everything wrong with free software
"obedience breeds foolishness"
=> the-fsf-doesnt-care.html the-fsf-doesnt-care
=> how-to-deal-with-your-raspberry-spy.html how-to-deal-with-your-raspberry-spy
*originally posted:* mar 2021
i was too young to fully appreciate the 8-bit revolution, but i really wanted a commodore 64. i did learn to load programs from cassette, before learning to code on a pc not too long after that.
one thing from before my time that explains so many things about the present, is the teletype. screens used to be a lot more expensive, the teletype was a device that already existed, and it did not even require a computer to be useful.
if you want to imagine a teletype, its very easy to do if youve ever sat at a typewriter. just start typing on the paper. it doesnt matter what you type, the letters and words dont actually do anything.
now type in the letters "ls" and hit the carriage return-- and imagine that a list of recent things youve typed out on the typewriter starts being automatically written to the same paper, showing you the names of documents, the number of characters in each and the date they were most recently worked on.
this is what computing was often like in the 1960s and sometimes even in the 70s, where a typewriter was used instead of a screen. of course you would be constantly feeding in a new sheet of paper if this were an ordinary typewriter, but along with the ability to type automatically it also used a continuous roll of paper.
eventually trees rebelled and gave the humans a videoscreen replacement. the "glass teletype" when it came down far enough in price, enabled people to have the same text scrolling upwards as lines were added (one after the other) that was possible before with paper. when integrated circuits made it possible to put 64k of memory into a plastic-housed machine you could lift single-handedly, consumers were now able to convert their own television into a glass teletype-- and the home computing revolution was in full swing.
today, it is still possible to type "ls" and a get a listing of files. but in the 1980s, a far more compact and inexpensive operating system known as "dos" (for "disk operating system") was in common use. it was derived from tim pattersons "qdos", for "quick and dirty operating system"-- which seems to imply that msdos is a slow and dirty operating system.
one of the nice things about dos, and there were at least one or two of them, was that it was ridiculously easy to "install" the os.
dos required just three files on the computer:
you could even replace command.com with a different program, as its job was to accept typed commands and run them. in (much) later versions of dos, msdos.sys became a configuration file, and io.sys remained as a binary program.
so maybe youre wondering, as i did-- could you simply copy the files to the computer to install?
this wasnt possible, because at least one or two of the .sys files needed to be in just the right place on the drive-- not just the correct folder (in fact folders, or as they are broadly known among the more technical, "directories" were not introduced until dos 3.x-- despite having existed in unix for years) but the right physical position on the disk.
there was a place on the drive called the "master boot record", and when you started your computer it would look for the mbr, the mbr would tell the machine where to go for the first program to run, and an operating system would make itself the "first program" so it would be able to load and run other programs.
but even though most computer operators didnt know that, and even though you couldnt install dos just by copying three files, it was still very easy to install. you simply inserted the disk, started the computer, and typed the following command:
sys a: c:
you would wait a minute or two, depending on how fast your computer was-- and dos would copy those three files to the computer and write the mbr for you.
one command, wait a minute, and its installed. is it any wonder then that people get nostalgic for these things?
the system that results from such an installation is modest indeed-- you type a command, a program runs. you quit the program, you can run another program.
but this is the sort of modest foundation (not necessarily dos, specifically) that great and powerful and reliable systems are built on, and there is a great deal of confidence that comes from knowing that if you hose your system entirely, you can simply sys a: c: and then copy the rest of the contents to the drive and get back to work.
its not rocket science, i can promise you that. speaking personally, i suck at physics-- and math. seriously. this stuff is WAY easier!
nor did i set out to learn terms like "master boot record" or "msdos.sys"-- these are simply terms ive encountered many times over the years, you pick some up along the way and you look them up sometimes if youre curious. i never attended a class that asked people:
what is the portion of the drive dedicated to information needed to start the computer?
a. startup record
b. master boot record
c. primary boot partition
d. virtual partition
thats a boring question-- you should just look up terms if you need to, or if youre curious what they mean. satisfying such curiosity is its own reward, but as a bonus it can be very useful sometimes. but i stress the fact that i never took such a class; a test like that might have taken some of the fun out of learning it, and this is one of the really nice things about learning-by-doing or learning just for fun.
note that mbr is called gpt now; i think the "pt" stands for "partition table" but often just knowing the acronym is enough-- fortunately they only changed that once that i know of in the past 40 years. you can still use mbr (the technology, not just the name for it) if you really want to. gavin says that buzzwords are used in part to obscure "magic search terms" like "master boot record" that help you find useful information more efficiently, and i think hes probably onto something.
the industry loves new terms, it makes it sound like everything they offer is new when it mostly isnt. new books, new classes! old technology, mostly. they could have just called it mbr 2, but whatever. now someone who misses the point will tell me "gpt is COMPLETELY DIFFERENT from mbr! its 10 times as advanced and useful!" thats not education though, its marketing-- even when theres some truth to it. the differences are more useful to salespeople; the similarities are more useful to understanding and learning about technology.
i am both tolerant and guilty of shorthand that makes it faster to type things ("coding" vs "programming") but then it becomes useful to teach both the shorter term as well as the original.
if i taught a history of computing class, i would only grade people via oral exams. if they preferred to write something, that would be suitable. history has a lot to teach us about computing, but getting people to regurgitate certain facts would be wasteful-- you will gain far more from computer history by cultivating a general interest than trying to follow a syllabus.
=> the-wonders-of-modularity.html the-wonders-of-modularity