximboldi | Hacker News Reader

Parent Full thread

"Holding back the industry"

No!

I think we, the software industry, are very irresponsible when it comes to __planned obsolescence__ [1]. Computers from 10 years ago are very capable of doing most things common users want to do nowadays. However, many modern operating systems and applications make these computers unusable.

We are forcing users to upgrade software, and because of all the layers of cruft accumulated in this software, they are then forced to upgrade their hardware too. This just adds to a spiral of consumerism that is going to destroy our planet.

Like most of us, I like developing with the latest high level fancy cutting edge stack; but we have to remember that it's not the society's duty to empower the tech industry but the other way around. It is our role to help people solve people's their problems, and in ways that are effective socially and ecologically and does not create waste. This also means, making sure that hardware does not become obsolete because of our carelessness, lazyness or selfabsorpion as software makers.

Also, alternatives do exist.

I know someone in Cordoba (Spain) that recently started a __slow computing__ cooperative, inspired by this article [2]. His goal is to bring tech experts and ordinary people together to build custom solutions adapted to their needs. They have users go to them and do an interview with them. Then, they make a GNU/Linux and libre software installation for them, often reusing old hardware, adapted to their personal needs, and do the teaching necessary. I find this a very interesting approach, that solves the security problem both by installing updated software that does not create hardware obsolescence, and educating users so they regain control of their machines.

There is no valid excuse for Microsoft here. Your argument is the perfect example of how proprietary software puts end users as servants to the interest of owners of the proprietary code that control their machines.

[1] https://www.wikiwand.com/en/Planned_obsolescence [2] https://newrepublic.com/article/121832/pleasure-do-it-yourse...

Whether the OS is proprietary isn't the important facet of this issue. The case for encouraging users to move off of Windows XP is the same as the case for encouraging users to move off of older GNU/Linux distributions: security updates.

Even in a slow computing initiative, outside invention will continue. Take ASLR, DEP, and other memory corruption mitigations. The architects of the legacy, fixed-address systems didn't expect the large amount of vulnerability due to buffer overflows in programs that handle external input. The authors of these programs failed to write tests that exercised some edge cases of malformed input. While long/malformed input unit tests might seem germane today, they weren't then.

Custom software initiatives make sense for people who have a high aptitude and the time to undertake them. This use case doesn't fit the vast majority of people who use computers.

While Microsoft certainly deserved the $10k judgement against them, software authors have the duty to help mitigate past imperfections, errors, and lack of foresight in their software. Authors of free and proprietary software share this responsibility.

> Whether the OS is proprietary isn't the important facet of this issue.

The OS being proprietary is an important issue. With libre software, lightweight forks that still work in old hardware can be maintained. Also, the software stack can be customized. Note that it's not security patches what makes software slow, but the zillions of layers of indexers, UI frameworks written in three layers of dynamic languages, etc that are included in a desktop environment. A modern GNU/Linux distribution function decently on old hardware as soon as you swap the desktop environment. With either you upgrade to a 20GB OS with a desktop environment that requrires 2 gigs of ram and so on, or you are stuck with an unsupported Windows XP.

> Custom software initiatives make sense for people who have a high aptitude and the time to undertake them. This use case doesn't fit the vast majority of people who use computers.

The average american watches 4 hours of TV a day [1] and they spends more time on Facebook than taking care of their pets [2] and takes weeks of training in order to get a driving license. I think the average person has the time to spend one or two days, when they upgrade their computer, pairing with a technician to get to understand the software they have and how it works and how to protect themselves. It's a matter of priorities. Sadly, for a lot of the software industry, an empowered userbase is not in their best economic interest.

[1] https://www.csun.edu/science/health/docs/tv&health.html

[2] http://www.bloomberg.com/news/articles/2014-07-23/americans-...

My friend sent me the link to the slow computing/technology sobereignty project:

http://tianguis.cordoba.cc/que-hacemos/soberania-tecnologica...

Upgrading software isn't (or shouldn't be) planned obsolescence. There's nothing inherent about software that means throwing away hardware. The only reason a Windows 7 machine wouldn't be able to run Windows 10 is if the hardware maker chose to not update their drivers, and at some point you have to stop blaming Microsoft for that.

There is no such thing as planned obsolescence when it comes to software. "Throwing away" software creates no pollution and Windows 10 runs on basically anything Windows 7 ran on.

>There is no valid excuse for Microsoft here.

There is no valid blame for Microsoft here. Windows 10 will run just fine on 10 year old hardware, as long as the device drivers are compatible.

How about 100% everything instead of "basically anything", without serious "if's" like "as long as drivers are compatible" ?

If a machine is running, say, windows xp, and it does in a manner that satisfies the owner's need, then it should be able to keep running forever in the same manner, provided that hardware lasts and/or is replaced with adequate spares, not necessarily anything newer.

It's ok if that computer becomes incompatible with newer software because of that.

It's ok if that computer becomes incompatible with newer outside/network services because of that.

It's not ok if that computer stops being able to execute exactly the same process that it could do 10 years ago.

