So I've been thinking a lot about cons cells and the relation between
the various mapping functions in Common Lisp. Poking around in the
Google archives of c.l.l. and in various Lisp manuals I have lying
around, I put together this chart:
Common Lisp | MAPLIST | MAPCON | MAPL | MAPCAR | MAPCAN | MAPC |
--------------------------------------------------------------------
Lisp 1.5 | MAPLIST | MAPCON | MAP | - | - | - |
Interlisp* | MAPLIST | MAPCON | MAP | MAPCAR | MAPCONC| MAPC |
ZetaLisp | MAPLIST | MAPCON | MAP | MAPCAR | MAPCAN | MAPC |
T | MAPCDR | - | - | MAP | - | - |
R5RS Scheme | - | - | - | MAP | - | - |
Based on these names I put together the following theory about the
history of these functions:
MAP (MAPL in common Lisp) was the primordial list mapping function.
Which is why it got the most basic name in Lisp 1.5 (and kept it up
until CLTL1 the Scheme dialects appropriated it for something else.)
MAPLIST came next and was so called because it combined mapping with
creating a list.
MAPCON was then a variant on MAPLIST.
Later on MAPCAR was introduced as sugar on top of MAPLIST.
Once we had MAPCAR, then simple consistency lead to the creation of
MAPC and MAPCAN/MAPCONC.
If anyone can fill in other rows in my chart (I couldn't find a
Maclisp manual, for instance) I'd be interested. Or if any one wants
to comment on or even refute my etymological theory I'm all ears.
-Peter
* The Interlisp mapping function took the list as their first argument
and an extra function argument to be used instead of CDR, if
supplied, to walk down the list.
--
Peter Seibel ·····@javamonkey.com
Lisp is the red pill. -- John Fraser, comp.lang.lisp
Peter Seibel <·····@javamonkey.com> writes:
[Whoops, let me correct the T row]:
> Common Lisp | MAPLIST | MAPCON | MAPL | MAPCAR | MAPCAN | MAPC |
> --------------------------------------------------------------------
T | MAPCDR | - | WALKCDR| MAP | - | WALK |
T also provided a MAP! function that was equivalent to this:
(defun map! (proc list)
(maplist #'(lambda (x) (setf (car x) (funcall proc (car x)))) list))
-Peter
--
Peter Seibel ·····@javamonkey.com
Lisp is the red pill. -- John Fraser, comp.lang.lisp
"Peter Seibel" <·····@javamonkey.com> wrote in message
···················@javamonkey.com...
> So I've been thinking a lot about cons cells and the relation between
> the various mapping functions in Common Lisp.
> If anyone can fill in other rows in my chart (I couldn't find a
> Maclisp manual, for instance) I'd be interested.
For Maclisp history, a good source is the archived directories from ITS at
ftp://ftp.its.os.org/. In particular, you'll want to look at the LSPMAI
directory (ftp://ftp.its.os.org/its/ai/lspmai.tgz) which contains all the
announcements and correspondence about MacLisp at MIT; the LSPDOC directory
(documentation) and the LSPMAN directory (the manual). I don't think these
go far enough back to tell you when the various map functions were
introduced, but they will tell you what was in Maclisp. There is also some
Lisp Machine documentation there. The dates on the files may or may not be
accurate.
So, looking at LSPMAN; MANUAL ASCII, internally dated March 3, 1979 (this is
the famous Moonual), the "mapping" section includes map, mapc, maplist,
mapcar, mapcon, and mapcan (and also mapatoms, but that's another story).
For earlier history, you can look at the A.I. Lab memos describing MIT Lisp
developments including PDP-6 Lisp (as it was called), which was the ancestor
of both MacLisp and BBNLisp/InterLisp. Memo AIM-98 = MAC-M-313 (Peter
Samson, June, 1966,
ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-098.pdf) lists
maplist, mapcar, map, and mapc (in that order, presumably because that was
the order in the interpreter source). Note that these documents are
archived as *images*, so you can't do text search on them.
