Dylan: Word from Apple
Net Bopper
------
Date: Mon, 28 Aug 1995 10:52:43 -0700
Errors-To: ser...@abs.apple.com
Reply-To: HAR...@applelink.apple.com
Originator: semp...@abs.apple.com
Sender: semp...@abs.apple.com
Precedence: junk
From: HAR...@applelink.apple.com (Harding, Pat)
To: Subscribers to <semp...@abs.apple.com>
Subject: Re: Dylan Update from Apple
X-Listprocessor-Version: 6.0 -- ListProcessor by Anastasios Kotsikonas
Given the recent semper.fi traffic regarding Dylan, Apple's Developer
Products Group would like to provide a status report.
A team, including members of the Cambridge Lab, is actively working to define
the future of Apple Dylan. These plans are still not firm enough for
full public disclosure, although DPG aims to disclose details soon.
As announced at WWDC, Apple Dylan will be made available as a technology
release by the end of 1995. This effort is on track, due to a tremendous
engineering effort by the team in Cambridge. Concurrently, marketing is
defining the packaging, pricing and distribution. The goals of this release
are the following:
o Make Dylan, broadly available to Macintosh developers for experimentation
and prototyping, thus building the necessary infrastructure for Dylan
adoption.
o Gain further feedback and insight from the developer community.
The Apple Dylan technology release will include these components:
- Dylan compiler and runtime
- Integrated development environment featuring incremental development and
advanced configurable browsing and viewing of code
- Dylan application framework
- Dylan user-interface builder
- Cross-language support allowing seamless access to existing C code and APIs
The technology release will be hosted on 68K-based Macintosh systems, and will
generate code for both the 680x0 and Power Macintosh systems. In addition to
PowerPC code generation, the release will feature greatly improved code
optimization (largely due to sealed classes), and significant bug fixes and
general cleanup. It looks like the compile speed will improve as well.
Specific details regarding where and how to order the Apple Dylan technology
release will be announced on the Dylan mailing lists, the Dylan WWW page, and
through Apple's standard channels later this year.
As discussed during WWDC, this release will be made broadly available, but long
term support will be limited. Based on strong customer feedback, we concluded
that the current implementation doesn't meet critical requirements. Since the
future success of Dylan depends on how widely it will be used by developers on
multiple platforms, our goal is a portable, licensable implementation, largely
written in Dylan itself. We will leverage our existing work significantly and
will leverage the other key players in the Dylan community: CMU and Harlequin.
We will locate much of the future Dylan work here in Cupertino. This will
assure that the Dylan development effort is in synch with other key
developments at Apple (e.g. Copland, OpenDoc, the OpenDoc Development
Framework, etc.) and that the whole Apple community, including management is in
close contact. Apple is working hard to retain current Dylan team members, to
work on both Dylan and non-Dylan projects.
We are pleased to announce that we recently hired Patrick Beard, the individual
who ported CMU's Mindy byte-code Dylan compiler, and Marlais, CMU's Dylan
interpreter, to the Macintosh. Patrick has implemented MacMindy as a drop-in
compiler for Symantec's 8.0 development environment; we look forward to having
this implementation broadly available very soon. We are also looking at the
Dylan framework and the User Interface Builder as key components of the future
Dylan product. We are evaluating compiler technologies, synergies between the
Dylan framework and other Apple frameworks and many other implementation issues
key to the final product. And last but not least we actively recruiting a
Product Marketing Manager for Dylan.
Apple is committed to Dylan as a technology but we still have many difficult
decisions to make as we go forward. We will work to keep all of you informed
of our plans and progress.
Pat Harding Peter Christy
Senior Marketing Manager Senior Director
Apple Developer Tools Apple Developer Products
-------------------------------------------------------------------------------
Additional sources of information about Dylan and future announcements
regarding its availability are:
o On the Internet, the Apple Dylan WWW (World-Wide Web) site is
http://www.cambridge.apple.com. The Apple Dylan ftp site is
cambridge.apple.com:/pub/dylan/
o On the Dylan WWW page at Carnegie Mellon University is
http://legend.gwydion.cs.cmu.edu:8001/dylan/
o On the Dylan newsgroup is comp.lang.dylan
o On Applelink, Dylan files are regularly uploaded to:
Developer Support:Developer Services:
Development Platforms:Dylan Related
Scott Fahlman
> Given the recent semper.fi traffic regarding Dylan, Apple's Developer
> Products Group would like to provide a status report...
>
> We are pleased to announce that we recently hired Patrick Beard, the individual
> who ported CMU's Mindy byte-code Dylan compiler, and Marlais, CMU's Dylan
> interpreter, to the Macintosh.
While we admire the Marlais effort, we probably should point out that
CMU had nothing to do with it. It is a nice complement to Mindy,
since Mindy is not very "dynamic" and Marlais is more so.
-- Scott
Robert A. Lentz
Net Bopper <gl...@netcom.com> wrote:
>Here is the full text of a message posted to the Apple Developers mailing list
I just hope they can get a PowerMac hosted version out soon too. I really
would like to be using Dylan to start programming on the Mac. I have been
looking forward to Dylan as a way to avoid C++.
-Robert
--
r-l...@nwu.edu http://www.astro.nwu.edu/lentz/plan.html
"You have to push as hard as the age that pushes against you."
-Flannery O'Connor
Marco Antoniotti
From: le...@annie.astro.nwu.edu (Robert A. Lentz)
Newsgroups: comp.lang.dylan
Date: 30 Aug 1995 15:54:29 GMT
Organization: Northwestern University, Evanston, IL
Lines: 18
References: <gleep-29089...@192.0.2.1>
NNTP-Posting-Host: annie.astro.nwu.edu
In article <gleep-29089...@192.0.2.1>,
Net Bopper <gl...@netcom.com> wrote:
>Here is the full text of a message posted to the Apple Developers mailing list
>...
