[ih] "How Gopher Nearly Won the Internet" Re: The Rise and Fall of the Gopher Protocol

[Jack Haverty]

Yes, that might explain why there was a CGI at all; perhaps it was 
necessary in Tim's original vision as a means to make changes to 
documents, and then NCSA saw a way to use that CGI for accessing 
supercomputer data and created a browser to do it?  All speculation on 
my part; it would be good to hear from more of the people who were 
directly involved about what, and why, they did what they did...

IMHO, the History of The Internet is full of situations where a piece of 
technology was developed with one use in mind, and then adapted by 
someone else to use in a very different way.

The Internet was built by McGyvers, rather than by committees with piles 
of documents.

/Jack


On 09/08/2016 01:11 PM, John Day wrote:
> One quarter we haven’t heard from are those who actually created it:  the web or the browser.
>
> One of the reasons NCSA was interested in creating a browser was that they were interested in the visualization/management of data to and from supercomputing problems. To some extent, the browser was intended to address their data management and visualization needs.
>
>
>> On Sep 8, 2016, at 15:30, Jack Haverty <jack at 3kitty.org> wrote:
>>
>> IIRC, the availability of a browser was necessary, but just one part of
>> the story.
>>
>> If you put a web server online, it wasn't very useful unless there was a
>> way to use it.   Yes, you could connect to port 80 and type "GET xxx"
>> but even techie nerds would find that tedious.   So, when I think of the
>> early web, it includes the browser client as well as the server as
>> critical elements.
>>
>> Gopher provided a slightly more friendly user interface than
>> tried-and-true FTP, and an index/search capability, but it was still
>> text-based.   That was fine for us techie nerds, but it wasn't exciting
>> to all the non-techie users sitting in front of PCs and used to
>> somewhat-graphical applications.
>>
>> Besides, we techies had all become quite comfortable using FTP, and the
>> "README" convention as a poor-man's index, for a decade or two.  By the
>> 90s NFS/SMB/Netware/Appletalk/etc made it much easier to have a "public"
>> repository (although AFAIK mostly on intranets) seamlessly integrated
>> into our computing environments.  Sales stuff was on the "SALES" server,
>> technical info on "ENGRNG", etc.  You didn't even know you were using
>> the network.
>>
>> In the broader world, even non-tech businesses had similar internal
>> systems, often a hodgepodge of proprietary environments.  They also
>> typically had "document management" systems, which kept track of
>> corporate documents and usually included capabilities beyond
>> index/search/retrieval.  In a business environment, you often need
>> mechanisms like review/approval, distribution lists, version management,
>> etc.
>>
>> So, as a document search/retrieval system, in the early 90s, Gopher was
>> interesting, but at best just an incremental improvement to existing
>> mechanisms, with no obvious way to integrate it with those mechanisms.
>>
>> -----
>>
>> As I recall, Tim B-L's original idea for the web was a sort of
>> collaborative notebook, where people could create and change documents,
>> and link them together, as well as read them.  There was no "search" per
>> se.
>>
>> But the initial thrust of the browser was to view the web as a read-only
>> system - you could "browse" an existing melange of documents, and set
>> bookmarks to help find them again.  No way (that I could see) in the
>> browser to make any changes to documents, new documents, etc.
>>
>> So the first days of the Web started with slightly-dynamic documents
>> (Tim's collaborative notebook vision), and with the advent of the
>> browser regressed to being primarily a static document repository
>> (browsers and servers, with documents created externally).
>>
>> People quickly started producing indices of what was on the Web.  When I
>> first encountered Yahoo, it advertised itself as "Yet Another
>> Hierarchical Officious Organizer" and was essentially a Gopher-like
>> index tree of interesting websites, one of many.  For a while, I
>> maintained a similar in-house list of interesting websites out on the
>> public Internet, and made it available within the company, along with
>> instructions on how to get browsers and servers.   As people in various
>> departments inside the company put up their own webservers, my list of
>> interesting Internet pages was expanded to include a list of interesting
>> intranet pages.   Search engines like AltaVista came along too.
>>
>> All of this could probably have been done with Gopher.   But it wasn't.
>>
>> The ease of getting the Web server and browser running, and of creating
>> simple linked-document internal corporate webs, and the ability to
>> create user-friendly documents with rich linking and formatting, all
>> helped push the Web forward.
>>
>> Gopher (IIRC) provided none of these, and no obvious way to integrate
>> Gopher into any of the existing systems.  At least I didn't run across