I think it would be interesting to put together a chronological index to
developments in MIT Lisp based on these memos.
Stavros
"Stavros Macrakis" <················@verizon.net> writes:
> "Peter Seibel" <·····@javamonkey.com> wrote in message
> ···················@javamonkey.com...
> > So I've been thinking a lot about cons cells and the relation between
> > the various mapping functions in Common Lisp.
>
> > If anyone can fill in other rows in my chart (I couldn't find a
> > Maclisp manual, for instance) I'd be interested.
>
> For Maclisp history, a good source is the archived directories from ITS
I can't tell you how frustrating this is.
The directories you direct people at are my directories, not Moon's, and
are not public domain. They are copyright MIT (though this was due to a
misunderstanding about how copyright worked back then, and really I believe
they ought always to have been my own copyright). I have since re-acquired
the copyright and have a version almost webbed.
I like nothing less than having worked a long time trying to get something
into a product form that maybe I can charge just a little money for only to
find yet another person giving away something for free. Especially when it's
the sweat of my very own brow that someone just "happened to have a copy of".
I thought I was told that that directory was not going to be made publicly
available.
Geez, this burns me and makes me want to just leave the industry altogether.
I can't tell you how amazingly fuming I am about this.
I just give up.
Kent M Pitman <······@nhplace.com> wrote in message news:<···············@shell01.TheWorld.com>...
> I like nothing less than having worked a long time trying to get something
> into a product form that maybe I can charge just a little money for only to
> find yet another person giving away something for free. Especially when it's
> the sweat of my very own brow that someone just "happened to have a copy of".
>
> I thought I was told that that directory was not going to be made publicly
> available.
>
> Geez, this burns me and makes me want to just leave the industry altogether.
>
> I can't tell you how amazingly fuming I am about this.
>
> I just give up.
I don't know if a reply is appreciated here... but though I have
expressed sympathy to free stuff, clearly it is in no one's interest
for you to kill your projects. I hope you remember I emailed
wondering what you had coming down the pike. So:
- Maybe the first thing to do is send a letter to the site's owner to
shut down the dir.
- Part of me is a consumer, and take your words as a threat that
you're not interested in supplying the market I'm in. And you're
really confusing me. I would expect that you'd be happy to pass some
time explaining the context of certain parts of that archive if we
were subscribing; now if you're dropping this project, this clearly
ain't gonna happen. Soo... what should we think? I think you need to
respect the people who would gladly be your customers. It seems
you're wanting us to feel guilty/resigned...
- Wouldn't it be happier to preannounce your intentions, even by
qualifying it with "This isn't my top priority, but a release is
likely"? Then you have the advantage of failing fast, or getting
immediate feedback. Even a competitor would warn you of impending
release to stave off community backlash. I would be fine with a
special mailing list for this purpose.
I don't know. I'll just pay for your previous usenet explanations.
Set up an amazon thing. While they're greedy, I more/less trust them,
and I'll pay the extra 15% or whatever usury they charge. It's just
this temperamental artist thing you're doing is driving me and no
doubt others nuts. You're really underestimating us and the value of
what you produce.
On 15 Dec 2003 20:18:37 -0800, Tayssir John Gabbour wrote:
> I don't know. I'll just pay for your previous usenet explanations.
> Set up an amazon thing. While they're greedy, I more/less trust them,
> and I'll pay the extra 15% or whatever usury they charge.
I don't know what "an amazon thing" is, but if you want to pay people
for Usenet postings, etc., you could do worse than to use e-gold.
They charge 1%, with a maximum charge of US$0.50 (and you can make
very small payments, as well)
Setting up an account is free: http://www.e-gold.com/e-gold.asp?cid=1099203
--
You don't have to agree with me; you can be wrong if you want.
