I am glad I came back and reread this -- I misunderstood it as thumb to middle finger at first, and thought it incredibly tiny.
Index to middle finger is 15cm for me, without stretching uncomfortably, and thumb to little finger ("pinky") is 26cm. Outside to outside of the fingers.
(For comparison I am a 1.88m -- 6'2" -- male.)
This is why I like large smartphones. The original tiny ones felt like using cocktail sticks as chopsticks to me: they were tiny and toylike and required superhuman precision.
6.5" phones are just about usable, especially with swipe-typing. Almost all Blackberry models were unusably small and cramped for me, but the Passport was OK.
(Copying my own comment from Lobsters, because I wondered why this didn't show up on here.)
It's interesting to me that this story hasn't had a lot of reaction.
It is the reason that Linux exists. It's the reason BSD came second. It's the reason for OS/2 and Windows 3 and Windows NT.
CP/M-86 was badly late: it followed over three years after the release of the 8086, the official successor to the 8080, the device that made Intel's fortune.
There was a single default OS for the 8080: DR's CP/M.
It was the industry standard, and the early business microcomputer industry ran on it and essentially nothing else. (But it needed at least a single floppy drive, which meant it was for £1000+ business computers, which in turn is why BASIC-in-ROM dominated in the inexpensive home-computer space: Commodore PET, Apple II, TRS-80, and their various 1980s successors.)
The Z80 was a tweaked 8080. Business was waiting on the 8080 successor that could handle more RAM -- which shipped as the 8086 in 1978.
DR didn't have an OS for it. It wasn't binary compatible: you couldn't run CP/M-80 on it, like you could with a Z80.
This article explains why DR didn't do that.
As a result, a cheap 3rd party compatible OS dominated, and that made the fortune of Microsoft, and that (alongside IBM's poor contract negotiation, which permitted MS to sell it to whoever it wanted) created the x86 PC industry.
DOS is why the 286 was a bit of a flop. That drove the 386.
(Aside: and, briefly, the Atari ST and Commodore Amiga -- and the Acorn Archimedes, whose CPU became the most popular best-selling CPU family in the history of humanity, and is why Intel is in financial difficulties now.)
The original all-32-bit 80386DX was too expensive. So, again, DOS drove the development of the cheaper 16-bit-bus 80386SX for cheap DOS 386s which didn't multitask and didn't run GUIs but ran DOS really well.
That market of cheap commodity 386 PCs is the soil in which Linux grew.
But in the late 1980s, the BSD folks were still focussed on minicomputers and things and they missed the PC gravy train.
286 protected mode gave you 16 MB of memory space, but it still had mandatory segments. Usable but it needed hard work and careful thought.
IIRC, that is why the 286 version of MCW Coherent OS still only allowed a maximum program size of 64 kB of code -- plus 64 kB of data. Very limiting. OK for early-Unix-style (say up to UNIX V6, of Lions Book fame and "you are not expected to understand this" comment.)
386 protected mode gave you 4GB of space, and you could choose the segment size, which meant everyone chose a segment size of ALL THE RAM PLS KTHXBYE. That meant you got flat addressing: all pointers could be flat 32-bit INTs. Simples!
The 386 also added the ability to seamlessly jump from 8086 mode to 386 mode and back.
The 286 intended that to be a one-way trip. DR and others found a loophole (IIRC the `LOADALL` instruction) and a way to exit 286 mode again but it forced a full CPU reset, so it was slow.
Not only did the 386 make it easy to get in and out of 8086 mode for DOS binaries, it natively supported multiple concurrent hardware-assisted 8086 VMs, plus memory management so you could keep most of DOS's 640kB free. So, win for protected-mode code, win for DOS code.
Intel aimed to please the DOS folks, and to alleviate jealousy of the 680x0 and its nice flat memory space, and make it easy to port UNIX and other memory-managed OSes over from minis and workstations.
TL;DR
The 286 protect mode was meant for 286 native OSes and it was not easy to run 8086 stuff, or change modes. Plus, segments, which everyone hated.
The 386 protect mode optionally got rid of all the stuff that people disliked about 16-bit x86 and offered alternatives that rivalled other platforms than x86.
Biggest change until x86-64 came along in 2001-2002 or so.
- In May 1990, Windows 3.0 running on a Intel 80386 could run multiple DOS VMs simultaneously
- In April 1992, OS/2 2.0 finally supported multiple DOS VMs. OS/2 1.x supported running only one DOS app at a time.
While MS Excel had been available on Windows for many years before Win3.0, Word for Windows had only been released a year before Win3.0.
So many PC users would still have been using the DOS version of WordPerfect (and maybe the DOS version of Lotus 123). Windows 3.0 allowed the user to run many of their DOS apps at the same time, which DOS and OS/2 couldn't do.
There were at least 2 separate and independent TV series, but I highly recommend going to the original: the 2 later novels from Douglas Adams, after he wrote the Hitchhiker's Guide to the Galaxy.
In the novels, a stuck sofa is revealed to have got there because a retired Time Lord, Professor Urban Chronotis, briefly materialised his TARDIS in such a place it provided a doorway. So the books are extremely loosely tied into the Dr Who universe.
I would love an AmigaOS desktop environment to run under AmiWM to make a Linux box feel like a Miggy. I'm puzzled it's clever been done.
reply