You youngsters!
By 1970 I had a remote access account to Dartmouth's Itty Bitty mainframe (around age 12-13).
The next year I had local access to a DEC pdp-8e, on which I learned ASM, FORTRAN, COBOL, FOCAL, and a couple other BASIC variants. Plus certain manual toggle entry code operations, useful later for setup of certain high end ESDI and SCSI controllers or video cards.
Before most of those toy computers mentioned here existed, I had a cp/m machine at home, and bought full doc's and upgraded BIOS from Micro-Cornucopia (other than National et al chip data books on my generic reference shelves), and as newer chips came out, doubled its clock speed with a PCB trace hack, and faster Zilog Z-80b CPU. That had a pair of Tandon TM100-2 5.25" floppies, each then worth more than a several TB HDD or basic SSD now. Some larger Japanese crooks ultimately bankrupted Tandon, with blatant Patent thefts, and far bigger budgets for crooked lawyers.
Next progression was getting online in the old days of ARPAnet, Usenet, BITNet, and comix characters named services. Unix was still AT&T's toll center toy, other than sharing with NASA (where some friends worked) to enable space flights on low memory systems re: workloads, and before the advent of Penguin stuffed animals at the better tech's workstations. As a gateway for better access than CI$ and before AOHell's rubber room for Trash-80 types who didn't realize how horrendous a real Digital Research cp/m manual (set) was (but it wasn't crippled to simplify options for low-tech users), I got a Racal Vadic 3400 modem (1200 bps, predating Bell 212), when others were just getting Bell 103 protocol (the last step where Baud = bps rates, but so blazing fast thanks to error correcting block sends of X-modem from Ward Christensen, Bellcore/WE "bandwidth must be 16 times Baud min for reliable data" standards were broken. That gateway access to the outer world was hosted on a VAX 11-780.
Over the years I've been in an underground bunker based Lucent IV-ESS tandem toll center like what we blew up in Bhagdad, and a telco DSC opposite the state's NOC primary of two "war room", with access to every SONET fiber path in the state, linking then still not independently homed data backbones, major customers, major data hub loop, and regular telco CO (mostly there Lucent V-ESS's, that work better than Nortel's or some others).
Yesterday I was doing some casual reading (gach!), of ITU, IETF, and CableLabs standards documents, trying to sort out some details of QAM-128 to 4096 Codeword Constellations coexisting for DOCSIS 3.0 to 3.1 transitions, with some suspicion that two major CATV's are scamming customers a bit on key details. Somewhere along the way there, I've done a few odd projects to help "build this city" of tech no one likely to be able to read this hasn't used. Or called out a few screwups, like when USR Courier firmware got miscoded adding CID and Distinctive Ringing features, so they had a 7 second dead period and caused call collisions (and ended up consulting with their Senior Firmware engineering team leader Joe F. after breaking through an obnoxious corporate blowoff structure, that wouldn't exist absent the 90% of support calls from incompetent, lazy, and crazy users previously mentioned, in a roughly 50-30-10% breakout). But, not a lot of users can identify timing windows of SS-7 and POTS interactions with high end modem firmware, certain common models of CO switches, and system tools, even among bright pro's, like the world leading engineers I had to ridicule as the "Useless Robots of Skokie" to break through business management and "customer service" abuses and get to pay attention.
But somehow I end up here playing some silly game? ;-)
As to DOS history, and hint, newer end of its life cycle actually overlapping OS/2 (IBM designed the linked files MS sort of stole) and OS9 (microcontroller oriented, not Apple), who recalls which versions were illegal for Gates to sell (and why?), or the subversions after MS lost big time in court, but extorted the victim to sell out for a third the court awarded damages rather than be buried in litigation costs and time delays?