(setq reply-to
(concatenate 'string "Paul Foley " "<mycroft" '(··@) "actrix.gen.nz>"))
From: Joe Marshall
Subject: Re: Etymology of MAP[LIST,CAR,CON,CAN,C,L]
Date:
Message-ID: <r7z2o7q2.fsf@comcast.net>
Paul Foley <···@below.invalid> writes:
> On 15 Dec 2003 20:18:37 -0800, Tayssir John Gabbour wrote:
>
>> I don't know. I'll just pay for your previous usenet explanations.
>> Set up an amazon thing. While they're greedy, I more/less trust them,
>> and I'll pay the extra 15% or whatever usury they charge.
>
> I don't know what "an amazon thing" is, but if you want to pay people
> for Usenet postings, etc., you could do worse than to use e-gold.
I think there'd be a better market for paying people *not* to make
Usenet posts.
My rates are negotiable.
--
~jrm
>>>>> "Kent" == Kent M Pitman:
[...]
Kent> I like nothing less than having worked a long time trying to get
Kent> something into a product form that maybe I can charge just a
Kent> little money for only to find yet another person giving away
Kent> something for free. Especially when it's the sweat of my very
Kent> own brow that someone just "happened to have a copy of".
[...]
Something being available for free does not mean you can't charge money
for it. It just means you have to be a little more careful what you
charge for (e.g. there has to be real added economic or moral value tied
to payment, you can't just set up a monopoly and charge unreasonable
amounts of money because it's only you who has the data).
To give you some real-life examples: I bought the SICP book even though
it is available "for free" on the Web. I'm reading Peter Seibel's book
chapters now, and I will buy the book, regardless of whether he leaves
them online in the final version or not (in fact, I'd recommend that he
does).
So -- I wouldn't worry too much.
--J.
From: Stavros Macrakis
Subject: Copyright of 1979 MacLisp Manual (Was: Etymology of MAP[LIST,CAR,CON,CAN,C,L])
Date:
Message-ID: <NmK_b.3178$ZV1.3100@nwrdny01.gnilink.net>
"Stavros Macrakis" <················@verizon.net> writes:
> For Maclisp history, a good source is the archived directories from ITS
"Kent M Pitman" <······@nhplace.com> wrote in message
····················@shell01.TheWorld.com...
> I can't tell you how frustrating this is.
> The directories you direct people at are my directories, not Moon's, and
> are not public domain. They are copyright MIT (though this was due to a
> misunderstanding about how copyright worked back then, and really I
believe
> they ought always to have been my own copyright).
KMP, Lisp colleagues,
Sorry for my very slow response. I am not a habitual participant in
comp.lang.lisp.
I have a pretty good sense of the ethics involved in the old ITS
directories, though not a very clear sense of the legalities.
From the ethical point of view, I have no qualms about pointing people to
these archives of project directories (not personal directories). They were
always publically accessible to anyone on the Arpanet when ITS was up, and
the culture of the 8th and 9th floors of Tech Square (the AI Lab, Macsyma
group, and others) always encouraged looking at files in all directories
(though you were supposed to avoid things that were obviously personal -- a
little like those notices on fax messages saying "if you aren't who I intend
you to be, stop reading this"). Moreover, almost all the files on the MC
machine were written by people employed by MIT on sponsored research
projects. On the other hand, from the ethical point of view, I certainly
respect the *authorship* of files -- people should get credit for their
work. If Seibel or anyone else uses these files, he should acknowledge them
properly. I don't have a strong ethical sense about publishing these files,
but I doubt it matters. Making the papers available in a
publically-accessible archive seems perfectly reasonable to me. There are
also many unpublished papers available in the paper archives at M.I.T. whose
copyright probably remains with the authors, not with M.I.T. (e.g. Ph.D.
theses).
It appears that the LSPMAN directory has been taken off the archive --
perhaps at KMP's request. This is unfortunate. It was a useful research
resource, though I doubt it was very heavily used....
Now some factual details backing up the ethical analysis.
The directories I pointed to, LSPMAI and LSPMAN, were project directories,
not personal directories. The files that are relevant to Seibel's original
query are mostly LSPMAI;Lisp News/Recent/Oarchi and LSPMAN; MANUAL ASCII
(dated March 1979). The LSPMAI files were edited and largely written by
JonL White, with some contributions from Guy Steele, and perhaps a few
others, so I presume KMP is talking only about the Lisp manual directory.
