4 ms·
A history of storage media (2017)
- ghaff 4y agoVery cool. I did a presentation with a colleague in a mid 90s context which was mostly focused on disk drives. Though we got into some earlier parts as well with the help of the MIT archive. It made you realize how the categories blurred. I started to put together a video at one point but only got part-way there.
- jl6 4y agoOne storage media topic I would like to hear more about is whether the stagnation in optical disc technology reflects a natural limit of the concept or a victim of market forces moving the most profitable use cases towards streaming.
- haimez 4y agoIf rumors are to be believed, S3 glacier might be an interesting data point
- ghaff 4y agoI've heard a variety of rumours about Glacier but I'm inclined to take the Occam's Razor answer of powered-off hard disk drives.
- aidenn0 4y agoWithout the content owners driving volume, it's unlikely to see a cheap optical option.
- layer8 4y agoI believe it’s a combination of the cloud and streaming being more profitable and also more convenient for the average consumer. Personally I’d be thrilled if there were petabyte data crystals. :) Somewhat related: https://en.wikipedia.org/wiki/5D_optical_data_storage https://en.wikipedia.org/wiki/5D_optical_data_storage
- Izkata 4y agoSearching for images of what that might look like are extremely disappointing. I was hoping for something sci-fi like a pyramid-shaped crystal.
- ianburrell 4y agoBluray uses violet laser. Higher density requires using a smaller wavelength laser. Up until recently, ultraviolet lasers were expensive but it looks like someone recently developed cheaper one. It might now be possible to make higher density disc.
- Dylan16807 4y agoIt's not all about frequency. With more effort we might have been able to get significantly more layers working.
- ggm 4y agoPeople modelled delay lines in acoustics as well as mercury. there's a story about how many clap echoes they could get stable.
- kbrannigan 4y agoWhile on the topic of storage media. Will knowledge be lost at some point? Terabytes of data forever locked away in encrypted devices. Maybe they'll only dig out small rectangles, mysterious bricks of plastic and glass. Since the younger generation is losing the concept of physical storage. Previous generation grew up with: cassettes, records, cd, dvd and usb drives. They had wired phones, wired everything. The current generation is growing up with: streaming and wireless transfers(airdrop).
- marginalia_nu 4y agoAlmost certainly. You don't need encryption or even particularly long time to lose knowledge. Stuff is falling off the Internet every day that isn't backed up or archived. Wayback Machine only captures some of it. The rest is gone.
- Dalewyn 4y agoPhysical storage doesn't last on its own volition either. Bit rot will strike at any binary storage medium eventually, certain form factors will literally lose the ability to be read eventually, file system support might be dropped in future OS versions, and so on. The only forms of data storage that have been demonstrated to last the tests of time are physical texts and images etched into cold hard stone or metal. That's it. Papers rot, inks rot, bits rot, plastics rot, magnets fade. You want something to last? You get a big chunk of metal or rock and etch your data in it. The thing will outlast any other form of data storage by at least several millenia.
- systemskid 4y agoProject silica by Microsoft Research seems like a very interesting project in this direction. Storing data on glass which is incapable of bit rot (and in more ways similar to the stone carvings from the medieval era)
- dbcurtis 4y agoOne of my profs had worked on CPUs with Williams tube storage, then mercury delay line memory. He later did research on core, and one of his PhD students from that era was my first boss. When I knew him he was chair of computer engineering and his PhD students were working on bubble memory. An amazing research career that almost defines the history of computer memory.
- gumby 4y agoWhen I was an undergraduate at MIT I learned of a prof who’d done interesting work in urban planning named Jay Forrester. This stuck in my mind because I already knew of the core memory inventor Jay Forrester. But you know, names are not unique, even in a small institution the size of MIT. Then a year or so later I learned that an MIT prof had come come up with an interesting theorem we used on solving, IIRC, large sets of simultaneous equations, a guy named Jay Forrester. OK, surely there couldn’t be three of them, and indeed there was only one. To my regret I was too intimidated to go meet him.