Hacker News new | past | comments | ask | show | jobs | submit login
In which I switch on a 30 year old computer and it just works (jgc.org)
71 points by jgrahamc on Aug 9, 2009 | hide | past | favorite | 28 comments



You know, the chances of an old 8 bitter firing up on the first try and loading its files from tape somehow does not surprise me. But you may want to worry a bit about the electrolytic capacitors, they tend to dry out over time.

If you did the same with a 15 year old 486 and read a bunch of 3.5" floppies (or worse, zip disks) I would be much more impressed :)

Eletronics back then were built to last, and the size of an individual bit on a medium from those days was large enough to withstand decay by auto-demagnetization better than anything new. Lucky you did not have any 'copy through' effects on the tape, that's the biggest killer trying to read old tapes.


Hmm. I have in front of me a Maxell MF 2HD 3.5" floppy drive with the following written it:

THESIS & PAPERS 15 FEBRUARY 1996 ZIP FORMAT / DOS

Externally it looks in good condition, but it's been stored in a box of other junk all that time. Tomorrow I'll take it to work and shove it in one of the machines there that still has a floppy drive and see if it's readable.


I wouldnt bet on a 3.5". However last year I found some old (like really old - 1992) 5.25" and that old P54C could still read them. In case you're interested there was simcity on them :)


Neat :) 13 years. I'd bet they'd work with some errors. Mostly dependant on the mechanical condition of the floppies, the 3.5" are usually pretty good because of the hard case.

I still have one of the first bernoulli 10MB removables here, I wonder how good they would fare. Not much point to try to find an adaptec 1542 just to satisfy my curiosity though.

You mention a bbc micro somewhere, do you still have that one ? (elite ???)


I have a BBC Micro Model B in working order with the associated tape player and a variety of ROMs. I also have some 8" disks with my code on them which was for the Research Machines 380Z (code written for CP/M and MP/M)


Wow... Better hold on to that one very carefully.

That's one of the nicest computers there ever was to start learning on. The lowest barrier to entry and some of the best books.

My 'growth path' was KIM-1, TRS-80 pocket, then the bigger trs stuff, then an english color computer clone called the Dragon 32, then a beeb, Atari ST and from then on IBM pc's.

At some point a company that I was partner in also had a PDP 11 but we did not do too much with it because of the power requirements, that was my first unix experience.


I think we have moved past that expection that a personal computer, a mainframe or electronics should last 30 years.

If you need a system that's reliable for decades (like a pension system), you probably should also think about how are you going to keep the system running after the current hw arch is discontinued and people knowing the specific programming language are retiring.

Nostalgia is great but I don't really care if my intel pc that I replace once a year lasts for 5 years or 30 years.


I wonder if the "designed to fail" trend will get worse or reverse in the coming years?


No, And my reason is because of the switch to lead-free solders. Lead-free solders aren't anywhere near as reliable as lead-based solders. Marketing folks claim they are. Reliability engineers can demonstrate that they aren't.


I hope - possibly against hope - that there will be some major shift around the corner, away from computing the way we do it today. There are plenty of contenders, a large number of which have passed the HN frontpage in the last couple of weeks.

So, I hope it will reverse, but I fear that it will get worse first...


I personally hope other things (eg price, increased performance) continue to be more important.


I have a ThinkPad which, judging from its survival so far, may still be around in a decade or two. Those things are built to last.


Modern PCs are pretty durable, IMO. Of course, if you go out and try to build a sub-$1000 PC, it's going to fail sooner or later. Invest in a durable case, a solid power supply and plenty of fans for cooling. I've found cheap DVD+RW drives fail very often, so invest in something higher quality than your $30 DVD+RW drive. I've never had a hard disk fail, but my motherboard has failed me several times. Now I have a Gigabit "gaming" motherboard (even though I'm not much of a gamer) that has lasted for quite some time.

I'm not sure how much LCD monitors last, but CRTs are known to work well for very long periods of time. I'm sure nobody wants CRTs these days, but I thought this fact was worth pointing out.

Oh, and I hear Apple's hardware is pretty good. In one year of heavy use, my MacBook has never ever failed me even once. This, of course, depends on how you handle your laptop. Mine never goes out of the house, so it's never exposed to shocks, dust or intense heat.

My point is, with a bit of thought, it's possible to build a PC that can last for quite some time.


I found this pretty impressive:

> Then I did the real test. After poking around and finding a tape of my

> old BASIC programs I typed LOAD again and explored the tape. 30 years on

> all the programs loaded from tape just fine and executed. I was able to

> spend a happy few hours playing character-based games that I wrote.

