I collect old devices, but almost never use them for work, so I treat philosophers philosophically about dead software stores, expired domains of vital services, broken links to firmware and other amenities of a museum worker’s life: he wanted it. But still ... On Reddit there is a community of buyers of things "for life", so that they serve as long as possible. There they discuss cookware, combs, boots, lawn mowers, boilers, dumbbells and audio equipment. But computers rarely discuss, and for obvious reasons. Is it possible to create computer equipment using modern technologies, with a life span of tens of years? I will try to answer this question as far as possible at the end of the post, but in general I want to talk about how modern devices degrade over time. For example, electronic watches, smartphones and computers of different years.
I keep a diary of a collector of old pieces of iron in a Telegram . The construction of the 386th computer is resuming; the next step is the difficult choice of a sound card.
Pebble
I collected a collection of six different models of Pebble smart watches last year , and I still use them. The first model of this manufacturer was released in 2012, and in December 2016 the company ceased to exist. Intellectual property was transferred to FitBit (which itself was recently sold by Google), device support was discontinued. At the time of the preparation of Pebble material in August last year, the situation with their performance was generally positive, but only because the community of developers and users of Pebble watches organized and created a mirror of the server infrastructure. Without a Rebble project, it would not even be possible to activate a clock after a reset. The latest versions of the Pebble application for iOS and Android have been adapted to work with this infrastructure: they managed to save the store of applications and dials, weather data transmission and even voice recognition of text in Pebble Time models. The only problem for the summer of last year was the shutdown of the Pebble website and forum: if something was wrong with your watch, before that it was possible to find a solution in the search using the error code. Along with the site, search results disappeared.
But that was only the beginning of the slow degradation process, which continues to this day. In the spring of 2019, the Pebble app disappeared from the Google Play Store. Google at the beginning of the year changed the rules by which applications on the Play Store receive permission to access phone calls and SMS. Under the new rules, only the application installed for working with SMS or default calls can access such data. If the app, in principle, does not have such an opportunity, that is, it is not a dialer or SMS chat, then it is removed from the Play Store for violation of the rules. Google can understand: many malicious applications used access to SMS. Pebble application requires this system permission not only to display notifications, but also to reset a call or reply to SMS with pre-prepared text. The current owner, of course, did not update the software, and it is not a fact that there was such an opportunity.
Okay, the life of Android phone owners who stubbornly continue to use the watch of a nonexistent company has become a bit complicated. Much more complicated is the life of those who use the iPhone, but not for long. In early August, the Pebble app disappeared from the Apple App Store. Since you can’t just download and download “APK from the Internet” in iOS, if you lose data or change your phone, it would turn out that your watch no longer works. Fortunately, in September, the application, through negotiations, was returned . It’s not a fact that it’s forever, so if you suddenly want to buy Pebble retro-clocks in the future, you’d better install the application right now. It doesn’t hurt.
And so it became interesting to me: what will make my Pebble watch unusable earlier - battery degradation (solved) or problems with software? Everything works on my current Android 9 smartphone. On Android 10, there are some problems , but it seems to be not fatal. This will not always be the case: the next major upgrade of the API can lead either to inoperability of the software last updated at the beginning of 2017, or to the inability to intercept notifications. I will replace the battery in the watch, if anything, and using several copies of Pebble in turn will most likely help to extend the service life. I will assume that at some point my small collection of watches will become another museum exhibit, and it will be stored complete with a smartphone with which these watches are capable of working. Perhaps they can help extend their lives to alternative smartphone applications.
The fate of Pebble watches will sooner or later be shared by many modern devices that simply do not work without a server infrastructure. Smart watches are not so bad, but what about smart home systems? In 2014, at that time, a separate Nest division of Google acquired Revolv, a manufacturer of home automation hubs. Sales of devices were discontinued almost immediately, and in 2016 the cloud service was turned off, which already bought and built the hub into their own smart home brought a lot of problems. The victims could return the cost of the hub, but the point here is not the price of the hardware, but the installation cost, and the simple fact that something worked for you at home, and then it stopped. In no case do I consider myself a luddite, but aren't we too often for the sake of progress forced to abandon the outdated and crooked technologies that we are used to and that work?
Smartphones
This is a topic worthy of a separate post, or multivolume about buried under millions of lines of code and business solutions by old devices. Let's start with the good news: thanks to the relatively conservative mobile operators, you can remove the SIM card from your fresh iPhone or Samsung, insert it (with an adapter for a different size) into your 2005 smartphone , or even your 2001 mobile phone , and it will work. A relatively new phenomenon called eSIM can break this compatibility: the subscriber identifier will become virtual, and there will simply be nothing to insert into the old device. I’ll guess on coffee grounds: I have ten more compatibility in my stock for ten years, and even more.
