Datasheet

GNU

❘

GNU

As mentioned previously, the GNU project was started in 1983 as a response to the closed source

software that was by then being distributed by most computer manufacturers along with their hard-

ware. Previously, there had generally been a community that would share source code among users,

such that if anyone felt that an improvement could be made, they were free to ﬁ x the code to work

as they would like. This hadn’t been enshrined in any legally binding paperwork; it was simply the

culture in which developers naturally operated. If someone expressed an interest in a piece of soft-

ware, why would you not give him a copy of it (usually in source code form, so that he could modify

it to work on his system? Very few installations at the time were sufﬁ ciently similar to assume that a

binary compiled on one machine would run on another). As Stallman likes to point out, “Sharing of

software…is as old as computers, just as sharing of recipes is as old as cooking.”

Stallman had been working on the Incompatible Timesharing System (ITS) with other developers at

MIT through the 1970s and early 1980s. As that generation of hardware died out, newer hardware

came out, and — as the industry was developing and adding features — these new machines came

with bespoke operating systems. Operating systems, at the time, were usually very hardware-speciﬁ c,

so ITS and CTSS died as the hardware they ran on were replaced by newer designs.

ITS was a pun on IBM’s Compatible Time Sharing System (CTSS), which was

also developed at MIT around the same time. The “C” in CTSS highlighted the

fact that it was somewhat compatible with older IBM mainframes. By including

“Incompatible” in its name, ITS gloried in its rebellious incompatibility.

Stallman’s turning point occurred when he wanted to ﬁ x a printer driver, such that when the printer

jammed (which it often did), it would alert the user who had submitted the job, so that she could

ﬁ x the jam. The printer would then be available for everyone else to use. The user whose job had

jammed the printer wouldn’t get her output until the problem was ﬁ xed, but the users who had sub-

mitted subsequent jobs would have to wait even longer. The frustration of submitting a print job,

then waiting a few hours (printers were much slower then), only to discover that the printer had

already stalled before you had even submitted your own print job, was too much for the users at

MIT, so Stallman wanted to ﬁ x the code. He didn’t expect the original developers to work on this

particular feature for him; he was happy to make the changes himself, so he asked the developers for

a copy of the source code. He was refused, as the driver software contained proprietary information

about how the printer worked, which could be valuable competitive information to other printer

manufacturers.

What offended Stallman was not the feature itself, it was that one developer was refusing to share

code with another developer. That attitude was foreign to Stallman, who had taken sharing of code

for granted until that stage. The problem was that the software — in particular the printer driver —

was not as free (it didn’t convey the same freedoms) as previous operating systems that Stallman had

worked with. This problem prevailed across the industry; it was not speciﬁ c to one particular plat-

form, so changing hardware would not ﬁ x the problem.

Free Software, Free Society, 2002, Chapter 1. ISBN 1-882114-98-1

024485c01.indd 7024485c01.indd 7 7/26/11 8:31:58 AM7/26/11 8:31:58 AM