SoylentNews is people
SoylentNews
Right before HTTP took off in the early 1990's, there was Gopher and for a while it, too, was growing exponentially. It was fast and hosted text, source code, graphics, and any number of other types of files, just not all mixed together in one and the same document. For a while it was winning out over HTTP and making grounds against FTP. But that changed eventually and the rest is history. The MinnPost goes a bit into the history of Gopher with the Rise and Fall of the Gopher Protocol.
This discussion has been archived.
No new comments can be posted.
The Rise and Fall of the Gopher Protocol
|
Log In/Create an Account
| Top
| 34 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Grelb's Reminder:
Eighty percent of all people consider themselves to be above
average drivers.
(Score: 0) by Anonymous Coward on Saturday August 13 2016, @12:05PM
Actually, the reason for his success was the UNIX vendors considered workspace OS too niche to get really invested in them, and everybody else wasn't good enough yet, or not marketable enough. This was also one of the reasons Linus wrote his OS -- a machine supplied with an installed UNIX was about 7000$, and the bare metal cost about 2000$. Billy G was either extremely good at predicting the future, or just plain old lucky. His products were the most cost-effective for the time, which allowed market dominance. Once achieved, market dominance is not surprisingly easy to maintain: drivers for devices are written for You by vendors because users demand it. Similarly, software is written because of demand, and more software attracts more users. This allows the OS to become a standard, which demands compliance from others.
One could similarly argue for the success of Linux in contrast to Hurd. It's not about what's best, but what becomes popular fastest -- and that one becomes the best.
(Score: 0) by Anonymous Coward on Saturday August 13 2016, @02:18PM
The problem with it is that even after many years you still can't build the system on itself. From time to time I give it a try. Last check was a year ago, one of the most important software for it (glibc, which is even GNU software) can't be build on GNU/Hurd without using external patches (which are hard to find in all the outdated documentation).
(Score: 0) by Anonymous Coward on Saturday August 13 2016, @07:05PM
It's GNU Hurd and it's GNU/Linux because GNU Hurd is genuine GNU software and Linux is not GNU.
The reason the Hurd has problems after all these years is all the people who could have worked on it chose Linux instead, and the work never got done.
(Score: 2) by Marand on Saturday August 13 2016, @10:36PM
The reason the Hurd has problems after all these years is all the people who could have worked on it chose Linux instead, and the work never got done.
That's a popular excuse, but according to Hurd's own history page [gnu.org], work on choosing or creating a GNU-approved kernel started back in 1986. A barely functional Linux didn't appear until late 1991, five years later. In that time, it went from "we'll use TRIX" to BSD 4.4-Lite [groklaw.net] before RMS decided Mach was the only option. (Also of note is that trend continued after Linux's existence, as well, with multiple attempts at using different kernels in place of Mach over the years since.)
It's easy to blame Linux, but if the GNU folks had just chosen something, anything, and started working instead of five years of design-by-committee, they probably would have been the ones with the developer effort and Linux would be the also-ran. Someone [wikipedia.org] made a kernel from ground-up in a fraction of the time it took the GNU folks to just make a decision and start hacking.
(Score: 0) by Anonymous Coward on Sunday August 14 2016, @03:03AM
His products were the most cost-effective for the time, which allowed market dominance
This was more true than anyone here cares to admit. At the time I could outfit a Mac dev for 25-30k. Full stack dev env. Apple thought a lot of its SDK, it wasnt bad but not terribly great ether. Unix (Sun, SGI, HP)? 30k+ usually with some sort of large yearly license attached and the SDK is a bit obtuse sometimes. Linux was nowhere near ready. BSD was still quasi legal and no one was really sure if you could use it or not, but free and basically HW cost. No one wanted to chance it. A Win/DOS box 3k tops. Full stack including MSDN and dev tools and a semi decent SDK that was fairly easy to understand. Everyone and their brother had one so the market was huge. That is all in 1995 dollars.
It has totally inverted. Linux and BSD are very free and cheap to setup. You can get a dev up and running for less than 200-500 bucks and they can be very productive. 2kish if you want them to have a mac, but not worth quibbling over as it long term is peanuts. Many devs use macs because they can jump between all the envs easily that they need to, and Apple cleaned up its dev env to actually be nice. Its also why the dev tools on MS are basically free now. Though they screwed up and yanked the trial OS software for devs.