The good news ends here. The most recent and most characteristic example of unplanned obsolescence is the Windows Phone platform. It existed since 2010, in 2017 it was quietly closed due to uselessness and lack of interest. On December 10, 2019, support for all devices on Windows Phone will be discontinued: which means that there is not much time left to buy a phone on this platform in the collection. I don’t have such devices yet, so I can’t assess the damage from the final closure of the platform, but I’ll assume that at least the application store will suffer, the maximum is the activation algorithm. In the best case, you will receive a phone without the ability to download new software (without dancing with a tambourine, which are always possible, but not always justified), in the worst, an unworkable brick. Most likely, the infrastructure will be disconnected not in December this year, but later, and in the near future any development and delivery of updates to already activated devices will simply end.
Okay, let's take a live platform, Android. I recently got the first version from the Samsung Galaxy Note bins: released in 2011, acquired by me in 2012. Seven years is not such a long time. The original firmware of this device is based on Android version 2.3.6. The last official one is 4.1.2, both versions are no longer supported. Unofficial builds are available up to Android 9.0 , but I used the maximum Android 7, and with it this dual-core smartphone with a gigabyte of memory works well, very leisurely. I installed the original firmware on it, and immediately appreciated the difference: the phone just “flew”, but most of the applications from the store refused to install. At a certain point, I stopped trying to pull the freshest software on this old man: neatly selected old versions of the software on the old firmware are much nicer than the modern software obviously designed for more cores and more gigabytes. Even updates to Google services, which can still be rolled on Android 4-5, significantly slow down such devices. But in such a scenario, this is again a museum device, which seems to be able to digest modern network services, but actually not. Although Android did not close anyone, and this is by far the most popular mobile platform, older versions of the OS will not differ much from Windows Phone in the near future.
Unwind a few years ago. My latest purchase is the Nokia E90, a 2007 monstrous keyboard communicator. How to determine if your piece of hardware meets the criteria for a retro device? Google the model name. If the first search result is a Wikipedia page, then yes, it is and matches. The device runs on Symbian OS 9.2, and since version 9.1 Nokia has introduced the practice of mandatory digital signature of applications. As a result, you can install a lot of useful utilities, as well as get direct access to the file system of the device only through a kind of jailbreak - although this term was not used in Symbian times. You can’t deliver officially distributed digitally signed applications either, because the certificates have expired, and there is no one else to update them. The clock helps, and you need to guess the correct time interval, not earlier and not later. Although smartphones are still being released under the Nokia brand, they have almost nothing to do with those old devices (and the Finnish company itself), and support for earlier devices has been discontinued. Nokia E90 compares favorably with other models by the presence of a small group of users who are so used to the device that they are not ready to part with it even 12 years after release. Therefore, there are solutions for renewing certificates for the normal (conditionally, of course) work of mail and the web. Otherwise, Everything is Very Bad: the Ovi application store is dead, the navigation does not work (you can screw offline maps of not the first freshness), there are no solutions for modern instant messengers.
The general conclusion about smartphones: they are not so seriously tied to the server side as IoT, and in general they do not turn into a brick, while maintaining at least telephone and minimal network functionality. Although options are possible, for example, here is the latest news about relatively old Apple devices. Some models (iPhone 5, iPad 4) needed to be updated via the network before November 3 to fix a bug in the GPS-receiver. More precisely, a flaw is present in the GPS standard itself, which uses a week counter with a maximum value of 1024. The first 1024 weeks in the global positioning system were counted from 1980 and ended in 1999. In April 2019, another batch of weeks ended, but this jamb caught up with old iOS devices only now. If you do not have the firmware installed with the patch, there may be problems with navigation or even access to iCloud, as time synchronization is also tied to geolocation, and it seems to work correctly with the cloud service. The fact that Apple took care of users of such old (2012) devices is good news.
The example with certificates in Symbian shows how security tools can become a rake for old-user users over time. Imagine what will happen to multiple protection systems in the same iOS in 10 years, when a similar bug with GPS slingshots in the code will work. Here you can’t do the firmware already. It is interesting what in the 30s of the twenty-first century there will be means of maintaining the operability of now still old devices.
Computers
In January 2020, support for Windows 7 officially ends , so I'll start with this “new retro”. My laptop Lenovo ThinkPad X220 in the seventh year of life successfully continues to work around the clock in the role of a home server on Linux. Initially, Windows 7 was preinstalled on it, but its specifications are sufficient for Windows 10, unless the screen resolution is “a bit big” by today's standards. Lenovo branded software suite is no longer supported - if earlier it was possible to download updates and drivers directly from the manufacturer’s server through it, now it would have to be done manually. The completion of OS support after three months will primarily lead to the termination of the supply of updates and patches, but will not affect the performance. Over time, Windows 7 will become the same unsafe OS for everyday use, which is now Windows XP - new vulnerabilities appear, and patches for them no longer come out. After another couple of years, software will start to break down: I fully felt it, trying to install Windows Vista on my semi-retro desktop . The next stage of degradation: the lack of drivers for the new hardware. For example, for GeForce 10xx video cards, there is no official support for Windows XP, it ended in 2016 before the release of this series.
