Hacker News Reader

Comment by nlh | Hacker News Reader

I think everyone who spent time computing in the 80s & 90s has at least one tale of woe relating to forgetting to save or crashing before a save or tripping over a power plug, etc.

In retrospect it seems pretty wild that users were expected to actuate an explicit “save” command and that word processors didn’t just handle this automatically.

I’m sure there were real reasons - I was never involved in DOS or Windows programming but I presume it had to do with the slowness of saving to disk (and that a background auto-save wasn’t technically possible?). Or did we just not yet collectively have enough experience to know that auto saving was something critical to writers…?

View on HN

Imagine the entire machine coming to a halt, ignoring input, and making grinding noises for 5-10 seconds. That's what saving a small file was like in the floppy disk era. If that was happening randomly while you're trying to type something, you'd go crazy.

Makes sense and I do remember that!

Apart from the lack of sophistication regarding background saving, the two main reasons were

a) writing to disk was inherently dangerous: file systems of the time were not robust, so anything going wrong while updating metadata (in particular) could trash not just your file but the whole filesystem. Floppies were slow and machines were crashier, making the whole operation riskier; and

b) using floppies reduces their lifespan because the disk physically rubs against both the read/write heads and the internal padding of its casing, so typically they were only accessed as part of an explicit user request.

Saving did take time, but multitasking was a rarity, as was multithreading. Machines only had so much spare cpu and ram, too.

I'd expect autosave to prevent typing, which would bogart the workflow. I can envision doing quick incremental saves, maybe, but then you'd probably get jittery behaviour.

If we're talking mid 80s, something like the c64 would need assembly to handle anything like that I'd think.

I remember writing a BBS in basic, and just typing for an afternoon and running out of ram. I had to create separate modules for the up/download area, the message board, file transfer, and so on, all loaded off of... there was a commodore 1M floppy drive I used.

Given that Mr. Carter received a special disk with which he retrieved the lost material, it seems more like he accidentally deleted it, rather than just failing to save it.

From my memory that would have been intrusive to the user. Saving a file meant the computer was saving the file and doing nothing else.

You have inspired more questions about incremental saves: speed was an obvious issue with personal computers, but was it implemented on any minis or mainframes? These were multiuser systems (i.e. multitasking), so they could do background disk I/O. Also, would it be a possibility on some personal computers? For example: 8-bit Commodore computers had their own CPU and could operate indepedently from the computer, while the 6809 was famous for being sophisticated enough to support multitasking (e.g. OS-9). Even less sophistcated machines could do I/O in the background if the disk controller signalled the completion of a read/write operation via interrupts.

I'm asking because it is also possible that such features weren't implemented because the weren't thought of or because the defied contemporary expectations.

Reply on news.ycombinator.com