>> anyone who had done it.
>>
>> -----
>>
>> None of this was "mainstream" technology at the time; no one was using
>> the web as a primary component of their business IT machinery.   I think
>> the CGI was at least one of the primary elements that moved the web to
>> dominance.
>>
>> At the time, Microsoft had a catchphrase "Where Do You Want To Go
>> Today?" or something like that.  It reflected the "browser" perspective
>> of the web as a library or document repository. It was interesting to
>> browse the virtual aisles, find an interesting thing to read, and go to
>> investigate whatever you found there.
>>
>> I gave a lot of talks/demos using the next-generation catchphrase "What
>> Do You Want To Do Today?" as a theme.  The CGI made that possible.
>> Instead of just browsing documents, you could have the server do
>> something, and create a document on the fly to show you the results of
>> whatever it did.   That document could have links that did something
>> else, and forms that allowed the user to pass parameters in to guide
>> that activity.
>>
>> Instead of just browsing, a user could now actually cause that remote
>> server to do stuff.  Essentially, the Web browser, with CGI, had become
>> a new universal GUI - graphical user interface, with a new API that
>> could be linked to any kind of program behind any server on any kind of
>> computer -- in effect a universal standard RPC mechanism that could be
>> used to access programs on any kind of computer that you could get a web
>> server on.
>>
>> Of course, some of this had been done ages before -- e.g., the Coke
>> machine on the ARPANET.  But in the 90s the WWW made it now easy with
>> the key components pre-installed, freely distributed, and mostly free
>> (IIRC, the early browser cost just $25 or something like that - a
>> bargain compared to typical software costs of the era for all those PCs
>> out there).
>>
>> In business, government, and other parts of the "real world", computers
>> are used to do stuff.  As techies, we understood that, but I think all
>> we really understood was using computers to write code, and maybe an
>> occasional document or two.  Telnet was good enough to get to that far
>> away computer.  FTP was good enough to move code and RFCs around.  Email
>> was good enough to hold endless debates until someone finally wrote the
>> code.
>>
>> Non-techies use computers for other things - order entry, market
>> research, inventory control, shipping tracking, customer support, etc.,
>> etc.   Much of that activity involves interacting with databases of some
>> kind.  So the web technology, integrated with databases through CGI,
>> suddenly provided a whole new way of doing all those activities, and
>> could be deployed without the massive disruptions often required to
>> introduce new systems.
>>
>> Security was also a concern of course.  With the web technology,
>> businesses, their customers, and their suppliers could all be
>> interconnected through carefully structured web portals into their
>> various internal business computers.   Their various intranets would be
>> connected and even use the Internet to do so, but without the risk and
>> exposure of IP-level router interconnections.
>>
>> Done carefully, web servers act as firewalls so that users could only do
>> what they should be able to do.  A web server is essentially a "router"
>> interconnecting two organizations or customers, but the "connectivity"
>> is not for IP packets.  Rather, connectivity is limited to a set of
>> activities that can be performed through that connection.
>>
>> I can order a book from Amazon through the web, but I can't cause Amazon
>> to order 10,000 copies of a book I wrote - assuming their webserver
>> designer didn't make that function accessible through a public CGI.
>>
>> I can interact with a computer deep within a corporation, but I can (if
>> they're careful) only do certain things that they want me to do.  I most
>> likely can't "ping" that computer, or open an FTP connection to it.  But
>> I can use it.
>>
>> The Web has enabled an internet of internets, with computers able to
>> interact across public and corporate boundaries.  "On the Internet" has
>> a new meaning (Dave, time for an update to that RFC...)
>>
>> The CGI made that feasible.  As I explained this to customers - usually
>> non-techie managers in non-tech companies - you could see the "light
>> bulbs" as they realized how the web technology could be used in their
>> business interactions - not just to browse and read documents but to do
>> the various activities involved in commerce, internally as well as with
>> their customers and business partners.   That was a very potent attraction.
>>
>> -----
>>
>> I imagine all of the above could have been done with a continuous
>> evolution of Gopher.  But it wasn't.  The WWW "won" because people took
>> the base technology of the early web, figured out how to use it in their
>> own world, and just did it.   Rough consensus and running code...!
>>
>> The 90s were interesting times....
>> /Jack Haverty
>>
>>
>>
>>
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
>