Because Microsoft didn't write the drivers for 100% of hardware... and tbh, if a company wants they can pay Microsoft to write those initial drivers, and provided the hardware for testing, iirc, they tend to maintain those drivers between versions. There are plenty of drivers that are written by MS, it's pretty common with Printers iirc, and works pretty damned well.

Because foo pro-audio, or bar cnc-board didn't want to update their drivers, or for that matter, more often than not, had a weird installer that simply didn't follow proper practices, doesn't work in the new version of Windows isn't MS' fault.

Nevertheless, because the chance of something like that happening is clearly higher than 0.0000000001%, Microsoft should simply never do such a drastic upgrade automatically. (So, even within Windows 10, they should not automate all upgrades, because occasionally they will have to include bigger changes.)

Also, users will not understand this, and will, rightly, blame Microsoft.

Any trivial update has a higher than 0.0000000001% chance of failure... if you need that level of specificity, you should disable updates, and network connectivity altogether.

That's exactly how it exists. What you're describing here is reality. If you have a machine running Windows XP and you never change anything, then nothing will change. Software doesn't magically bloat over time, those changes happen with updates that add new features. If you never update things, you'll stay on the same software revision and it will run the same as the day you bought it.

A company I used to work for in 2012 had a Windows 3.11 workstation running some ancient software to support an auxiliary system. Many banks run COBOL on mainframes and haven't changed anything since the 70s.

I'm really confused as to the argument here. If you never update your software and don't expect to browse the Internet, old software works exactly like it did in the past. How do you expect it would change?

The whole issue is that some machines been force-upgraded to windows 10 despite users explicitly (or so they thought) responding multiple times that no, they do not wish to upgrade; and then failing to work correctly because of e.g. driver issues.

This is not acceptable behavior - I can understand automatic upgrades that are fully backwards compatible, e.g. a security patch; but upgrade to windows 10 isn't like that.

Upgrades aren't always entirely painless and without cost.

If the UI significantly changes, for instance, there will be a time investment needed to re-learn what you need to get work done. (This was a particularly nasty issue for Windows 8, not quite as bad for 10 but there are a few changes for sure.)

If the upgrade fails in any way, there may be significant time spent getting the computer back on-line. Sometimes this takes a very long time.

Software upgrades may actually introduce bugs. In the worst case scenario, some OS software upgrades have actually "bricked" devices (see: Apple iOS 9.3.2). These further impact productivity / take time to work around.

Some very old legacy programs may not work at all. It takes time to implement a new solution for the problem in a new tool. (This isn't a big issue for 7 to 10 from what I've heard, but you do hear this about other upgrades in the past.)

This cost in time is not nothing. If there is no real net benefit, it could end up being a big waste.

Even if people want to upgrade, IMHO the way Microsoft is pushing out the upgrade is quite bad. Usually in the enterprise world, IT departments tend to test new OS releases (and even patch updates) to make sure everything works, before rolling out the updates in the wild. They don't want a patch to break some mission critical software. The cost of that could be huge in some cases.

Even for Windows 7 retail to Windows 10 retail, I could see many people preferring to be cautious. Perhaps they are wanting to wait for a slightly more stable OS (usually the first release of a Windows version is a bit buggy). Some people also might want to test things out first, to make sure there are no problems. Of course there's those that don't want to update because change bad, but I don't think that's all cases.

Furthermore, a fair bit of Windows retail I'm sure is businesses - these tend to be even more cautious than overall consumers. A software update going wrong could cost them money, after all, and "if it ain't broke, don't fix it". I honestly don't think the "rolling updates" model being pushed works for enterprise... even small ones.

If the machine is on the Internet, it needs security updates. Let's take that as an inviolate law of nature:

If the machine is on the Internet, it needs security updates

This isn't a proprietary vs. open source thing, this is a "level of effort" thing. Microsoft right now, if they discover a security bug, has to create a patch for:

* Windows XP SP2 (thanks to all those "extended support" buyers; the US Navy has purchased support through June 8, 2017, they are not the only ones)

* Windows Vista (yup; also "extended support" for another year)

* Windows 7

* Windows 8

* Windows 8.1

* Windows 10

Why does that matter?

* It requires (up to) 6 times the development time to develop a fix for the bug

* It requires 6 times the QA time

* It makes Microsoft far less responsive to bug reports because of the time required to come up with a fix

* It makes Microsoft set the bar higher on what bugs are "fixable" because of the cost of coming up with a fix

It makes life difficult for everybody, and it wastes a ton of time Microsoft could instead be spending making their product better instead of constantly digging-around in 17-year-old code.

Let's say you're Ubuntu, you're all open source-y and have that warm fuzzy feeling. You want people to use their 10 year old OS installs, no problem. Let's be generous and only include the long-term support versions, because I want to save on typing:

* Dapper Drake

* Hardy Heron

* Lucid Lynx

* Precise Pangolin

* Trusty Tahr

* Xenial Xerus

That's also six versions you have to update every time a security fix comes in. (Potentially) 6 times the code to create the fix, definitely 6 times the QA time, etc. All the same problems Microsoft has, but this time in your open source dreamland.

That's a huge, huge, waste of effort that could go towards actually making the OS better.

You really don't understand Linux.

Reply on news.ycombinator.com