Gopher server revived after 15 years of downtime

Blogging site MetaFilter has restored its Gopher server, after 15 years of downtime. Explain to kids these days that there was a time when the World Wide Web was just one of several competing ways to navigate the internet and they'll stare at you blankly. Tim Berners-Lee gave the world Hypertext Transport Protocol (HTTP) in …

[x 7]

something I never got my head round........whats the difference between Gopher and an FTP server?

[lpcollier]

Gopher (almost) seamlessly spanned multiple servers with a single interface. FTP requires the user to log in (even as a guest) to a separate FTP server for each site you want to visit.

[AustinTX]

Gopher vs FTP

An FTP server generally displays folders full of files and symbolic links in the order which they actually appear in real folders on the server. Like an HTTP directory listing. The FTP server generally only displays files and folders on one server - it doesn't span servers (though this can be accomplished). Also, FTP is technically a command-line interface though this is masked by using a GUI FTP client. FTP directory contents are fundamentally bound to real accounts existing on the server.

A Gopher page's content is arranged at-will and contains hot-links to pages and documents on various servers. Just think of Gopher as the web without embedded images or self-launching widgets (though there can be entry fields and submit links for search engines and such).

[x 7]

Re: Gopher vs FTP

thanks to those who replied to my question, its nice to finally know after all these years......

"Just think of Gopher as the web without embedded images or self-launching widgets"

sounds like heaven doesn't it!

[Down not across]

Re: Gopher vs FTP

"Just think of Gopher as the web without embedded images or self-launching widgets"

sounds like heaven doesn't it!

It is.

I used to store lot of docs in the 90s in my gopher server. Need to dust off that old Ultrix box and see if it still boots. If it doesn't thump on its side should cure it anu of the disks have jammed (some of you lot may remember some disks had a habit of not spinning back up after being off for a while due to non-ideal lubricant in the bearings).

[Doctor Syntax]

I'm not sure why the downvote for x7. It's perfectly reasonable to ask if you don't know.

Perhaps it's time to consider whether there might be use cases better met by Gopher than by WWW - and if so, can adverts be kept out?

[Anonymous Coward]

"can adverts be kept out?"

I expect this is the very reason it lives on and someone requested it be turned back on at MetaFilter.

[Jim 43]

I would have paid good money to be sitting next to the network/firewall tech who picked up the service ticket to allow that traffic.

[AdamWill]

pfah

I have a colleague who's still configured to accept UUCP-routed mail...

another one I can think of, John Carmack's .plan file was updated up until 2005: http://www.bluesnews.com/cgi-bin/finger.pl?id=1 . Someone else at id updated theirs in 2007: http://www.bluesnews.com/cgi-bin/finger.pl?id=476 . Sadly seems like you can't connect any more, though. :(

[Anonymous Coward]

Apps

I recall using telnet to get to a VAX and then running gopher. One day my boss asked me to look into this www thing (around 1992ish.) It looked a bit chaotic to me compared to the neat menus in WAIS and gopher.

Shows how much of a visionary I was ...

I used to telnet a VAX from my PC then do something involving a X25 PAD then telnet on to NSF (I think) in the US then telnet to CERN and then move my cursor around "links" or something. Couldn't find any kittens anywhere for quite some time although I did find some other interesting image files that I downloaded, one by one, from someone's temporary shared nfs mount until it suddenly vanished for some reason. I kept on using up my homedir quota - the whole process took ages.

[ddogsdad]

Re: Apps

In February 1993, the University of Minnesota announced that it would charge licensing fees for the use of its implementation of the Gopher server; this killed Gopher. It was released to GNU in 2000.We played with it a bit when it first came out but Mosaic killed it quick. We had been using FTP for nearly 20 years before Gopher came out on Cyber Mainframes. I used FTP to connect to our Mainframes with My Osborne running CPM starting in about 81. Before that it was dumb terms and Apples.. Gotta luv the internet at 120 Bits per second. Damn thing still works too.

