I started my career working on paging transmitters. Some of the Chinese customers were still using 25-year-old transmitters and we were still making the same model beause they didn't like retraining their techs. It made no business sense for us, which is one reason that that company is out of business. The circuit boards had to be hand-assembled with through-hole parts. There was an ID code for each transmitter and instead of being stored in a ROM you had to program in the code by cutting out diodes with a pair of dykes. I had to redesign some analog filters for this transmitter when parts went obsolete. This was in the day when a 3KHz-cutoff filter would normally be done in a DSP but I was designing analog active Butterworth filters. Anyway, that was a product that had a long lifetime, and maybe long after I "updated" some of the parts, though by then the era of pagers was fading out.
Aerospace designs often have a long life. I think they were using ferrite-core RAM in the shuttle for a long time.
"And what's the longest-living product that you have ever helped design?"
I got my first training as a radio tech in the grandly named British South Africa Police radio section in the then Rhodesia. We got some new 40 MHz AM vehicle 2-way radios from a company called Emcom in New Zealand. The modulation stage (for AM so it was just an audio amp with power output comparable with the RF output) had a pair of power transistors driving a centre-tapped transformer. There was a fairly simple line-up procedure to set and balance the quiescent currents in these transistors, but it involved getting your meter probes into two tiny holes in a 9-pin socket in the corner of the (very crowded) board. So I made a tester with two small and old but identical meters that measured the currents simultaneously, and while I was about it put in a switch to measure some other parameters - RSSI, battery voltage and the like, that were on other pins. It connected to the radio via a 9-pin plug on a cable, with a long bolt as a handle. I proudly labelled it "ASH'S EMCOM TESTER". It was incredibly useful and was always getting borrowed by other techs.
Over ten years later I went back there to visit a friend and there on a bench was my Emcom tester. My friend said there were still a few Emcoms around and the tester was still used regularly. I guess ten years is not a lot but for a quick and dirty tester it's pretty good.
If you really need a formatted document to last a long time, use the open-source text formatter LaTeX, and don't use a word processor. LaTeX (created by Leslie Lamport) is built on TeX (created by Donald Knuth to format his books "The Art of Computer Programming" so his books would never go obsolete). These are standard in mathematics, physics, and engineering for writing journal articles, and are specifically intended to never change. You write source code in plain text, and then run LaTeX to format your paper. Keep both source and formatted versions.
Decades ago, I got in the habit of keeping comprehensive unofficial informal plain-text diary files, indexed via HTML files. These diary files are not laboratory or project notebooks, and are not offical formal documentation either. Notebooks are intended to be permanently unalterable once written, for future use in patent lawsuits. My diary files are editable anytime by me, and record everything I think about for a project. If it's worth scrawling on the back of an envelope, it's worth writing up in a diary file. I number the diary files sequentially, with an unchanging three-letter prefix and four digits of numbers incrementing for each new diary file. One HTML index file links to the diary files sequentially by number. Each project gets its own HTML file linking to diary and other files relevant to particular projects. Result: instant internal HTML site of all my work for the company. I put all the files into one directory or folder. The files are organized through the HTML files, not through a directory structure. This means that one file can appear in more than one place in the organization as seen by a web browser. No subdirectories means no searching subdirectories for files. When I need to copy a file, I use the web browser to verify the filename and the command-like copy command to do the copying. In my last twelve years of working, I created 2638 diary text files at work. Since retiring in March of 2010, I have created 1459 text files at home so far.
I've created some products that lastet much longer than initially expected:
- An individual industrial automation project that was operated from 1991 to 2008 (when the business unit, excluding this installation) was sold.The installation itself was built around 1977, so some parts were 30+ years when it went out of operation.
- Some measurement unit was originally developed for a single project - 10 units (plus 2 spares).It ended up as the only system applicable for the task - resulting in 3000+ units sold during 20+ years. Plus some supporting products that sold in the hundredsWhen the microcontroller used was no longer available, my ex-employer decided to go for something "totally new" (and a bit overengineered). We'll see whether this new system will also have a production life of 20+ years.
- Another system (same microcontroller) shared about the same fate: 20+ years of production.They went for something "super-duper" - until now nothing is available :)
- There are more examples of this kind:Do it right the first time and start to wonder: one solution I kindly considered "inferior" is regarded the "reference solution" these days :(
The company I work for used to have hard-copy project notebooks. They've all been scanned to pdf. I'm not sure that this is an improvement. And I sincerely hope that the old paper copies are securly stored somewhere....
The newer scans with OCR aren't too bad but the older ones are impossible to search.
Agree 100%. Historians fear that there's going to be a "Black Hole" in histrory from about 1950 onwards, because it will be easier to read Sumerian business records from 3,000 years ago than the numerous proprietary and obscure document formats from then onwards.
It might also be prudent to ensure that there are dead-tree versions available, preferably either loose-leaf or continuous. They may be miserable for reference,, but at least OCR is reasonably reliable.
Save long-lasting documents in plain-text pure-ASCII form, or maybe UTF-8 extension of ASCII. Avoid proprietary character sets. Microsoft programs are liable to save-as-plain-text in a proprietary Microsoft variation of ASCII, with its own special versions of punctuation marks.
I knew a guy who worked for Siemens who knew the old ISDX PABX's backwards. We had one at work and he provided support long past when they stopped selling them. He also eventually retired and I think Siemens (with some notice) advised remaining users that support would be ceasing. It was a lovely PABX to work on and I shed a tear when we finally switched ours off.
I've had similar experiences sorting out a serial cable problem on a leather cutting machine. Several others had tried and failed, but I sorted it out in 20 minutes with an RS232 breakout box and a soldering iron. Everyone just thinks USB these days.
I wish I had been a COBOL programmer. They still get good money.
Yes, we still get the print edition. After the change in productin locations, it took us dozens of calls over several weeks to cuustmer service to get the carrier to properly deliver that paper. Finally, the message got through to the carrier. Once the papers were delivered properly for a week, I left the carrier a little incentive to keep doing so.