Unsurprisingly I got a 503 error. I'll be more curious how many requests it can handle if it is really only on floppy (no caching etc..). 1.44 MB is plenty of space to host a nice website !
Yes sorry, I meant I'll be curious to know how many requests per second that floppy can handle. Well for sure it's not enough for instance since I still get 503 :). Since I am curious so I'll just bookmark for next time.
I published my first website in 1994 via floppy disk.
I didn't have the internet at home so would design the website there, save to floppy disk, then go to the local library that had internet terminals you could hire for an hour. From there I'd upload the website from floppy disk.
I did the same in 1996 (hosting at my univesity). On an Amiga. Which didn't even have a TCP/IP stack (I think you could get a commercial implementation).
Once back in the early 2000s I connected the floppy drive incorrectly and the entire thing melted inside with some smoke. It was rather funny at the time, mostly because it wasn't my computer. That said, I'm hoping bigcat.space r&d headquarters are not on fire and the website will be back up soon :)
Pragmatic approach would be host it from floppy disk, but have some caching proxy before it. It could cache into RAM, most modern machines have enough memory to cache many many many floppies. Even Raspberry Pi.
I was wondering how he prevented the kernel from caching the disk blocks in memory, but then I saw he is running it on a machine with 4MB of main memory, and presumably most of that is eaten up by the TCP stack, web server, and parallel ethernet drivers. Not to mention memory needed by the OS itself.
Even more surprising, his page isn't just pretty, it is functional! You can get a copy of the demo for Wolfenstein 3D off of the site (supposedly, I did not try).
The pragmatic approach would be to not host it from a floppy. If you are going to host it on a floppy then just having it all constantly in cache somewhat defeats the point.
It's hosted on a machine with 4M of ram on windows 3.11, and specifically uses a web server which doesn't use caching (And I don't think Dos 6 caches files at the OS level, at least by default)
I can't remember how buffers in DOS work, but that page seems to indicate that for input purposes they're used as some kind of a read-ahead buffer. That doesn't sound quite the same as keeping recently or commonly accessed data in RAM to speed repeated access.
DOS did have SmartDrive [1], though, which IIRC provided a more typical disk cache. I wouldn't think it was loaded in the default configuration, though I'm not sure.
As said in that page and the documentation, the read-ahead is the [,X] part, which is actually disabled by default in every DOS past 6.x (incl FreeDOS). The first part is the number of primary buffers which is indeed "for keeping recently read data in RAM".
Smartdrv, like Fastopen, just allows disk buffers to survive a Close File call. Normally DOS only does close-to-open consistency i.e. buffers are flushed on close. Smartdrv specifically also allows write-back rather than write-through.
And, anyway: yes, Smartdrv is enabled by default in 6.x installs with enough memory.
Nice, although if it is set to 15, that's only 7,680 bytes of buffer, and even 99 is just 50,688 bytes. It's not loading now, but I suspect a single load of all those images will flush that data
When I ran my BBS in the 80s, it was hosted on an 8 bit machine with 2 3.5 drives; however the BBS software itself which included all pages of the ‘site’ where on a ram disk of 256kb so they would load ‘instantly’. The 2 disks contained (weekly refreshed) public domain software and demos.
Now I am curious if I put it online (I still have the machine) how it would work. Seems fun to do.
What is causing the 503 errors? Could the server be reconfigured with eg. a longer timeout to avoid them? (On that note, is there a limit to maximum timeout on the client-side?) Or is the actual hardware getting overwhelmed by the traffic?
The whole machine only has 4MB of RAM so presumably at some point requests start getting dropped. Then again, you can store a lot of requests in even 1MB.
Which reminds me of a catastrophe movie, where all countries launched all their nukes, and earth became radioactively poisoned. A submarine survived, looking for signs of life, and they received sporadic messages like "the whales are fine" (which made sense to them since whales live under-water). They went to the origin of the signal, only to find a computer turning on and off because of its solar panel, sending messages at every boot.
Which reminds me of the http://solarprotocol.net/ I stumbled up on earlier today. It' a network of solar powered servers and the content of the site is delivered by the peer with the most sunshine.
I do have a fully solar powered VPN gateway/LTE router/web server/Wi-Fi relay - in my case, it’s out of necessity, as I live in a deep valley, and the only way I could get internet access here was to stick a mast on top of a hill with an IP67 box of goodies and a solar panel and battery, and an LPDA and yagi antenna for LTE and 802.11n respectively. The web server isn’t public, but fulfils a few important functions for me.
Sounds like our local ham relay station. On the top of the mountain, solar and deep cycle batteries, and a radio relay. All housed in a cheap little shack.
I was last up there around '98 to help top off the batteries with water, but it's still running strong today.
It's fully solar because a long time ago the sun got hot and the solar system got put together and that combination resulted in life and then later some sample of that life built a website with materials extracted from what the sun did to the original mess.
Not really. The transuranics are believed to be mostly created during supernovas (core-collapse supernova nucleosynthesis) and neutron star mergers.
Those events are not particularly similar to our sun's normal solar radiation.
I think you are grasping at straws (although I'll be honest I don't know what that literally means). And I think the word solar really only applies to our own sun.
Those two things can easily both be true, and it doesn't change the fact that shallow comments are bad for HN and therefore against the rules here.
It's perfectly possible to write a non-shallow comment about work that you consider shallow. More likely, though, there isn't much that's interesting to say about it, in which case you shouldn't post anything.
But that doesn't detract the artistic merit of the submission.
Agreed, looked at in that light I'd suggest that it actually adds to the artistic expression. I was one of a few, apparently, that the site loaded for (sans images) and it was a tiny thrill to see it load. So, it felt like watching a performance of sorts, for me at least.
Lead with the positive curiosity of wanting to know how the author did this rather than your negativity of “what’s the point?” The author could very well post a follow-up article detailing how
What is there to talk about really? Hook up a USB floppy drive to an IIS machine, select the floppy as the source for the site and voila. Takes 1 minute to set up if you have a USB floppy drive hanging around.
It’s great that you already have an idea of how this was done; others may not. There’s a lot to talk about there for people unfamiliar, don’t you think? Saying otherwise would be very close-minded.
It seems to me from your other comments in this thread that you’re intent on being negative / reductionist to others work without offering much to the conversation, so I will happily disengage with you now.
If you really checked my comment history, you’d see that I actually gave a warm compliment just a few days ago. I apprecite the work of others, but not necessarily when the work is trivial and keeps popping up on HN quite often.
The site loaded for me. So I can tell you; it does not use IIS. And it's running on a 386 so I doubt there's any USB involved. What he pulled off is a lot more interesting than what you're describing.
It was a fun thing he did to see if he could and when, to his apparent delight, it worked he decided to share his results. As someone who grew up in the era of floppy disks and constant worry about disk space and performance I found the whole thing very entertaining.
Sometimes we need to take a break from all the serious and just have a little fun. Items like this guy's post are why I come to HN.
Thank you for the clarification. I too find this interesting, especially because it runs on win 3.x. The floppy part is a bit unorthodox, but cool nonetheless!
One point could well be to see how your setup handles extreme overload?
I guess different OSes and server software would handle it very differently, so using HN to trigger a huge overload would be a smart way to get some science done along those lines.
Reminds me of Hotline Connect file servers, sometimes people would host directories linking to different drives in their system, including the A: drive
