Gopher was great. I remember telneting through university networks, exploring, and by chance finding fascinating stuff; more than I could ever read.
But I think the author misses the real killer feature of the web. The difference in hypertext models (external vs internal) is mentioned, but glossed over. Gopher was stuck with hierarchical indexes. It suffered from all of the limitations of hierarchy and had no way to acknowledge the relations between things. And documents in Gopher were dead ends. The web's ability to link from anywhere within any document to any other (and even locations within it) is a huge differentiator.
That was critical - instead of a limited hierarchy leading you down to a dead-end, you had an unlimited neverending web of related ideas. Imagine the C2 Wiki or Wikipedia in gopher format - all the same content but with no links. It wouldn't have anywhere near as much value.
The simplicity, cleanliness, and order of Gopher has beauty, but the much messier web has real-world value. The network effects of linking would naturally cause it to grow exponentially faster.
The organization of knowledge is an old problem. Trees and indices are very handy ways to locate books, er, files. There were three things that killed gopher: multimedia, licensing, and search.
Inline images really set a webpage apart from the typical wall of text common at the time. University of Minnesota's licensing scheme made the web more attractive.
But the killer app was the web search engine: first DEC's AltaVista and then Google. Before web search engines, finding info you wanted on from a website was terrible - unless the pages were modelled after the traditional trees and indices. Links were fun (remember webrings?) but not nearly as useful for finding information until Google figured out ranking pages by weighting links.
I totally agree with this. The web reflected the messy organization and connections of the real world in a way that let us discover and link content in profound ways. Great supporting example with C2 Wiki w/out its links. That would be horrific haha.
NCSA httpd supported an option that allowed every user on a system to put html documents in a specially named directory and have them instantly served up to the world. User created content exploded--undergraduates (like me) were instantly addicted.
I'm not even sure who came up with ~/public_html (Rob McCool or Ari Luotonen are the primary suspects) but whoever it was will always be an undersung hero of the web.
That's funny...I've never thought of it that way, taking Github and AWS S3 so much for granted these days...but one of things students really liked in a class I taught on using the Unix shell was the ability to automate the pushing of files to their school account folder and have it show up immediately at stanford.edu/~theirusername
It's still pretty cool in this day and age of PaaS, and well, social networks...but was much more revolutionary in the pre-MySpace days (and post-MySpace too, if you had something you wanted to post that deserved a little more gravitas than a social media wall post)
I remember when someone started running httpd on our university network, and the universal reaction was 'seriously? It's reading stuff out of our private home directories and broadcasting it to anyone who asks? WTF, get this thing turned off immediately!'.
\*nix home dirs aren't really "private". They usually default to world-readable, and in a university environment (at least one like mine) there was a culture where people expected friends to look at them.
> world readable ... there was a culture where people expected friends to look at them
Same here. The default user home from /etc/skel (or some equivalent) had a user-access-only area that we were told to put sensitive work (individually marked projects where we were explicitly expected/told not to collaborate on) and other such into. We could protect other directories a little more too if needed, or even open up bits of that one to "group" or "other" (at our own risk) and there were specific groups setup for explicitly collaborative projects (these days I would do away with that and control such collaboration through git or its ilk).
One of my most magnificently grand stupid ideas was to create a strange boondoggle thing that kind of reinvented Gopher combined with the web and linked data.
The exciting action starts at 1:30 if you want to see the outcome of software developed at great expense with really no useful purpose.
Ladies and gentlemen: "NeoGopher" (New Gopher)
Essentially it let you browse through linked lists. Each list item could link on to another list or to a web page. Each list was created by about 15 lines of JavaScript which could be loaded from any web server and the data came from any HTTP web API.
(turn the sound down, the words serve only to confuse)
Ahhh... so many meaningless words in the demo. Even I, who designed it, found it hard to explain what it was - today I would say "It's a linked list browser.", which is what Gopher was.
Of all the stupid ideas I have had (there have been many), re-inventing Gopher was probably the worst. What was I thinking?
At the end of 1993, gopher was much more popular than the web, gopherspace was bigger, had search engines like veronica and so forth. You would hear about the web, and download lynx, and connect to a dull University of Kanas web page, or perhaps CERN's web page. Gopher just had a lot more content, was searchable with veronica, was more enjoyable to browse etc.