[Bloakey1]

Re: Apps

<snip good stuff>

"Couldn't find any kittens anywhere for quite some time although I did find some other interesting image files that I downloaded, one by one, from someone's temporary shared nfs mount until it suddenly vanished for some reason. I kept on using up my homedir quota - the whole process took ages."

Whatever happened to all that ASCI porn? (yep I know it is still out there) I have often gone to a uni dot matrix printer looking for a POP 11 program printout only to find some weird ASCI stuff and a knackered ribbon.

Gopher and WAIS were good stuff back in the day I am sure they would be happily used now if only the "yoof" <sic> knew about them.

[AustinTX]

Good Gopher Times

Back when I was a college boy, I got my Internets for free by dialing up the local university's free Gopher-only dialin for library book availability. You dialed in just like it was a BBS or CompuServe, only you just got their Gopherspace.

I could maneuver my way into a real free Unix account provided by cyberspace.com by using Gopher search engines (Archie, Jughead and Veronica were the Google, Yahoo and Bing in those days) to find a "gopher to telnet gateway". I typed the destination into the gateway's Gopher page field, and if it was agreeable, my screen turned into a telnet window.

Cyberspace gave free trial accounts to anyone who applied online, so from that point I had a real commandline and tools like Lynx and Pine. Pine got me my email and newsgroups and Lynx got me my web pages sans images and file downloads. If I wanted anything on my local machine, I had to mail it to myself at a local BBS (9JACK9) which connected periodically to the Internet since the nature of my connection prevented X/Y/Zmodem from working.

[AustinTX]

Re: Good Gopher Times

BTW, if you want to visit Gopherspace, there are still, well, dozens of servers to connect to and thousands of relatively updated links. Google 'em. You'll likely find that your browser no longer supports gopher:// addresses, but if you install Lynx on a Unix/Linux system, it still supports it. Lynx actually makes a handy file manager if Midnight Commander is too heavyweight for ya. See paragraph on bottom right here: io.fondoo.net

[glen waverley]

Re: Good Gopher Times

Ah, lynx. Memories of the old days

[Pirate Dave]

Re: Good Gopher Times

"Ah, lynx. Memories of the old days"

Old days? I just used it last month on a Linux box. It's still handy for those of us who don't like GUI on our Linux boxes that are doing semi-important stuff. Although I do admit with all the new advancements in "web" stuff over the past 15 years, pages sometimes don't look as good as they used to.

[Michael Thibault]

OMFG! My browser supports the protocol natively.

Not much of a loss not having known, though: I've long since ceased lamenting the decline of gopher as a protocol, as the ever-more-fruitless sessions, attributable to decreasing deployment of it, were the writing seen on the wall not many years after the protocol launched.

[John Brown (no body)]

I hear it's part of the undernet and it's used by terr'ists and peados 'cos no one else knows it's there. No need to waste time and resources encrypting stuff when the security services don't even know it exists.

[Mark 85]

No need to waste time and resources encrypting stuff when the security services don't even know it exists.

Well.. they do now thanks to you reminding them.

[AustinTX]

Gopher is all cleartext anyway, man.

[Nik 2]

Ah, but then there was 'Tbcure', which had many of the same documents protected by a ROT13 cipher.

[Mike 16]

Could you use it legally?

IIRC, although there were freely available gopher clients, the gopher protocol, and the right to run a gopher server, had to be licensed. Or have I gotten it confused with some other "we will be billionaires by selling this nifty new networking tool" scheme?

[Anonymous Coward]

Re: Could you use it legally?

But of course you could, and can. All was paid with USA tax money and developed at a land-grant university. If memory serves me right, the gopher protocol, the reference client and server were part of a distributed database project paid by DARPA. It wasn't all about killer robots back then.

You must be remembering roxen, which included gopher as a paid extra. Or was it aolserver? Did not take my fish oil pills today. :-P

[Mark 85]

Re: Could you use it legally?

Unless it was paid by AOL, I don't recall getting charged for using Gopher or any tools back then. ISTR that there were some shareware tools available but being shareware most people never paid for them.

[Deltics]

Re: Could you use it legally?

You're sort of correct.

University of Minnesota (where gopher was developed) indicated it would charge to license it's implementation of the server. This effectively put the kybosh on the protocol as this created concern that even independently developed implementations might fall foul of this licensing requirement.

[Anonymous Coward]

Re: Could you use it legally?

Why yes. I remember now. It was one of the first cases I was aware ---of course there are a lot more and previous--- of a US University administration trying to cash in research paid for by the US gov with taxpayer money, that by that country's legislation is in the public domain. They have managed quite well patenting and cashing in materials science, CS and molecular biology research, btw.

[Anonymous Coward]

Re: Could you use it legally?

Wouldn't put it past those SOBs. The story I heard was that the gopher server was originally developed at UMD, and was called "bulldog" before the main campus caught wind of it, jumped on the coattails, and renamed it after Goldy.

Sure, "gopher" was a better pun, but still, that's kind of a Borg move by the Main-U.

Anon, just in case they ask for my diploma back. It does say U of M on the top, even though I went to the Duluth campus.

[Ole Juul]

needs some work

I saw the story a couple of days ago on Hacker News, and went to have a look. It's not working properly. Now I see the story here (kudos to El Reg) and try the gopher again. It's still not working properly. Is this a case of bragging without checking?

I'm using a classic WATTCP DOS application and still the site is extremely slow and most items just return a "(null)". It's an unfinished mess. I don't know what they're running this on, but a period appropriate floppy system is much faster. There are still a bunch of gopher sites out there that actually work. I think this is embarrassing for MetaFilter.

[AustinTX]

Re: needs some work

Could be there were problems which they fixed.

I'm browsing it using Lynx browser and it's lightening-fast.

[Ole Juul]

Re: needs some work

Thanks. Now that you mention it, I see that I can browse it in Lynx as well but that's using 32 bits which is cheating. The site does not work with one of the original Gopher clients which I'm running in DOS 6.22 on bare metal. I have no problem with "real" gopher sites, so this is not as retro as they're suggesting. I guess it's OK, but I'm disappointed that us vintage guys won't get any use of it.

PS: try gopher.floodgap.com

[45RPM]

Gopher is awesome - and makes the Internet useful for users of elderly computers. I'm delighted each time a new server for this venerable but efficient protocol is (re)started.

[Flywheel]

Bandwidth

The great thing about gopher is that it uses so little bandwidth - very handy if your roaming and need to access data, but don't want to waste money on all the WWW flab.

Oh, and speaking of memory usage, my own gopher server is consuming 2Mb of memory on my 512Mb Raspberry Pi Model B.

[davidp231]

It was munching on 2-3 megabits of RAM?

[The Sod Particle]

Now we've started,

Just need to resurrect the finger protocol now. For the IoST* obviously.

*Internet of sex toys

[allthecoolshortnamesweretaken]

Re: Now we've started,

NOW you're talking! Kickstarter?

[To Mars in Man Bras!]

Gopher Broke

Does anyone else remember having to dial into gopher via cheesecloth? And then hot-wire the jumpers, so you could piggy-back over the transatlantic cable to UCLA's rounding thermals server?

One wrong keystroke while in the flapjack transfer mode and you could fry your computer as well as set fire to the local exchange.

Great days!

[Bloakey1]

Re: Gopher Broke

<snip>

"One wrong keystroke while in the flapjack transfer mode and you could fry your computer as well as set fire to the local exchange.

Great days!"

I say old chap, I do believe that your banter is broken.

[Anonymous Coward]

"a file system is a good model for locating documents and services"

Written by someone who has never gone hunting for a document in a filetree, methinks.

"I know we have an expenses form. Is that in the "HR" folder? Or is it under "Finance"? Or "Policies and Procedures"? Or "Templates"? Or "Misc"?

Nope: it's in Alice's the Accountant's personal shared folder"

[Jonathan Richards 1]

Re: "a file system is a good model for locating documents and services"

> hunting for a document in a filetree

Yup. For effective information retrieval, you need either (i) a filing system with associated fileplan, or (ii) indexes and and a unique document ID/location. There's a reason that many libraries implemented the second of those two concepts hundreds of years before computers were invented. Mis-filed documents in plan (i) are effectively lost, and as the quoted post says it's almost certain that human error will result in mis-filing, and soon. Compiling and maintaining the indexes for documents is resource-intensive, but provided that the documents are on the literal or metaphorical shelf where you're expecting them, then retrieval is much more precise.

HTTP became the protocol of choice across much of the Internet, but it needed search engines (web indexes) really to bring it to life; Lycos, Alta Vista, and some other outfit that I mis-remember...

[Luiz Abdala]

Re: "a file system is a good model for locating documents and services"

don't you forget also those precious gems in file systems used by more than one person:

New Folder

-> New folder

->New folder

and my favorite:

New folder

New folder(1)

New folder(2)

Copy of New Folder(1)

Copy of Copy of New Folder(2)

... repeated ad nauseam all over the system tree.

[Anonymous Coward]

Re: "a file system is a good model for locating documents and services"

> Copy of New Folder(1)

> Copy of Copy of New Folder(2)

> ... repeated ad nauseam all over the system tree.

Long, long ago I had the idea to write a Windows 'add-in' of some description that would allow administrators to define regular expression rules that would be enforced on file system directories. So if you had a policy that files in the "accounts" directory must start with a 3 letter abbreviation of the department name, followed by the year, followed by 3 initials of the author, or whatever, then you could define a rule and filenames that didn't match wouldn't be allowed to be saved.

You might get a lot of support calls form people saying "my file won't save", but at least you would know who the dimwits needing a bit of assistance from a 4"x2" clue-stick were.

I never did enough Windows programming to become sufficiently proficient and it got forgotten about when the Web and document management systems became a bandwagon.

However, now you know, so if anyone fancies giving it a go...

[Jonathan Richards 1]

Re: "a file system is a good model for locating documents and services"

> Copy of New Folder (1)

Ah, yes. An essential element of my plan (i), i.e. a filing system for document retrieval, is that you NEVER let the users create a folder; there has to be a very controlled process for that, along with updating the file plan.

Slightly OT: this is where the metaphors get a bit mixed. In the formal filing systems which I have in mind, viz. the UK Civil Service of thirty years ago, paper documents were sequence numbered by a Registry and inserted into folders (colour coded for security classification). One or a series of folders constituted a Registered File. In the computer metaphor, most often, a "file" is the same thing as a "document".

[Anonymous Coward]

Re: "a file system is a good model for locating documents and services"

When I started here, I found these folders on the ICT shared folder...

Cisco drivers

New Cisco drivers

Latest Cisco drivers

The folder called "Intel wireless drivers" was in the "New Cisco drivers" folder!

[Anonymous Coward]

Re: "a file system is a good model for locating documents and services"

well, where else would you expect the Intel drivers to be?

[Anonymous Coward]

I remember when I first got on the internet (AOL in the mid 90s) gopher (as well as WAIS) were still prominently featured in the menus, but it was as far as I could tell, already obsolete. I made a few attempts but never managed to find anything interesting in there. I always felt vaguely cheated by that. If I'd got on just a few years earlier when gopher was still cool... something, I donno.

But now, after all these years it almost sounds like gopher is cool again, after a fashion. Perhaps I will yet be able to fulfill my ambition to find something interesting through it some day.

[Wayland]

Gopher cool again

It just needs a suitable task to tackle and it will be cool.

[CarbonLifeForm]

Gopher memories

I remember the "Searched all of Gopherspace" missive fondly. I used gopher often for my Ph.D. literature searches.

It may have more applicability now-a-days than the author gives it credit. Being able to search docs worldwide without dealing with the Web's bloviatory excesses does have a certain appeal.

Hook it up to a NoSQL server and see what happens.

[queenslight]

Official HTTP Gopher Proxy, For Meta Filter

If anyone ever sees this or it wasn't posted already, I just wanted to make sure I shared this with yall here:

http://gopher.metafilter.com:70/