Accusations, accusations, always accusations
anicolao at cgl.uwaterloo.ca
Thu Oct 21 13:45:27 UTC 1999
Richard Stallman wrote:
> I find the two quite similar actually. Cygwin32 is gcc & company, bash,
> and a standard set of file utilities (ls, tar, ...). From the user's
> perspective, it transforms their NT system into a system where the
> shells look and act just like UNIX shells.
> You're talking about "the user's perspective", what the system
> looks like. I'm talking about what it is made of.
> Judging from your description, adding Cygwin32 to Windows makes is a
> system that looks somewhat like GNU/Linux (or equally, somewhat like
> Unix). But it still has all of Windows in it.
No, I'm talking about what Cygwin32 is made of, not what's in the whole
operating system. And Cygwin32 is made of a porting layer that supports
UNIX-style tools on top of the win32 API. All of the UNIX style tools
shipping with cygwin are GNU tools: it is the GNU sytem sitting on top
of the porting layer to make it run on win32. Now, let's look at the
name: Cyg-Win32. So, the Win32 part is the Win32 API - not exactly the
windows kernel but widely considered the standard windows programming
interface (I'll ignore the fact that it is not a real API since every
version of windows subsets it: it is, after all, from Microsoft and
can't be expected to be a high quality specification). So win32 gets
credit. Cygnus gets credit for extending win32. Where's the credit for
GNU? All things considered, it should be called GNUwin32, since what it
does is allow GNU programs to run on top of the win32 API. It does this
by providing the kernel interface that the GNU programs expect. Linux
does the exact same job for naked hardware: it provides the GNU system
programs with the right API so that they can run.
So, you see why I think it is so similar. If anything, there's a lot
less owrk in Cygwin32 than in Linux, so the GNU system portion is
proportionately much larger. It really is GNUwin32, or GNU/Cygwin32, but
it isn't called that.
Now, I don't think it needs to be called that. I think that it is
sufficient to give the GNU system credit in the software. But I think
that for both cases. So that is why I am puzzled that the proponents of
GNU/Linux don't insist that it is called GNU/Cygwin32: the two are very,
very similar. The only real difference is that in the cygwin case
there's also windows, but that's not a problem since the win32
developers have already got their credit - right in the name of the
> While doing this work, one can't help but
> notice that Cygwin32 is effectively a method for porting the GNU system
> onto the Win32 subsystem.
> Well, you might be able to do that. And if you deleted all the
> Windows DLLs and graphical applications, you might end up with
> something which is the GNU system plus the kernel of Windows.
> That would justify the name GNU/<kernel-of-Windows>. (I don't know
> the name of the kernel of Windows; I have never been a Windows user.)
> I would rather use GNU/Linux, since Linux is free software.
Sure. I would too - in fact right now none of my machines boot windows
(although I am typing this on my Solaris box, and it is running on my
Linux box). But what we'd rather run isn't the point we're debating -
the point is what we call what we'd rather run.
You agree that if you threw out most of windows you could call it
GNU/<kernel-of-windows>. Although it is inaccurate, since win32 isn't
the kernel but only its API, this sounds a lot like GNU/win32. But it's
called Cygwin32, and it seems like Cyg has supplanted GNU. Perhaps Cyg
came from the acronym Cyg's Your GNU? Although it's cute, I'm sure Cyg
came from the fact that Cygnuswin32 sounded silly, but Cygwin32 has a
ring to it...let's call it that.
> > I will ask someone to tell me what CodeFusion is. I have heard a few
> > people mention it recently, but no one has described it to me.
> As I understand it, CodeFusion is a GUI that wrappers the GNUPro tool
> suite (gcc, gdb, etc).
> Those programs, the compilation tools, are just a part of the GNU
> operating system. So this is quite different from the situation
> with GNU/Linux.
> As for whether they are treating us properly, giving credit properly,
> that depends on the details of what they say. I haven't seen it, so I
> can't judge. (It is a tangent which I think we need not go down.)
Well, I guess we must agree to disagree on this point. Even the code
fusion case seems similar to me - lack of credit for a portion of the
GNU system used - and it isn't tangential since my whole point is that
there's no effort to equitably persecute everyone who doesn't give
credit to the GNU system in the product name. The effort is all directed
at the single most popular packaging of open source software since its
inception. The fact that the criticism is biased in this way is exactly
the point I'm trying to make, by citing other examples that have
received no similar criticism.
> I guess the root of my confusion is that I just don't know how much
> credit is enough.
> How about giving the GNU Project equal billing with Linus?
> Since we did much more of the work, and had the overall vision
> too, surely it is not unreasonable to ask for that much.
This too is unfair. Wars are not won by commitee. Glamour attaches
itself to single individuals, not to groups. There are many others who
have given huge amounts of effort to the cause who get almost no
recognition: Donald Becker and Dave Hinds, for example. Linus and Alan
Cox are widely recognized as leaders of the effort, and as a result get
virtually 100% of the credit and glamour that goes with being leaders.
These people don't get the credit by going on TV, issuing ads, or
They get the credit because the media fastens on them as the driving
force behind this mysterious, new market phenomenon: that free software
can run 50% of the servers on the "information highway", and that maybe
Microsoft isn't the only source of software after all. The fact is that
if systmes could get the credit the Linux system would get credit; but
Linus gets it instead. If there is a case wehre the software system
should be held up and idolized, it is linux. It doesn't happen simply
because abstract concepts are not glamourous: individuals are. So it is
that Linus is trotted out by the press as the david out to slay the
Gatesian goliath. So it is that McNeally is shown pitted against Gates.
Man versus man is the interesting conflict; man as an idol is the
interesting subject to heap credit on; and software will never get equal
billing with a human being. (Substitude Woman if you wish in the
So, I would claim that the GNU project will never get equal billing with
Linus. An individual person - a major contributor like yourself, RMS,
could - but the abstract concept of the project as a whole simply isn't
the kind of entity that people glorify. So it is that when I think of
the greatness in computer science I think of Donald Knuth, James
Gosling, Richard Stallman, Brian Kernighan, Dennis Ritchie, and so on (I
don't mean to exclude anyone here, but nobody wants to read my list of
personal CS heroes). I don't think of TeX/MetaFont/Art of books,
Emacs/News/Java, GCC/GNU system/HURD, and UNIX/C. These software
packages are great achievements, but the greatness lies in the
individuals that produced the packages.
Projects like the GNU system will never command the respect and credit
that the single name Richard Stallman already commands - and by going
for credit for the system you simply erode the true credit that is
already your due, at least in my mind.
More information about the License-discuss