That his programs loaded after 30 years on a cassette tape makes some of those cheap flaky CDs' lifespans look pretty bad.


I'm quite certain that if you take highest quality CD-R medium (he mentions that he had used highest quality tapes), burn it carefully, and then store it away from sun, excess heat or humidity, it will last 30 years.

Wouldn't say same for the drive used though, and that's where the real problem kicks in.


I recently found a GoldStar cassette tape with my Atari hacks (in BASIC and in 6502C machine code) back from elder days of 1988-89. Externally and visually the tape is in normal condition, no chipouts of the ferromagnetic layer. Now to try it I need an Atari 800, anyone can send me one with a tape recorder ? ;-)


You could probably get the data with wav2cas - http://home.planet.nl/~ernest/atarixle.html


Oh my!!! Long ago I thought of possibility to create such converter, but did not know someone already implemented it. This is so cool, thanks for the link.


I almost linked blueprints or specs instead. I've been thinking about this sort of hardware emulation since PixelVision cameras came up elsewhere months ago - they recorded black & white video on audiocassettes.


Luckily for the computer cassette tapes, the FSK'ed bits lie way down in the sound spectrum, IIRC circa 1000 Hz.

I recently digitized some old favorite audio cassette dubs (TDK SA, most made on really good gear) and apart from atrocious wow-and-flutter, the high-end was completely gone on most, down to about 8K anyway.

Tape may still be the safest bet to store data long-term. I've read recently that (burned) CD's over 10 years old are getting iffy. Whereas people are still recovering NASA data off tapes made 45 years ago.


Oh I have a great anecdote about my Sharp MZ-700.

I got into the habit of loading up games then entering the debug console and poking in random points in memory. As they say: It's ROM, you can't break it!

But I broke it! I rebooted and found the caps lock in the debug console was broken (but did still work in BASIC). caps lock never worked again in the debug console.

I think it's pretty bizarre to accidentally rewrite a ROM chip, but amusing too.


When you say "Caps Lock" are you referring to the ALPHA key? I don't recall there being a Caps Lock on that machine.

Of course, it's entirely possible that writing to the ROM was decoded separately and used for some control signals. The Apple ][ I believe did some interesting address decoding to get control signals from the address bus.

Looking at the memory map for the MZ-700 it appears that it implemented bank switching so that the ROM could be paged in or out and replaced with the video RAM and I/O ports. Some of those I/O ports are connected to the keyboard.

http://www.sharpmz.org/mz-700/coremain.htm#banksw


I have bad memories of my Sharp MZ-700. The problem? For one, the display wasn't bitmapped, it had a fixed character display with just 2 256-letter fonts. Any game had to make good use of these fonts - like Angband.

Further, there were virtually no games for it.

When I got a Commodore 64 for my birthday, I was very glad to ditch the Sharp and enjoy a REAL computer.


I still use my 27 year old Sinclair ZX Spectrum now and again. Last year we used my CDTV as the living room CD player for 6 months as we didn't want to buy a new stereo or stereo speakers (my wife blew them).


I have a Compaq Deskpro Portable from the mid-1980s that still boots. autoexec.bat ftw!


The rate of change of technology has been "adopted" by hardware designers, now with a considerate built in replacement time. That way we don't forget its time to upgrade again. Moore's Law was really gathering momentum/social acceptance in the mid 70s.


Moores law gathering social acceptance ? Moores law is a derivative law, not one that is to be followed by anybody.

It's more of an observation than a 'law' as such.

In its simplest form it says 'technology gets cheaper, smaller and faster over time', there is nothing you can do about it, it will happen. Social acceptance and momentum don't enter in to it, it either is like that, or it isn't.

Hardware designers do the very best they can, if they could they would prove GM wrong whenever they could, but they can only work so hard and the market is only willing to pay a certain premium for a certain improvement in speed/size and so on. After the early adopters have paid the way the tech becomes mainstream, the price drops and the cycle repeats (it really is executed several times in parallel with staggered outputs, one 'thread' for every big chip manufacturer, that's where all the leapfrogging comes from).

This is about as market driven a process as you could possibly find on planet earth.


I read a great post by Kevin Kelly on the history of Moore's law that contradicts this opinion. The standards hardware engineers worked towards were based on assumptions of a new transistor density, and the assumptions invariably became self predicting prophecies. Achievements of anything less than Moore's law improvement weren't acceptable! I'll dig up the link if you can't find it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: