Visi-calc started the PC revolution.
I was using email by 1984, so whatever program that was was extremely influential.
Win-95 was a useful OS, but it was still based on DOS with a GUI running on top. But XP was a totally new OS (which, however, had to include code to run DOS-based programs). My son worked in the NT group (he programmed the interfaces with all the third party hardware). It really was a new beginning and I am going to go with that.
Which still could be the most influential if every other language basically started with “whatever we do, let’s not model it after COBOL!”
I just got to this post, DOH! And I thought my previous post was pretty funny, oh well.
Wow, my post really was unoriginal.
emacs, le duh
I was a computer science major at the University of Michigan, class of '79. They had “mini courses” in several languages, and I took SNOBOL, 360 Assembler, and LISP. I told my counselor I wanted to take the COBOL course but it wasn’t offered that semester. He said, “We list it, but we *never *offer it. Don’t take COBOL. If you really want to take COBOL, I think they teach in the business school. But don’t.”
Seems to be two camps forming here between those who measure “influence” by users affected, others by technical attainments. Sometimes we have a lucky merger, such as Mosaic, sometimes it’s more one than the other.
In the COBOL/FORTRAN debate, I see Shodan’s side - people who knew nothing about computers held their products in their hands every time they received corporate correspondence, this made possible as many businesses used COBOL.
But I understand the other side too - regardless of how many used it, COBOL was an evolutionary dead-end as far as languages go, and its long-term influence will eventually wane as its installed user base wanes. And people held FORTRAN-generated bills in their hands too.
I would put neither one @ #1, but when I do compile a list of the top programs (imho) from all these suggestions, the higher ranked language will be… well, just wait. 
And remember - since one of the “criteria” was “I used this one a lot in my life”, there are really no wrong answers here. Maybe more correct answers, but no wrong answers.
Except Windows ME.
I think the SABRE system, founded by American Airlines in 1960, and now handling 100,000 customer transactions every second, is worth a mention.
There are probably a fair number of these examples of vast corporate software “packages” which run complicated processes which we are completely unaware. Like the software needed to run a, say, an automobile factory probably hasn’t been mentioned here, but I would not be surprised if many auto manufacturers use variants of the same system.
With Unix kernel, utilities, shells and C compiler bundled together any other nomination seems like a joke! … (by my criterion).
When someone starts the Worst computer program ever thread, there’s an auto factory system I’ll nominate!
Probably worth a mention in its own right because of its importance to airline travel, but also, I believe that airline reservation systems like SABRE pioneered some of the basic concepts underlying packet switching.
I’m a little shaky on the details so take this with a dash or two of salt, but I believe that for many years SABRE used what was essentially a datagram protocol called Airline Line Control (ALC). “Datagram” means that information packets were sent independently and with no acknowledgments between the airline terminal equipment and the airline host systems, mainly because each packet contained a complete transaction and if it was lost, it would be obvious to the terminal operator and it could just be manually re-sent. Technically this would be called a connectionless protocol, whereas a connection-oriented one operated with additional handshaking over a session-based virtual circuit. Packet switching, which provided for both, allowed the efficient use of long-distance leased lines without having to have expensive physical circuits to each location, and later formed the basis of commercial packet-switching services based on the ISO X.25 packet-switching protocol, and indeed formed the basis of the network layer of all important future networking protocols. A network router is essentially a packet switch (although X.25 was a very specific paradigm that encompassed physical, data link, and network layers; SABRE itself eventually migrated its proprietary ALC layer over to X.25). So in some ways SABRE was the granddaddy of some important modern networking concepts.
SABRE was so important that the city of Buffalo eventually named its NHL hockey team after it.*
- I should be allowed to just make shit up every once in while.
When the vi people come charging in, I’ve got your back.
And of something even less obvious, that microprocessor and computer they use to run SABRE was almost certainly designed using EDA software written in C. And maybe C++ now.
Well, COBOL was nothing if not wordy. 
The mouse, or whatever software allows the mouse to work. Apparently, its inventor never recognized its potential and failed to copyright it.
There’s no particular single piece of software that allows a mouse to work. Not sure where you got this idea, I’m sure there are numerous copyrighted mouse drivers for each of the many different types of mouse that have been made. One early mouse used in the Mother of All Demos mentioned above was patented but the patent expired before the devices came into common use, and several different technologies have been used to implement the devices.
And you don’t do anything to get copyright other than write something down. The copyright is attached as soon as you write the code.
I was referring to this guy: Mouse inventor 'never made money' - BBC News
Perhaps it doesn’t qualify based on the OP criteria, since the invention is more hardware than software?
ETA: apologies, the article clearly mentions a patent and license
The local community college offered COBOL classes, but in its Data Processing department, which was entirely separate (and in a different building, on a different computer) from the CS department, that offered only FORTRAN.
When I went to Cal in the early 1980s, there was a list of what computers could handle what languages; COBOL, along with BASIC and ALGOL, were listed, but the answer was, “None.”
What was the first one that was a dedicated OS?
Windows 2, which I think was actually called “Windows 386,” was not only not standalone (you have to start it from a DOS prompt, and when it crashed, it returned you to the prompt), but it did not support color icons, and I think that was still when you couldn’t resize the windows (I am guessing because somebody was afraid of a lawsuit by Apple); you could have one large window, two side-by-side, two top-and-bottom, or four one-in-each-corner windows.
Is that still around? I think one reason SABRE was popular was, one of the first systems that people could use directly was a simplified, non-windowed version on
CompuServe called Eaasy SABRE (two As in “Eaasy,” for American Airlines). I think the first time I used it was in late 1991, to book a flight to Australia in January 1992.
I also remember that SABRE was still using the “fixed size windows” from pre-3.0 Windows even after versions of Windows where the windows could be resized and overlap had been released.
It’s amusing to remember the long-standing criticism of unix was that unix wasn’t a real operating system, was just a toy etc. because (mostly) it didn’t provide proper isolation and virtualization. It’s easy to see where the name-calling about Windows 95/98 came from
The problem is, you write that and then people think it means this:
Now, Hari Seldon goes on to write this
Which is true, and a valid contribution to the discussion
But it’s got that wrong bit in it: “DOS with a GUI running on top”.
Win 95, for all it’s problems, for all it’s calls into 16bit drivers, wasn’t “a GUI” and didn’t run “on top of” DOS.
And of course even though it provided a buggy virtual 16 bit machine for DOS programs, and a buggy virtual 32 bit machine with 16 bit memory allocation for Windows programs, even though it connected to 8 bit devices like the keyboard and some old disk drive controllers, it was fundamentally not DOS, not on top of DOS and fundamentally a new 32 bit OS that provided buggy virtualization of the DOS and Win3.x environments.
SABRE was enormously influential and successful when it was used only on dedicated terminals used only by professionals. If that was all it ever was, it would still be remembered for reforming the industry.