Some of this behind the scenes stuff mattered, like University of Minnesota's announcement of licensing fees. But the release of the Netscape browser is really what changed things for a lot of people, with its inline images, live cameras like the Fishcam, and this sort of thing. Mosaic was a harbinger of it, but the release of the initial Netscape browser in 1994 is what really set things going, and then the release of Netscape 2 and then 3 in 1996. Then Microsoft release of their then-primitive browser with Windows 95 helped as well (as well as them including an IP stack in Windows 95 - Windows 3.1 needed things like Trumpet Winsock).
From the article, "After its creation in 1991 at the University of Minnesota, use of Gopher exploded."
I'm a Gopher. That is, I'm an alumnus of the University of Minnesota. My first Internet account was based on my alumni relationship with the university, which gave me access to their large dial-up modem pool, and a distribution of a software suite called Minuet (Minnesota Internet Users Essential Tool).[1] Minuet included FTP and other standard protocols, and also had a Gopher client built-in (of course).
I used Minuet to browse around the Internet for a few months after getting my account, and definitely used it at first to access Usenet newsgroups. On the Usenet newsgroup about homeschooling, I noticed that one frequent poster had a funny-looking string of symbols in his signature block, which I learned was the URL for his website about homeschooling, which is still alive and well.[2] So I downloaded a Web browser (finding advice about how to do that in one of the first popular books on using the Internet, borrowed from my friendly public library), which must have been an early version of Mosaic. Wow, wow, and wow. The integrated images in HTML documents are what won me over to preferring the Web to Gopherspace, as well as the nonhierarchical hypertext links. Following the URLs given in the library to some early cool websites was all I needed to get hooked. An early version of my own website about homeschooling[3] was on the Web by a year later, and I had my own ___domain for it a year or so after that. Now I live online, as you see here. Gopher got some people started in reading reference documents online, and then Web browsers finished that job and also got us talking in new discussion formats that have largely supplanted Usenet.
> new discussion formats that have largely supplanted Usenet.
I still wish that message boards/forums had a standard format and protocol, like email does, so that we could use any OS-native client apps to access them, like we do for email.
Apparently Usenet was too limited to support modern "necessities" like avatars, profiles, stickies, votes and polls, but surely they could be worked in as extensions or even an entirely new standard?
I used Gopher (and Archie) pre 1991, when I spent a year in the military. During the spring of 1992, on a visit to my university, I saw an early version of a web browser (Mosaic, if I'm not misremembering things) and was completely blown away. I did like Gopher and used it a lot, but in my opinion it could not compete with the new web browser.
I haven't really had that kind of experience after this. The web really was the thing and it has completely changed how we learn and work.
I think it wasn't just about mindshare. It's easy enough to emulate gopher with a website: it's just a list of links: <ol> and <a> to the rescue.
I still remember the lab where someone first showed me Mosaic. I was familiar with the Internet: FTP, Usenet, xtank, archie. I said, 'what's that?' He said, 'it's the World Wide Web!' I asked, 'what's that?' and he replied, 'it's like the Internet, but even bigger!'
There used to be an early version of this essay online -- not the one that opens the still-copyrighted _A thousand plateaus_. But there's a video where it's read to illustrative diagrams (an "illustrated audiobook"):
A few years ago, I needed an automated way to pull files into an Apple IIgs (well, emulated Apple IIgs), which has a TCP stack but not much in the way of TCP applications. So I wrote a gopher client. And server, for the host computer. Around the turn of the century, I (ab)used the finger protocol to copy files, so gopher was a step up in that regard.
But I think the author misses the real killer feature of the web. The difference in hypertext models (external vs internal) is mentioned, but glossed over. Gopher was stuck with hierarchical indexes. It suffered from all of the limitations of hierarchy and had no way to acknowledge the relations between things. And documents in Gopher were dead ends. The web's ability to link from anywhere within any document to any other (and even locations within it) is a huge differentiator.
That was critical - instead of a limited hierarchy leading you down to a dead-end, you had an unlimited neverending web of related ideas. Imagine the C2 Wiki or Wikipedia in gopher format - all the same content but with no links. It wouldn't have anywhere near as much value.
The simplicity, cleanliness, and order of Gopher has beauty, but the much messier web has real-world value. The network effects of linking would naturally cause it to grow exponentially faster.