Post a Comment
DSL didn't really die, the "owner" (in quotes because he didn't do much as long as I was involved with the project) kicked everyone out. In other words, DSL didn't die, it was killed. But, Robert Shingledecker, the main developer of DSL still actively maintains a "spinoff" distro, called Tiny Core Linux. It works very well on older systems, and is even smaller than DSL was. It's very nice for old systems.
Do you by any chance know how they do X in Tiny Core? Last I heard, they were using KDrive, which is hopelessly outdated and not compatible with lots of hardware, or so I thought. Or I guess Xvesa will technically run on just about anything...
EDIT: I remember my incompatibility gripes were about the keyboard and mouse functionality, which broke for some older hardware of mine after Nokia submitted certain patches upstream. Xvesa always worked on whatever display, though it was crazy slow.
Edited 2011-07-12 21:47 UTC
But back then, it was nearly impossible to install any other applications. It was too bad, it worked so well on my K6-II 150mhz. I had to OC it to 550Mhz with some home made watercooler circuit to use Mepis.
Modern Debian can still work fine on P1 with 16mb ram, but you have to dig quite deep to clear the bloat.
It was the motherboard hard set limit, I was still running at 45C on load. I could have reach 700mhz easily and still run stable. Not bad from something machined with a wood drill and using a bunch of fish tank water pump. Those water blocks still work to this day.
Too bad high school geek look to be almost extinct. Those totally absurd challenge between us/them always end up with awesome result (and even more awesome catastrophic failures).
Your memory fails you (and rather early?). K6-2 certainly started above 200. With 300 to 400 being most likely / they seemed to be by far the most popular ranges IIRC (and Celerons of such ranges also had such overclocks)
550 was a standard top model; large part of slower ones were probably a silicon mostly capable of that speed but "factory underclocked", to fill the demand for mass-market chips; something which was and is a common practice (plus temperature isn't everything; MHz scaling is largely an architectural limitation; even K6-2+, which shifted to 180 nm process, wasn't capable of much above yours "easily and still run stable" ...which you can't know; nvm how "stability" has a very fluid meaning among pc hardware cargo cults)
Oh I'm so glad the kind you specifically refer to is almost extinct. Because there are plenty of HS geeks around, but today they seem to strive to do something with their computer (or with some fairly standard embedded board; or with almost run-of-the-mill robots) instead of superficialities (vs. building one in the style of, say, Galaksija) inside their computer.
That, and it's nice to be able to ask for (and readily get) an after-market cooler which is primarily as quiet as possible (one near which it would be easy to focus, one which won't get in the way of doing things with a computer) without getting the looks from salesmen like they have just seen an alien; without their surprised requests for clarification, without experiencing how hard the cogwheels inside of them are working when confronted with an entirely alien way of thinking about computers. And still getting IMHO grotesquely (vs. the requests) loud one, out of limited selection available (primarily targeted at "enthusiasts")
Modern Debian can still work fine on P1 with 16mb ram, but you have to dig quite deep to clear the bloat.
Exactly! I have never been a fan of projects meant to revive old computers because of that. The second reason is that I have never kept a computer for longer than 3 years. Else my house would look like a computer museum. 4 notebooks and 5 desktops since 2002.
I can second that. I tried DSL and then Arch Linux on an old Pentium II notebook. It had 128 MB of RAM. I had to buy a PCMCIA card for ethernet (€ 10 ;-)).
DSL did succeed to run, however it had some nasty quirks and, most importantly, the selection of applications is quite pure. After that quite joyless experience I decided not to give up the machine but instead try Arch.
With Arch Linux, I had really good performance, using fluxbox window manager with Nautilus as file browser, Epiphany as browser, Abiword for word processing. It ran really well. Web browsing was smooth and you could even listen to songs on Youtube. RAM usage was exceptionally low!
In conclusion, it was a nice experience to see that I can use my main operating system also in this special scenario. I thought that DSL would naturally be a better fit, but it had many shortcomings that Arch did not share. I did not perceive any disadvantage in using Arch.
Note that Arch Linux, however, will not run on the 486.
