I ordered a 3TB hard disk yesterday for the princely sum of $130.00.
I remember in the early ‘80s when we got the first PCs at work, they had 10MB hard disks. At that time, the hard disk alone cost about $1000.00. I also remember wondering “what in the world would a person ever use 10MB of disk space for?”
So the 3TB disk is 300,000 times larger than that 10MB unit.
And the unit cost has gone from $100.00/MB to about $0.000043/MB. In other words, the 10MB hard disk was about 2.3 million times more expensive.
Due to inflation, $1000 in 1983 dollars would be worth $2348 today. So that 10MB disk was actually not quite a million times as expensive.
Surprisingly, somehow, this does not make me feel old. How many other things have improved by a factor of one million in three decades?
Haha, yeah those aren't big enough yet, this is for backups so raw capacity wins. These hybrid drives also look interesting, have any experience with those?
In the late 90s I worked out that a device with the capacity for a lifetime's music should be available before 2020.
Mp3s go at about 1MB a minute, so that's about 40TB. If you filled a 40TB media player with music then you could spend your entire life constantly listening to music (including sleeping time) without repeating a track.
CowJam:
In the late 90s I worked out that a device with the capacity for a lifetime's music should be available before 2020.
Mp3s go at about 1MB a minute, so that's about 40TB. If you filled a 40TB media player with music then you could spend your entire life constantly listening to music (including sleeping time) without repeating a track.
Nice. Sounds like that prediction is pretty much on track.
Depends what compression is used. I use 256K bit, songs take about 8-9 Mbyte.
128K only sounds okay when listening in a car when road/tire/engine/airstream noise kills fidelity.
The new Samsung laptop I just bougth has 1 TB of storage - body is very thin, no CD/DVD, don't think it has a hard drive.
Taking it back tonight - I don't like the offset keyboard (from the number pad) and I really don't like the chiclet style keyboard, just awful for typing on; and the front edge has a sharp edge where the body parts meet. Is nice & quiet tho. Haven't loaded any software (like the IDE or Eagle), the i7 processor seems to be pretty quick. Installed chrome - happened so fast I didn't think anything had happened initially! Made me wonder if it was preinstalled and then just enabled.
CrossRoads:
Installed chrome - happened so fast I didn't think anything had happened initially! Made me wonder if it was preinstalled and then just enabled.
It ought to be. What's the use of a PC without a Web browser?
CrossRoads:
The new Samsung laptop I just bougth has 1 TB of storage
My latest company issue standard laptop has 1/4Tb of SSD and 1TB of hard disk (and a mere 8GB RAM - I really must insist on upgrade to 32Gb )
Back in Memory Lane, I do remember particularily a local pc store announcing (a bit longer ago than yesterday) "Never run out of disk space - we have 800Mb on a single drive!". Yes, there were bigger disks, but this was in the affordable range for a household pc. That space did last a long time. (I upgraded from an 80Mb.) Backup to floppy was no longer an option - I had to invest in a tape drive (using those DC-cartridges, mini sized)
In a story from the 1920s I read, a father explains to his children that "Everyone is talking about a Millon this - a million that. Can you visualize a million? Here is an aid. Get a stack of millimeter paper. (ie with milimeter squares on it). Cut and place them side by side on a board that is a meter on each side. Now you can see 1 million square millimeters. To get an idea, you can colour each carefully. See how long it takes to do one thousand. And how little that is of 1 million."
I am trying to extrapolate this to 1 Giga and 1 Terra. That is 1 million of theses 1 million sheets. A piece of paper is (typically) 0.1 mm thick so we need one hundred meter long stack of these square meters sheets to get to one Terra.
POP!
The DNA inside a single (human) cell is less than a CD, by the way.
I think data management is going to be an increasingly big problem. Say you have 10 files (like I used to, 35 years ago). You can easily keep track of those 10 files.
Then as disk space increases you end up with thousands of files. And you get a bigger/newer PC. So to save trouble, you just copy all the contents of the old PC onto a (fraction of) the new PC's hard disk. And you do that again every 5 years or so.
Before you know it you have millions of files, some you want to keep, others you don't, and a lot of duplicates.
And then you worry about "which is the latest?". Good version control can help with source files, because before that you would tend to make a copy "just in case" and then in 5 years wonder which was the original and which was the copy.
The one he shows is somewhat smaller than the ones I used to work with. They were around the size of a washing machine and stood on the floor. I can't offhand find a picture that does it justice, but I would estimate that the removable platters were around 1 foot deep and maybe 1.5 feet across. There was a removable cover that you twirled around it to "grip" the platters, and then you lifted it out of the drive mechanism and stored it, and put another one in its place.
From memory an entire hard drive held the compiler, which you had to physically install into the device, then copy the compiler into memory, then you would power down the drive, remove the platters, and then go get the other platters which held the source code. Using the compiler now in memory you would compile the source, store the object back onto the second disk, and then remove that. Then you would get a third drive onto which you copied the compiled object code. This third drive held the "production" object files.
The whole process took maybe 15 to 30 minutes. That's to compile one program.
Actually this page shows the general idea:
Since you unscrewed the plastic cover during installation, the platters were exposed to the open air.
The IBM 3390s were a very successful line of drives, very reliable. I seriously doubt they were the world's most expensive but I'll allow Dave some artistic license there. IIRC there were four spindles per cabinet. I don't remember the part about Halon being pumped into the disk assembly. We had Halon-protected machine rooms but I think he's wrong about it being connected directly to the equipment, it was actually fairly similar to overhead sprinkler systems, just using a gas instead of water.
Some of the reliability was due to the fact that the 3390 disk packs were not removable. Previous generations had disk packs with platters that looked very similar, but that could be moved from one drive to another. This led to huge disasters: If a drive experienced a head crash, it would of course ruin the disk pack as well as the heads (which were part of the drive). This could be quite a spectacular thing, a very bad noise and material being scraped off the platters. The disaster came in if the operators misdiagnosed the failure. If they thought it was just an electronics failure, they might start moving packs around to different drives in an attempt to read them. Of course moving a crashed pack to a good drive usually ruined the heads on the good drive, and moving a good pack to a crashed drive would ruin the good pack. I forget what our record for crashed drives/packs was, but I remember coming into work in the morning to find the system down and the operators hanging their heads once they realized they'd multiplied the problem many times over! Ah those were the days but I certainly don't miss those recovery scenarios!
My Uni years in mid 70's, programing in fortran, (first language HPBasic, very forward thinking maths teacher at high school).
You had to get a licence to operate the card punch machine, or put your program sheets into computer office and wait till next day to have it punched in.
Programs run on 'small" computer (with huge Winchester drives) and line printer communicating with KRONOS 120kms away in Melbourne.
Completed runs then put in pigeon hole to collect.
God help you if you used 1H1 in a format statement and got caught in a large FOR..NEXT loop, paper used to cascade out of line printer.
Uni also had MINI computer, 6ft rack, ITT teleprinter and paper tape punch for storage running BASIC.
Encouraged to use MINI for any job to get program practice, my first non Uni prog. OSCAR amateur satellite prediction charts.
Only problem, you had to book for MINI time, and only in 1 hour blocks each day.
Then SCAMP came along, 2K of RAM I think, then Z80 evaluation boards and tape storage.
I still have a Z80 eval board.
The thing is when oldtimers disease kreeps in, that's all I'll remember. Even a Meg of RAM was budget breaking.
And that's how the dust/hairs got in and caused the famous "head crashes".
I wonder if there's a video on Youtube with the screaming sound of a head crash on one of those? You could replay behind a sysop and give them flashbacks to the old nightmares.
){kind=link}