That is, 15 years after the release of Windows XP itself. By the standards of smartphones, and even more so modern IoT devices - it's an eternity! What will happen with Windows 7 even later, you can see the example of Windows 98, the first Microsoft operating system, web-oriented. The modern web in Internet Explorer 4 simply does not work. All links from multimedia add-ons over the system are dead. A WebTV component that works in conjunction with a TV tuner and downloads a program guide from the network does not work. The page with the description of this program does not work either. Updates for Windows 98 can still be found on the Microsoft website, but in the form of an unstructured dump of artifacts. Technical support pages are also deleted, but the web archive saves. On a computer twenty years ago, you will have problems with basic network standards, such as WPA2 for WiFi (but nothing is impossible). Best of all is the good old-wired Ethernet and old-school network protocols: you can file a file ball and ping Yandex not only from Windows, but even from DOS on a computer forty years ago !
And here I got to the most beloved period in the history of IT: computers based on 386 and 486 processors, and the corresponding software, for example Windows 3.1 and 95. It was an era when, for a relatively short time, a personal computer was truly personal. It was preceded by the time of mainframes and generally inaccessible computers of gigantic proportions. After it, to this day, the time of cloud and Internet services continues, when our data is scattered on hundreds and thousands of servers around the world. But in the period from the late eighties to the mid-nineties, both the computer, and software, and data were stored in your home, in the form in which it is convenient for you. It is not easy to find working hardware and programs thirty years ago, but if it is found and started, it will work. Enough for my age. The beauty of those times was that the Internet (or Fidonet, or another network) was important, but not required, your personal computer was not tied to it in any way. They don’t do that anymore.
Buy it for life
I will not even try to discuss the current situation from the point of view of programming approaches or methods of conducting a modern IT business. I will be glad to discuss this topic in the comments. I will say this: if innovation requires the abandonment of Legacy, then go ahead, ruin the hell with him. I am not a supporter of a position like "Symbian was not so bad, you just had to finish it." Maybe so, but this hardware and software design has already collapsed, it was not supported by either the business or the developers themselves. Even the short-lived Maemo platform has survived the test of time better, although it was also a dead end branch of development. The more disruption there will be in IT, the more new-old devices will replenish my collection, so I do not complain.
But I hope that in high technology there is enough dough and knowledge to satisfy a very small (possibly only from me) demand for products with a particularly long lifespan. Imagine a phone with guaranteed support for at least ten years. My Samsung Galaxy Note is approaching this milestone, and enthusiasts are still sawing firmware for it. But what if someone does it in a more organized way? With a guarantee of quality. With the promise of not turning the interface upside down every year. With a small but useful set of accessories such as the same smartwatch. I have no doubt that such a solution will be expensive: we are now selling smartphones for $ 1000 with a two-year update guarantee, and here we will have to pay the labor of developers, admins, technical support five times longer.
This is also an interesting technical task: how to develop software so that its support does not drag the manufacturer to the bottom for many years? How not to devour all available memory, even a gigabyte, even ten? Is such developer minimalism justified, almost like in the good old days, when the program had to be shoved into 48 kilobytes of memory? Is anyone interested in such an approach? Is it possible with the prevailing methods of teamwork? Is it possible to integrate such a device into an existing network services infrastructure? I have no answers to these questions, but it seems to me a personal device with a rather limited, but still wide functionality, in which some basic things such as a calendar or messenger are brought to the ideal. With applications for doing business, reading news, working with mail, which are really convenient, about which books with recommendations have been written, and these books do not expire after six months.
Ahem. It seems I will rather wait for a direct interface between the computer and the brain. I know only one device that is as close as possible to the ideal I described - this is the Psion 5MX , a device 20 years ago with an incredibly convenient software part. The IT industry has not yet settled to such an extent that the approach I described has become widespread. The complexity of a modern smartphone or smartwatch is incredible - no 486th computer from the 90s can do that. Cool, but sometimes I get the feeling that all this triumph of technology is stuck in the alpha version for a long time, but I want to see a beta, or even Release Candidate. Maybe it makes sense to stop and think about what will happen to this code or that hardware in 15 years? Is it possible to make it continue to work? Not even so that it really functions in 2035, but to create the right atmosphere. It seems to me that two options are possible. Or some basic modern designs will settle down and stop changing at breakneck speed. Or there will be another revolution, which will send all our smartphones, watches, smart homes, email, video streaming and payment systems to the dustbin of history, and restart the squirrel wheel. I continue to observe and slowly build my 386th computer. If I get tired of a series of innovations, I will include it, there is the Prince of Persia, stability and peace.