Edited 2011-07-12 11:27 UTC
Not a flame... I am fascinated by the subject and a hobby-tinkerer myself.
Every time an article of this sort is printed, one of the issues raised in the flames that you sought to preempt is power consumption. Is it cost effective or "green" (man, I have come to hate that word) to run an older unit with all of the advances in power technology?
There is also the question how much of a demand is there for such a computer (a power guzzler with a DSL desktop and no horsepower) outside of the hobby field? I don't think very much, although feel free to correct me if I'm wrong.
As it stands, I recycle much of what I come across, retaining only the newer stuff to tinker with. I have tried offering older systems for free (with Puppy, DSL, etc...) with no takers. While it kills me to take a perfectly working computer (which if you think about it, is a really amazing and sophisticated device) and turn it into scrap, the lack of a more compelling option makes it so.
Even if the prognosis was more positive, it seems like a stronger case would be made for a unit to be repruposed for use outside of the paradigm of desktop computer. Certainly the obvious is understood - print/file/media server, internet router, home automation server, etc...
But what else? I've had some ideas, mainly surrounding reducing the footprint of the case by eliminated hardware from the scanario. One idea was to build a motherboard (with integrated everything) and a compact power supply into a desk. now only a monitor, mouse and keyboard would need to be visible. In this scenario, a thin-client solution would be probable, so it ultimaltely would still be a "desktop" in one sense. The value-add here would be no visible computer in the room.
I'd love to see an article reviewing some of the things that people have accomplished using older PC hardware, with a focus on thinking outside the box. Maybe if I get some time I'll research it myself.
Actually, you're not the only one who keeps wondering about that. Very old PCs might not support ACPI at all, and even if they did there is a high likelyhood that it's very barebones. And these old PCs eat huge amounts of energy even without any extra devices connected to them whereas you can get a light modern PC running fine with as little as 120W PSU.
So, if the old PC is actually used a lot wouldn't it be more cost-efficient to get a light, modern PC instead and save on the electricity bill itself?
What? I couldn't hear you over the whirr of the fans of my workstation, HTPC, server, and Xbox, while my car is running outside to get the airco going before I drive off to do some groceries down the street.
Good point, but this remains a true question. Even if an old computer consumes 3x as much power as a new computer, it may still cost more to recycle it and buy a new one than to make it run for a few more years.
If we count in energy, let's assume that a computer costs 1 GJ* to produce and 2 GJ to recycle. During its service, its average power consumption is 30W (counting the time during the year where it's shut down in the average). Now, 3 years after, we have a new computer, which has same production and recycling costs, but this time consumes 10W.
Keeping the old computer for 6 years : 9GJ
Changing computer after 3 years : 10GJ
The computer which consumes three times less power (which is pretty nice for three years of hardware evolution) loses, given sufficiently high production and recycling costs.
I'd love to see someone do this calculation with actual numbers from computer manufacturers.
---
* 1 GJ = 300 kWh, for those who prefer these units
Edited 2011-07-12 08:50 UTC
I don't know how applicable that is. After all, new PCs are crammed out anyways all the time, regardless of whether or not you're buying one. Should the energy spent on manufacturing one be counted on it? And is the energy spent the same for all PCs manufactured anyways? I honestly don't know.
The "PCs are produced anyway" argument is a bit of a perverse one. As weakly linked to each other as they are nowadays, offer is still supposed to follow demand to some extent. If people in large number tried to make their PCs last longer, we would produce less new PCs as a long-term result, since they would sell less.
As for the cost argument, however, I certainly agree that all computers are probably not equal from a production/recycling point of view. The question is, can we approximate these costs for all of them with a weighted average cost ? Or is this model not precise enough ?
As I said, someone with the proper data and good knowledge of statistics should work on this production/recycling cost problem. As it stands, we can argue all we want, but still cannot tell which of our intuitions is right. Your guess is as good as mine, as they say.
Edited 2011-07-12 11:37 UTC
As I said, someone with the proper data and good knowledge of statistics should work on this production/recycling cost problem. As it stands, we can argue all we want, but still cannot tell which of our intuitions is right. Your guess is as good as mine, as they say.