The MacLisp manual (the Moonual) was originally written by Dave Moon. I do
not know which file that corresponds to. A "-read- -this-" file dated
January 1981 says that the files in the LSPMAN directory "belong primarily
to KMP's new manual", so I suppose that is what KMP is annoyed about.
As far as I know, JonL, Steele, Moon, and others have not objected to these
files being made available by Alan Bawden (another old Lisp hacker from that
group). JonL in fact wrote me "Tnx for the pointer" when I told him about
them (May 2003).
No one's name appears in the Manual ASCII file -- not Moon's, not Pitman's.
There is no copyright notice or other indication that it is not in the
public domain. Files on ITS were publically readable by all users, and in
fact by any user on the ARPANET, with no password protection. This was not
some sort of lapse in security; in fact, it was widely known and taken
advantage of (and in some sense was the inspiration for the FSF philosophy).
I believe that everyone who contributed to these files was employed by
M.I.T. as a staff member or research assistant on government or
industry-sponsored research grants (ARPA, DOE, Macsyma Consortium), and the
work on maintaining and documenting MacLisp was specifically covered as part
of these grants. I do not know if KMP in particular was paid by AI/LCS. I
do know that these grants gave rights to the granting organization; in fact,
the 1982 version of Macsyma (copyrighted by MIT) was distributed by the DOE,
and later DOE released Macsyma under the GPL. So I would think that DOE has
rights to other software and documentation written under these grants.
So much for the ethical analysis.
Now for the legal analysis.
In 1982, anything that was published without a copyright notice was
automatically in the public domain (this was before the US ratified the
Berne convention). The question is then whether making something available
on a widely-known, freely-available FTP address constitutes "publication".
I don't know. Legally, publication involves making a relatively unlimited
number of copies available to the general public. Distributing to a list of
known people, or asking people to sign for copies is probably not
publication. But in this case, the document was available by anonymous FTP
for many years. So it is not so far from publication. I don't know if
publication as a Technical Report is considered publication for copyright
purposes, and I don't know if the Technical Report version has a copyright
notice.
Next question: who is the author? Moon's manual was originally issued in
1974, and Pitman's was issued in 1983, so I suppose the 1979 version could
be a version of Moon's manual, or an early draft of Pitman's. This should
be easy to check, but I don't have either available right now.
Next question: if there is a copyright in this material, whose is it? It
was developed on equipment owned by MIT and operated under research grants.
I don't know if the author(s) were employed by MIT at the time, and I forget
what the status of "work for hire" copyrights was at the time. In any case,
whether it is MIT or the author who holds the copyright, MIT presumably has
rights to publish. After all, "The Revised MacLISP Manual" by Pitman was
published as an MIT/LCS technical report in 1983, and is in fact still
available from MIT (http://www.lcs.mit.edu/publications/specpub.php?id=863).
As for:
> I have since re-acquired the copyright and have a version almost webbed.
I don't know what you mean by "re-acquired the copyright". If the document
was once in the public domain, it is always in the public domain. But
perhaps you mean that you acquired publication rights from MIT?
So I don't see that there's a legal or ethical problem in making these files
available on a public FTP site (as they always were when hosted at MIT), and
I *certainly* don't see a legal or ethical problem in posting a link to
these files!
Now, as for:
> I like nothing less than having worked a long time trying to get something
> into a product form that maybe I can charge just a little money for only
to
> find yet another person giving away something for free.
I have nothing against charging money for things (I am not an FSF radical).
And even if the originals were in the public domain, or have an MIT
copyright, you can certainly copyright your additional work (though you
might also need an MIT license).
Sadly, I doubt that the 1983 MacLisp manual, no matter how well revised, has
much monetary value. But I wish you all the best.
