Huge spike in GitLab imports after Microsoft reportedly acquires GitHub


I always found that floppies were eaten by the floppy disks... and so, a bad floppy drive would eat your disks.
 
I always found that floppies were eaten by the floppy disks... and so, a bad floppy drive would eat your disks.

Oh, I've certainly had two floppies that experienced binder failure. (Squeaking sound and, when the disk was pulled out ASAP, it revealed a transparent ring where the drive heads had wiped the magnetic oxide off the surface) Thankfully, neither of those were important. (A home-made DOS 4 boot floppy and a Delrina DOSFax/WinFAX pack-in that I'd done a "Zip up the contents" backup of years ago.)

Is it possible your "bad floppy drive" example just had its read/write heads gunked up with magnetic oxide build-up? (It actually reminds me of the Iomega Zip Drive's infamous click of death, where a bad drive would mess up a disk, causing other drives to click as they repeatedly retracted and re-engaged their read heads while trying to get a read of it.)

As for my EA Kids disks, I tried them in two different floppy drives (A USB floppy drive on my ddrescue-enabled Linux machine and a traditional floppy with VGACOPY/386 on my retro PC), both of which work beautifully for everything else. (I did have a handful of disks with one or two bad sectors where ddresdcue had to retry several times, but the bad EA disks had a forest of bad sectors. When dumping was attempted using VGACOPY/386, which has a visual read-out, it clearly showed a pattern of bad and weak sectors indicative of crappy media rather than the one or two isolated problem sectors you'd see from age related degradation of good media.)
 
@ssokolow: Gaah... nightmare scenaro: transparant floppies! About magnetic gunk: Might very possibly be my case. Lots of dust on the heads, and never cleaned it properly (http://retrotechnology.com/herbs_stuff/clean_disks.html)

The scenario of floppies was on the university desktops. And sometimes, even new floppies died at first time use. Floppies were expensive back then, but also bad quality.
 
Ahh.

One interesting tip I read was to create your own non-abrasive cleaning disk by taking advantage of the manufacturing standards within the floppies themselves. Cut open a floppy you don't need, remove the data surface, and rotate the felt sleeve so that the gap no longer lines up with the shutter, then put it back together. Induce the drive to attempt to read this disk for 20 seconds and the heads will wipe themselves off on a material spec'd to be soft and safe.

Source: http://goughlui.com/2013/05/19/project-kryoflux-part-6-dealing-with-difficult-disks-and-drives/

(Though that guy uses Vaseline to lubricate drive rails while David, The 8-bit Guy, prefers lithium grease when restoring drives.)
 
Self-hosting is generally a bad idea. You have no idea how many times I've been frustrated to find that old Pascal resources I want to for writing utilities for my retro-hobby PC have fallen off the web because the site were they were hosted is gone and the Wayback Machine hadn't gotten around to archiving them. Git as opposed to bog-standard files on an HTTP site will just make that loss of history worse. (If the Wayback Machine doesn't archive something as significant to the late 90s and early 2000s as its FTP sites, I seriously doubt it's going to archive git repositories.)
It may be pertinent to note that git repos hosted on a web server just makes the web server serve the files, so they're accessible to a standard web browser under http/s access. They're not easily browsable, because it's the contents of your .git folder from inside a repo, but you can clone them and then checkout any version you like.

That doesn't mean the archive services will pick it up of course, especially if nothing links to it.

Also, I looked about a month ago and a hell of a lot of the gopherspace is still up. An awful lot more than any old deep website links you had from about then, anyway.
 
It may be pertinent to note that git repos hosted on a web server just makes the web server serve the files, so they're accessible to a standard web browser under http/s access. They're not easily browsable, because it's the contents of your .git folder from inside a repo, but you can clone them and then checkout any version you like.

That doesn't mean the archive services will pick it up of course, especially if nothing links to it.

Even if The Wayback Machine does pick a git repository up, I haven't heard of any way to work around the link indirection they use to inject their history-browsing header to avoid having to manually download each individual file in the repository in order to reconstitute something the git client will recognize.

Also, I looked about a month ago and a hell of a lot of the gopherspace is still up. An awful lot more than any old deep website links you had from about then, anyway.

True. I just haven't gotten around to poking at gopherspace yet.
 
Self-hosting is generally a bad idea.
Nonsense.

You have no idea how many times I've been frustrated to find that old Pascal resources I want to for writing utilities for my retro-hobby PC have fallen off the web because the site were they were hosted is gone and the Wayback Machine hadn't gotten around to archiving them
So that's the reason you dislike self hosting? Curious to think that only "self hosted" files on the internet can be removed.
 
Nonsense.


So that's the reason you dislike self hosting? Curious to think that only "self hosted" files on the internet can be removed.

Oh, I'm well aware of that. Half of them were on a big programming resources site that lost interest in their old catalogue... but it's far more common for a self-hoster to lose interest or not care about keeping their old URL schemes resolving than for a big hoster to close up shop. (And, when they do, you tend to find behaviour like with Google Code or Vine where, for PR reason if nothing else, they keep the URLs resolving but switch to a read-only setup.)

Heck, even corporate self-hosters still have that problem. For example, if it weren't for people who upload FTP dumps to the Internet Archive, we'd have lost a lot of official update patches and supplementary files for older games.
 
Nonsense.
If I may add one addendum: Deadman switch. You don't pay, the host-provider puts you offline.

I recall an urban legend, where friends self-hosted their own email server, with different domains for their mail enjoyment. The one who pays the bills suddenly dies. Even the others being tech-savvy, could not keep the server alive (move from automated payment to manual payment and change owner).
 
Last edited:
Back
Top