Indeed, it would be really nice to see some actual, in-depth research on it. I simply do not have enough knowledge of these things to be able to make any meaningful estimations, I'm simply throwing questions in the air.
Another related issue is the method of disposal for abandoned electronics and the envionmental impact of that. How much winds up in landfills, and what is the procetag on PCBs leeching into the water supply? And who can forget the stories about "recycling" programs that involve burning electronic waste to reclaim valuable metals? What is the impact there?
There are many factors beyond actual power consumption from the wall from a "green" (uhgh!) perspective.
If 1GJ = 300kWh then at 15c/kWh that is $45 for my rate. In China it would be way less but then we all pay for that in their CO2 and other.
And why is recycle 2x as much, the PCBs are likely (sadly) hand disassembled in Bangladesh or Chinese backyards. Most of the weight of PCs is plain steel, aluminum or plastic, recycling the bulk metal should be low.
A newer PC that is more integrated and has less components, comes in smaller package and retails for lower price almost certainly uses far less energy in manufacture and recycle as well as in use. So energy to manufacture is likely a near constant % of the manufacture cost (or they'd go broke). I'd hazard a guess that a $40 mobo for me has only $1 of energy and a few $ of labor cost in it. The chips and case probably same rates. The info is probably out there though.
After 3 years of use, it is reasonable to think about moving on. Just as incandescent bulbs are going the way of the dodo (by law in US and EU), perhaps most all energy clunkers should be retired since energy costs are going up as fast as PC energy use can go down.
I wish more PC stores would stock something like the tiny mini-box picoPSU that can plug directly into the mobo ATX connector. These are close to 95% efficient and can handle up to 150W and let you shrink a basic PC down to the size of the motherboard and a few cm thick.
I just replaced my wife's 140W 20" CRT with an 20" LED display, uses 14W, that will pay for itself every year. Shame that the CRT looks so much better for TV though and still could last many more years.
In the free download book "Without the hot air" by Prof David MacKay, there is a good analysis of energy costs for almost anything you can think of. Watts, calories, BTUs are really all equivalent. The market more or less can trade $.. for any of those within reason.
The only problem with that last bit is that, for some people, amortization of cost is not a realistic option. For many, to keep a computer that they have already or can acquire cheaply and not spending, at minimum, several hundred dollars on a new computer makes a BIG difference to their bottom line.
Even in cases that are not so dire it may not be so attractive to spend a chunk of money today to save comparatively trivial amounts on the electric bill. Especially if they don't use the computer for much and the productivity gains to be expected are minimal.
In these cases, where productivity gains and long term financial savings are not a suitable carrot, what would be the motivator? We're back to the "green" arguments again, and as your discussion with Neolander is revealing, those are murky waters.
Finally, from the tinkerer perspective, it seems natural to make use of what you can find. Older, power hungry desktops are easily obtained. New, power efficient hardware, not so much. If I had to fork out bucks for every experimental idea that I wanted to mess around with I would give up the idea altogether. There is also a satisfaction (and dare I say beauty) that comes with taking something that already exists and repurposing it rather than destroying it.
Pentium III, 256MB RAM laptop with Damn Small Linux. Runs fine. Problem is when you try to browse the web... No matter how lightweight your browser is (with GUI at least), it's unusable for me. Not to mention using Flash and playing around with YouTube video clips
Too slow for WINE and MS Office. It's fine to use text editor or Abiword (which I have no use). For developers maybe? So far no use for my old computer, unfortunately.
You just reminded me of the long and drawn out discussions I used to have with my geek friends back in high school, regarding 486 vs Pentium. I was the poor kid with the 486 SX50 laptop with 8MB RAM, they all had Pentium desktops with 16MB or more, and were shocked that I could run Doom without lag just as they did. Of course that had little to do with number-crunching, something that their Pentiums would spank my 486 on, but all we ever did was play games and dial up to BBSes and Compuserve.
Man, I miss the 90s!
