Slashdot Mirror
All of Gopherspace Available For Download
An anonymous reader writes "Cory Doctorow tells us that '[i]n 2007, John Goerzen scraped every gopher site he could find (gopher was a menu-driven text-only precursor to the Web; I got my first online gig programming gopher sites). He saved 780,000 documents, totalling 40GB. Today, most of this is offline, so he's making the entire archive available as a .torrent file; the compressed data is only 15GB. Wanna host the entire history of a medium? Here's your chance!' Get yourself a piece of pre-Internet history (torrent)." Update: 04/30 00:16 GMT by T: As several readers have pointed out below, our anonymous friend probably meant to say "pre-Web," rather than "pre-Internet."
Beat me to it. The summary should read "Get yourself a piece of pre-world wide web history," since gopher came AFTER the birth of the internet (1981) but before the widespread usage of the web (circa 1993).
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
To a lot of people, WWW=Internet. Us old greybeards who remember when the Internet was telnet, FTP, e-mail and Usenet know better.
This ain't rocket surgery.
In gopher, everthing is either a link or text. There is no way to embed a link into a body of text -- what is now called "hypertext".
So, yes, Usenet preceded the Internet in the sense that it did not rely in IP, though both generally evolved around the same time.
But, there was a rather vibrant pre-WWW internet where the protocols of choice were smtp (mail), ftp (file transfer), and gopher and archie for repositories of places to find stuff. News could be carried via nntp (net news transfer protocol).
What some may not know was that sendmail could work over transiently connected points as well, rather like usenet. Anyone still remember bang path notation? One would address mail using the sequence of hosts required to get it from one's own to the destination, using names understood by each successive host in the sequence. One of the reasons sendmail configuration files were so horrendous was to permit relaying between networks using different host naming conventions.
In Liberty, Rene
Yes.
In 1997 we had a 100Gb disk array holding the research data from our lab, all of which was available via gopher (and ftp, and the web). We moved to a 200Gb array shortly after, and then a 400Gb after that. And then 3Tb, around 2008.
Sometime around 2007 or 2008 the SunOS system that ran the gopher server died permanently and was replaced by a virtual linux server without gopher. Even without that server, I found not long ago that I was still creating .cap files -- which were gopher, as I recall, but maybe archie.
Quantitatively, online currently I have more than 15Gb of data for just 1997, all of which was gophered at the time. In 1998, another 18Gb.
So, I would say, had the gopher scraping been done in 1997 instead of 2007, the result would have been a lot more data. In fact, a few months earlier in 2007 and it might have BEEN a lot bigger.
No, just another ten years of November.
I believe you mean September.
What do you mean, "was"? Gopher still works fine. There are dozens of servers out there. See quux.org or just install your Linux distribution's gopher package and fire it up.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
I've been around a while, and I can't think of any time a Slashdot editor fact-checked, spell-checked, or proofread a submission. Look at it, they put the entire thing into a quote. That way they can just say they're quoting the submitter and that's what he said.
They might add the "UserXXX writes," part themselves, but a couple characters of perl could probably do that part just as well.
The original pre-RFC HTTP states that a response is an HTML message.
Here's my explanation in graphic form: http://parseerror.com/images/explain/internet-vs-web.jpg