Stavros Macrakis
········@alum.mit.edu
Macsyma Group, MIT, 1971-1977 (······@MC)
Developer on GPL Maxima, sourceforge.net
Developer on Gnu Common Lisp (GCL), gnu.org
From: Christopher C. Stacy
Subject: Re: Copyright of 1979 MacLisp Manual (Was: Etymology of MAP[LIST,CAR,CON,CAN,C,L])
Date:
Message-ID: <uptc4wfis.fsf@news.dtpq.com>
>>>>> On Tue, 24 Feb 2004 15:47:57 GMT, Stavros Macrakis ("Stavros") writes:
Stavros> No one's name appears in the Manual ASCII file -- not Moon's, not Pitman's.
Stavros> There is no copyright notice or other indication that it is not in the
Stavros> public domain. Files on ITS were publically readable by all users, and in
Stavros> fact by any user on the ARPANET, with no password protection. This was not
Stavros> some sort of lapse in security; in fact, it was widely known and taken
Stavros> advantage of (and in some sense was the inspiration for the FSF philosophy).
The book you are referring to was published with a Copyright notice.
The fact that you were able to obtain an illegitimate pre-publication
copy of the book from the author's computer does not place it
in the public domain.
Have you consulted an IP lawyer about this theft and your
theory of how that places a work into the "public domain"?
Chris,
Please re-read the previous postings, because (a) I didn't say what
you think I said and (b) things aren't at all as clear as you think
they are.
-s
Detail:
······@news.dtpq.com (Christopher C. Stacy) wrote in message news:<·············@news.dtpq.com>...
> The book you are referring to was published ...
First of all, the document I pointed to was (I'm pretty sure) Moon's
Maclisp Reference Manual (not KMP's revised manual), which was never
published as a book, just as a technical report.
> ... was published with a Copyright notice.
I don't know if the technical report carried a copyright notice.
> ... was published ...
It's not clear to me whether distributing a technical report counted
as "publication" under 1982 copyright law. What's more, KMP's book
was published with MIT's copyright, not KMP's.
> ...you were able to obtain an illegitimate pre-publication copy of
> the book from the author's computer...
The document(s) in question were *not* obtained from "the author's
computer", but from a site with an archive of files from MIT-MC.
MIT-MC was an MIT-owned and -operated computer, and both Moon and KMP
had access to it as research staff (not academic staff) of MIT working
on sponsored research. The documents in question were not in personal
directories, but in project directories. So it is plausible (though
not certain) that these were official project documents owned by MIT
and/or the research sponsors, even if they were written by Moon and/or
KMP.
> ... your theory of how that places a work into the "public domain"?
I never claimed that either the Moonual or KMP's manual was in the
public domain. In fact, if you'd re-read my message carefully, you'll
see that I said that I didn't know what their legal status was.
Perhaps public domain, perhaps MIT copyright, perhaps Moon's or KMP's
copyright. There are both legal and factual issues here.
> The fact that you were able to obtain... does not place it in
> the public domain.
Does the fact that the author made the document available by anonymous
FTP (on MIT-MC) on a project directory for years on end constitute
"publication"? I don't know. If yes, then under copyright law
prevailing in 1982, if it does not carry a copyright notice, then it
is in the public domain. Was it written by Moon or KMP as part of
their paid employment with MIT? I don't remember what copyright law
said in 1982 about works for hire, but certainly MIT would have some
rights to the documents, in particular the right to publish them.
> Have you consulted an IP lawyer about this theft...?
It seems you are characterizing putting these documents up on the FTP
site I pointed to as "theft" (I assume you're not characterizing my
mentioning the FTP address as theft). Theft is a strong word -- and
technically inappropriate for copyright infringement, but I'd have let
that pass if you didn't mention IP lawyers. It implies not only that
the maintainer of the site didn't have the right to put the documents
up, but also that he knew (or should have known) that he didn't have
that right. I have no reason to believe that he put them up in bad
faith. Quite the contrary.
As for how much I know about IP law, no, I am not a lawyer, however, I
have studied legal treatises and the case law relevant to many of
these issues.