>The technology release will be hosted on 68K-based Macintosh systems
>...
I just hope they can get a PowerMac hosted version out soon too. I really
would like to be using Dylan to start programming on the Mac. I have been
looking forward to Dylan as a way to avoid C++.
Am I wrong or I did hear that Apple was actually using MCL to develop
Dylan?
--
Marco G. Antoniotti - Resistente Umano
-------------------------------------------------------------------------------
Robotics Lab | room: 1220 - tel. #: (212) 998 3370
Courant Institute NYU | e-mail: mar...@cs.nyu.edu
| WWW: http://found.cs.nyu.edu/marcoxa
...e` la semplicita` che e` difficile a farsi.
...it is simplicity that is difficult to make.
Bertholdt Brecht
Patrick C. Beard
In article <420npv$p...@cantaloupe.srv.cs.cmu.edu> Scott Fahlman,
s...@CS.CMU.EDU writes:
>While we admire the Marlais effort, we probably should point out that
>CMU had nothing to do with it. It is a nice complement to Mindy,
>since Mindy is not very "dynamic" and Marlais is more so.
For the record, Marlais was originally written by Brent Benson, known
recently
for his excellent library for embedding Scheme interpreters in C
programs. Joseph
Wilson and I have hacked on Marlais a bit, but the code still owes a lot
to
Brent.
I am told that Marlais got its name because Dylan and Thomas were already
taken,
and Marlais is Dylan Thomas' middle name.
--
// Patrick C. Beard
// Core Tools Group, Apple Computer
// be...@apple.com, http://www.bdt.com/home/beard/
Joseph N. Wilson
In a previous message, "Patrick C. Beard" <be...@apple.com> wrote
[...]
> For the record, Marlais was originally written by Brent Benson, known
> recently
> for his excellent library for embedding Scheme interpreters in C
> programs. Joseph
> Wilson and I have hacked on Marlais a bit, but the code still owes a lot
> to
> Brent.
This summary is perhaps somewhat misleading. To say that "the code
still owes a lot to Brent" is certainly true. The Marlais interpreter
was initially develped by Brent Benson. Versions through 0.3 were his
work. Marlais-0.3 consists of 9811 line of C code that provide a core
model of evaluation for prefix Dylan and a protocol for extending the
interpreter.
To characterize the contributions of others, however, by saying
they've "hacked on Marlais a bit" is extremely misleading. I used
Marlais-0.3 as one of the language vehicles for a class in OOP I
taught at the University of Florida in spring 1994. I found that the
interpreter had various omissions and a few errors, so I contacted
Brent about modifying it. He was enthusiastic about the possibility
that someone might "breathe new life" into the project which he
characterized as something of a few-weekend hack. I produced Version
0.4, which fixed the method specificity checking and added the
behavior required by many of the Design Notes. Steve Strassmann of
Apple encouraged (prodded) me to produce a version of the interpreter
for the nex infix syntax. I thought I'd be able to convince a student
to do the project but, unable to find an appropriate victim in the
period of time I allocated myself, I went ahead and developed an infix
front-end to the prefix interpreter core. I used the grammar of Paul
Haahr as a starting point, modifying it to work with a LALR parser
generator and generating prefix code corresponding to the infix input,
then directly evaluating the prefix code. This allowed the
possibility of supporting both prefix and infix syntax. Around that
time, Patrick Beard got involved in the project. He developed a Mac
version of the interpreter with appropriate modifications and provided
much feedback to the project. Patrick's most valuable contribution to
the base implementation has been the development of the tail-call
strategy that is used in the interpreter. He added big integers to
the Mac version, however, I've been hesitant to incorporate that work
into the base implementation because it requires the use of C++ rather
than C and I'd like to preserve the high transportability of the
interpreter. Much of Patrick's work hasn't migrated back into the
original source tree, particularly changes providing a Mac Application
interface.
It is true that Brent's few weekends hack has turned into a many
weekends hack by others, but I'd like to point out that the
contributions those others have made are significant and worthy. I'm
not just some minor dude who rearranged some statements and while not
all of Patrick's work has migrated back to the original source, he
hasn't just been pointing and clicking at random. Many others have
provided code, bug reports, ideas, etc. and I've tried to acknowledge
them all in the ACKNOWLEDGMENTS file that goes with the distribution.
Please consult that for a relatively complete list of the
contributors.
Version 0.5.11 of Marlais consists of 25898 line of C code and has
tracked most of the significant developments in the language. There's
no doubt that it's been hacked on a bit. Brent's (heavily modified)
interpreter is still the heart of the implementation.
Having said this, however, I would like to point out that the Marlais
interpreter does not reflect a "major" implementation effort as one
writer characterized it recently. I take no umbrage if Marlais is
referred to as a "toy" implementation because I know that the
interpreter is a result of the freely donated work of a small group of
people and does not compare in scope of effort to the other
implementations that are available. There has never (to my knowledge)
been any funding for the effort whatsoever, and I have no reason to
believe that someone is waiting, ready to pour money into the project.
Thus, Marlais is likely to remain the work of a few committed
individuals.
>
> I am told that Marlais got its name because Dylan and Thomas were already
> taken,
> and Marlais is Dylan Thomas' middle name.
Marlais is also the name of a small village in France. This makes me
wonder if Dylan's parents gave him the name in reference to the
village. I have yet to find a source citing the origin of his name.
>
> --
> // Patrick C. Beard
> // Core Tools Group, Apple Computer
> // be...@apple.com, http://www.bdt.com/home/beard/
Joe Wilson
<A HREF="http://www.cis.ufl.edu/~jnw/">j...@cis.ufl.edu</A>