Where Tunes is going.
Thu, 21 Oct 1999 17:27:46 -0700
To all actually interested in the Tunes project:
This is not a question. This is the answer.
Tunes is not for coders. The Tunes philosophy does not support the needs
of C-programmers to continually re-write the same things over and over.
Anyone out there who thinks that when Tunes exists, that they will "code"
in it is drastically wrong, and I'll explain why:
Tunes is about unifying languages (even domain-specific ones), and
providing for automating this unification process. This still _requires_
human intervention, but the intervention required is not that of coding,
for the simple reason that Tunes is about obviating that need. Consider
what happens when you've fully formed your C or Lisp or Forth or Smalltalk
program. It's done, and it works. But without Tunes meta-programming
framework, how does that code get migrated? Changes outside the language
framework are _manual_. Lifting the idea from the code is _manual_. And
there's no place for a meta-framework that Tunes is in a language where you
must specify everything that the language evaluator needs. To counter
those of you who suggest that Tunes is about features, I disagree, because
ultimately you will either encode those features in some language, or you
will place it within Tunes. It is the ultimate chicken-before-the-egg
issue. If you want to support Tunes, you have to give up the language
idea, and that's why I am working on Arrow.
So where is Tunes going? Arrow is Tunes. Any who disagree are those who
read my words (as codes) and mistake my explanations for the idea.
Many of you believe that we must have an OS to boot Tunes from in order to
have Tunes. I agree on the fact that Tunes must make an OS framework (a la
OSKit) in order to extend its usefulness, but requiring an OS to be on hand
before we build Tunes is absurd! If the OS-code is not available as Tunes
objects, then it's not useful, and therefore not Tunes.
I want to offer the Arrow development to you. What it involves is
modifying a Lisp environment so that it supports Arrow ideas. This
involves some technical points which I will lay out in detail. There are a
little over 100 people who receive Tunes mail, and I believe that over the
past year, I've learned what nearly all of your goals are, via email or
irc. From the information I have gathered, what I offer is not what you
want. If you really want Tunes, you will be flexible enough to accept the
tools that I have found to work with. But I do not see this happening
without appealing to people outside the group.
My point is that _I_ have done the research, _I_ have learned the ideas,
and _I_ am working with the very necessary theory first-hand. My
disadvantage is that I am a one-person development team, as I have been for
six or seven years now. All aspects of Arrow (and therefore a good portion
of Tunes, I argue) are mine, and unless you do something for Tunes other
than write OS-code or your own programming languages (codes), then you are
not contributing to Tunes.
A common argument about Tunes is that it requires AI, yet your very efforts
completely avoid the necessary issues. You effectively "pass the buck" to
the research community at large, which is slow and cumbersome. Yet you all
have intelligence sufficient to stop following your silly language ideas
and learn beyond them, and Tunes *does* need that.
I have much more to say, but I'm sure your egos are sufficiently insulted
to listen to me now. You will either choose the future that Tunes must
have to succeed, or you will choose to continue to waste the world's time
with your pathetic lies about "what Tunes is".
Finally, I have a request for Tril and others concerned to make a concerted
effort to show *publicly* the results for Tunes that I am working on. The
Arrow introduction certainly needs some work, I admit. But that work has
been stalled because again _I_ am doing all the work. I need people who
will take these ideas and present them to the world via the web site, and
organize a *real* development effort to support Tunes.
I'm sure that I will receive some very un-educated responses, as well as a
few wise ones. Just keep in mind that I've done the work, and that you _as
coders_ do not have the answers. I have learned to take on humility when
it fits the situation, and this is time for you to do such. I have
listened to all your ideas for far too long with too much respect, and all
I have been is disappointed. Tunes needs no more "half-way" solutions.
I've picked the tools for their simplicity of use in various aspects, and
I've worked out a great deal of theory, particularly for the basic
Here is the line. Either cross it, and move on with Tunes into history, or
detract from it as you have been wont to do. I will accept nothing else.
I want Tunes, and I want it more badly than Fare or Tril or any of you do
by any stretch of the imagination. Deal with it.