A paper that is well worth reading in analyzing the adoption and nonadoption of systems is News, Bad News, and How to Win Big "Worse is Better." Its thesis is that while LISP provides extensive functionality, typical LISP-based attempts to provide "100% solutions" fail because the Unix principle of looking for the 90% solution provides something "good enough" more quickly.
(Note that the paper is sometimes difficult to locate; an alternative source is cbl.leeds.ac.uk.)
Richard Gabriel has further elaborated on this on his "Worse Is Better" page. He has waffled rather a lot, holding various contrary positions. The ultimate point may not be that Worse Truly Is Better, but rather that it is a constructive way of thinking about systems.
This may be further extended to the notion that Good Enough is Best. That essay, which unfortunately appears to have disappeared from the web, set forth the the further thesis that Windows NT is "good enough," and therefore will probably win the "OS Wars." This of course assumes that NT is in fact "good enough," which stirs up controversy as to precisely what it is "good enough" for. Common assertions go along the lines that the installation CDs are "good enough" to use as coffee coasters...
This essay was written by the legal counsel for the Free Software Foundation, and is extremely well-written.
Metaprogramming and free availability of sources is a (possibly not yet complete) paper that presents the thesis that further improvements in software are likely to be driven by metaprogramming, that is, by the use of tools that can analyze and manipulate programs. Source code availability is obviously pretty necessary for such analysis and manipulation, and furthermore, restrictive licensing will restrict the permissibility of doing this sort of thing. As a result, "open source" software can benefit from developments in metaprogramming, while "closed source" software is very restrictive in this regard.
From a Netscape employee comes Setting Up Shop: The Business of Open-Source Software
Linux et le developpement decentralise This document is a translation of an older version of my essay into French. I am very pleased with what I see of the quality of the translation.
Various things that have been learned thus far by those that have participated in the Debian project. Many principles that are useful to consider for anyone that wants to work on a substantial free software project.
The major point of his thesis is that IBM is licensing their software using arrangements that allow them to terminate essentially at will your ability to legally use the software. Only a seriously unwise person would use software on that basis...
There has been some controversy, possibly overstated, over the "openness" of Apple's Darwin source code release.
Problems of the Apple License (APSL) (Essay by RMS on the revised APSL)
The Ultimate Argument Against Linux - a very good essay; the final argument is quite compelling...
An essay that explains a probable "Microsoft perspective" on why they'd consider free software to be a bad idea.
IT News - Booch-speak: keep it simple, stupid - Grady Booch denigrating "free software"
Is Linux Ready for Delphi?, or, Linux? Are you people crazy?!?!! by Danny Thorpe, Delphi Research and Development
Arguably a somewhat sinister proposal...
Many software packages have initially been targeted at Linux, and
may be installed and used on other systems. The following is, to my
mind, exemplary of what should be done by those creating software "for
Pyrite - (was
PalmPython) (a package to allow connection
to 3Comp Pilots) requires Python and
pilot-link or newer. It is developed and
tested on an i486 running Debian GNU/Linux 2.0, but should compile on
any Unix-like system capable of running
pilot-link. It may be portable to other
platforms as well, and the author would appreciate compatibility
Writing portable software is a net benefit for the GNU system:
Portable software is more robust; testing on multiple systems exposes silly assumptions
Portable software is better encapsulated; e.g. the GUI is split from the main engine
lots of people using proprietary systems are good programmers and contribute useful patches. more audience means more code contributions.
automake or autoconf plus Unix standards mean it isn't especially hard to write software that works on most systems
If you refuse to accept portability patches, you'll get flamed and people will fork the project; this hurts development way more than supporting those systems does
A lot of GNU support comes from people using proprietary systems.
<email@example.com> Havoc Pennington
01 Oct 1998, plus some subsequent comments...
A discussion of the use of free software with Artificial Intelligence applications...
The Time Trading Development Technique - Controlling Time-Waste in Open Source Projects
I want feature X in some open source system. I offer 10 hours of my time for 1 hour of time from one of the really skilled developers.
Love him or hate him, he has been pushing to promote a community using free software for a long time now, with the attendant significant influence that results.
He is quite tremendously stubborn, which was quite necessary when people couldn't believe the GNU project could be a credible thing. But his positions sometimes cause people great frustration. There are a group of people that wax fairly much worshipful about him, much as there are people that disagree so vigorously that putting them together is an almost explosive thing. (Sort of like the way matter and antimatter annihilate one another...)
Love him or hate him, just about everyone has some kind of opinion about him...
"Pictorially showing the events in the competition for world domination between proprietary and free (open source) software"
An essay that pretty much takes the position that there is no moral principle surrounding it, and that it's pretty much about unrealistic, unAmerican, "pirates" trying to do their thing.
An essay on the notion of running "free software" on Windows. It may not be pretty, but you can do it...
How the "Linux games company" rose and fell. A lot of it related to business problems that hadn't the slightest to do with Linux or free software.