I would be happy to be corrected on both the facts and the law in
these cases, but I do ask that you read what I've written more
carefully next time.
-s
From: Christopher C. Stacy
Subject: Re: Copyright of 1979 MacLisp Manual (Was: Etymology of MAP[LIST,CAR,CON,CAN,C,L])
Date:
Message-ID: <uu117kpth.fsf@news.dtpq.com>
>>>>> On 2 Mar 2004 10:16:02 -0800, Stavros Macrakis ("Stavros") writes:
Stavros> First of all, the document I pointed to was (I'm pretty
Stavros> sure) Moon's Maclisp Reference Manual (not KMP's revised
Stavros> manual), which was never published as a book, just as a
Stavros> technical report.
I was really only commenting on KMP's book; I don't know much about
the status of Moon's document. I am quite familiar with the MIT
computers in question, since I was in charge of them at the time,
and I do not personally subscribe to your interpretations. The legal
answer to the copyright questions, as I suggested, would be discerned
in a court of law. In any event, I am very comfortable discouraging
people from taking documents that do not belong to them, copying them
without explicit permission, and declaring them to be public domain.
That's really all I have to say on the subject.
······@news.dtpq.com (Christopher C. Stacy) wrote in message news:<·············@news.dtpq.com>...
> I am quite familiar with the MIT computers in question, since I
> was in charge of them at the time, and I do not personally
> subscribe to your interpretations.
Which 'interpretations' exactly do you disagree with?
> The legal answer to the copyright questions, as I suggested,
> would be discerned in a court of law.
Yes, if the copyright holder(s) sued. I do not know who the copyright
holder for Moon's manual is (Moon? MIT? no one?), and as far as I
know, they are not suing. The copyright holder listed for KMP's
manual is MIT. KMP claims that this is an error; I have no way of
evaluating that claim.
In any case, it is perfectly reasonable to try to understand what the
law says about the situation without going to court.
> In any event, I am very comfortable discouraging
> people from taking documents that do not belong to them,
Since a major issue is precisely who holds what rights to these
documents, this is begging the question. Even if MIT or KMP is the
copyright holder, it is possible that others have the license to
redistribute. This is not at all theoretical. For several years,
MacLisp and Macsyma development was funded by the same government
agencies (DOE et al.). MIT is the copyright holder of Macsyma (the
1982 version, called DOE Macsyma), but the DOE has the right to
redistribute and sublicense, and in fact 1982 Macsyma (renamed Maxima
to avoid trademark issues) is now licensed under the GPL.
> declaring them to be public domain.
I never declared anything to be public domain. Did anyone else? The
maintainer of ftp://ftp.its.os.org put these documents up on his FTP
server for anonymous FTP. Is this in your mind synonymous with
declaring them public domain? How is that different from putting them
up for anonymous FTP at mit-mc.arpa? Did *that* declare them to be
public domain?
By the way, these directories are no longer available at
ftp.its.os.org, presumably at KMP's request. I think that's a pity
(regardless of who if anyone holds the copyright).
-s
········@alum.mit.edu (Stavros Macrakis) wrote in message news:<····························@posting.google.com>...
>
> By the way, these directories are no longer available at
> ftp.its.os.org, presumably at KMP's request. I think that's a pity
> (regardless of who if anyone holds the copyright).
>
> -s
I think it was Kent himself in his Slashdot interview who said he felt
it would be a shame if in the future people were to look back on the
birth of computing, and they were able to go back to circa 1990 but no
further (since all those paper-based records and personal recollections
would not have been made available on the web prior to that time). He
urged people to make all their records available on the web to prevent
this at the time.
Whilst I'm not trying to say copyright is unimportant and we should have
free access to all this stuff (although that would be nice...), I just hope
that *somebody, somewhere* knows who owns it and is taking steps to ensure
its continuing availability (even if that availability is limited, and for a
price).
It's hard to avoid reinventing the wheel if there's no record of the wheel
having existed.
-Duncan