I didn't think I would have to elaborate, sorry. My estranged biological father bought a used laptop and mailed it to me in my senior year of high school, as a sort of "I'm still in your life, acknowledge me!" kind of thing. I have no idea how much he paid for it, but I felt extremely lucky to even have a computer at that time. I spent every dime of the money from my part time job upgrading that thing for the first few months I had it.
All my friends with their custom built PCs were from the wealthy side of the tracks, so it felt good even being in their league, regardless of the source. Plus I could fit it in my backpack and ride my bike to whichever friend's house was hosting the gaming sessions, since I was the only one who didn't have a car. Fun times!
No reason to be sorry, I figured it was probably some kind of handed down scenario.
It's just about slight comedic effect of seemingly somewhat contradictory "poor kid with a laptop in the 90s", particularly in the context of international forum ;p (for just one example: with my part of the woods - a fairly typical ex-Comecon country with shock of the 90s (but still one of more decent places to live), EU member a decade later - PCs in general can be hardly seen as a thing of the 90s; laptops, definitely only 2nd half of "noughties"; available part time jobs, a car, what are those?
)
http://www.connochaetos.org/wiki/
The same developers as the old deli linux
Based on ArchLinux recomplied for i586
My regular Arch installs use less than 12Mb booted to CLI and 40 Booted to Icewm or FVWM2 for instance. Room for a light browser like dillo or Links and so on!
Another use for DSL is turning those old laptops with a still functioning LCD into a digital picture frame:
http://sacbhale.net/dpf.shtml
Must burn fossil fuels faster!!(?) ~3 million years of production burned every year is not fast enough!!(?)
(seriously, while getting a long use out of computers as computers is certainly worthwhile, digital picture frames in themselves seem like icons of few things wrong with us; hey, fitting they usually show... us ;p )
A 486 with a CRT screen will cost far more in the long term than a more modern machine with an LCD screen due to electricity costs.
When you consider that P4 desktops and 17" LCDs are throw away items in western countries there is absolutely no point maintaining old machines except as a hobby.
Giving someone a 486/PI/P2 is like donating a gas guzzling 1950s car. Generous in theory but totally impractical.
Yeah right.
So what's the probability that someone who is so poor that they are still using a 1991-1994 vintage computer just happens to have a spare LCD monitor and VGA adaptor cable lying around?
Close to zero.
It is hard to even find a working CRT monitor. I haven't seen one for sale for at least three years.
They don't get thrown away. They get sold in lots and get bought by importers from third world countries like mine. P4s are sold for between 50 and a 100 USD here. Most people can't afford newer systems so these old ones sell like hot cakes. Even servers with power hungry netburst architecture dual xeon chips are sold here. All brought down from the west. Trash yes. Useless no.
Also you are wrong about p4s not being powerful enough. P4s have about as much power as netbooks running atom CPUs and we know how popular those are.
Also you are wrong about p4s not being powerful enough. P4s have about as much power as netbooks running atom CPUs and we know how popular those are.
I agree. Even if I have nothing less powerful than a Core 2 Duo T8300 (a notebook, which I am going to sell very cheap) P4s should be powerful enough to run almost every OS, including Windows 7. In the Western World we are spoiled, myself included.
In Australia most people buy white boxes with cheap noisy power supplies. They normally throw them out when the power supplies die after 4-5 years. The labour costs are far too high to repair or recycle them.
I actually said that 486s aren't powerful enough.
P4s are quite powerful but they are hot, noisy and very inefficient - especially when combined with a CRT monitor.
The same could be said of Haiku OS (and on a personal note, with none of the strangeness of AmigaOS.)
I've had it running on a couple of P3-800 with 512MiB of RAM with no issues for half-a-year or so.
The big thing with older hardware today is simply the quality of life issues, notably noise and heat and such. A slow, noisy box simply isn't attractive compared to other more modern systems.
We had stacks of some very nice rack mounted 5-10yr old Sun SPARC and AMD machines that went to the recycler. It was hard to see them go, but today most of what they did could easily be consolidated in a VM architecture on modern hardware.




