a searchable index to the Common Lisp standard
Steven M. Haflich
to be excellent as a language standard, it is not particularly easy to
use. Details about a particular element of the language may be
scattered in various obscure places. Without an exhaustive
index, any conclusion about language semantics is uncertain.
I recently created a searchable index to the language standard
primarily for my own benefit, and it's proven so useful to me that
I'm making it generally available on the Franz website:
http://www.franz.com/search-ans.html
Check it out.
Craig Brozefsky
> Why is it that with Common Lisp I get The HyperSpec, CLTL2 and the
> actual standard available online with a searchable index but with C,
> C++ and Java there are no complete, definitive or qualitative language
> references available online (that my Altavista time has turned up
> anyways), when these languages are so much more popular?
There are references for the java languages, the java virtual machine,
and also class references, all available in various forms online. I
have used to extensively both when doing Java programming, and then
doing research into it's method selection and reflective properties
for some doodling in Kawa. They were complete, and of respectable
quality.
java.sun.com should have a top level link to documentation. Java
actually is a rather well documented language IMO.
I can't realy say the same for C, but that could just be that I have
never had a need to look really hard as I picked it up sorta like a
programming creole thru years of massaging packages into compiling and
running on various platforms.
Kent M Pitman
> Why is it that with Common Lisp I get The HyperSpec, CLTL2 and the
> actual standard available online [...]
Not to detract from your other points, but the actual standard is not
available online. It is available only in hardcopy and only from ANSI.
Both the HyperSpec and the similar document from Franz (for which I
don't see a catchy name) are derived documents. The HyperSpec is
derived (with permission) from the standard itself and the Franz
document (as nearly as I can tell) from something that is the
technical (but not legal) equivalent of the standard, but that is not
owned by ANSI. The differences in both cases is, in practical terms,
(between the document in question and the ANSI standard) include
cosmetic issues and the slight risk of errors in mechanical
translation. But in spite of their similarity to the standard, it is
an important legal distinction that these are not the ANSI standard
and they do not purport to be. For a copy of the definitive
reference, you must contact ANSI.
The online CLTL2 is also a fine feat of technical engineering, but the
underlying document is not conformant to the ANSI standard. It is
perhaps 95% similar, but it is easy for a knowledgeable party to point
to discrepancies.
Nevertheless, you're right that there are a wealth of very usable online
documents for Lisp, and it's surprising that the other languages don't
have such things.
Maybe it's harder to make such things in other languages.
;-)
Barry Margolin
Kent M Pitman <pit...@world.std.com> wrote:
>Nevertheless, you're right that there are a wealth of very usable online
>documents for Lisp, and it's surprising that the other languages don't
>have such things.
I think it's mostly because you took the initiative to work out the legal
issues with ANSI that allowed you to create the Hyperspec, and folks in
those other language communities have not.
I should note that a late draft of the C++ standard *is* available on the
web. The final version isn't, but I expect that the differences are
minor. It's available at <http://www.cygnus.com/misc/wp/>.
--
Barry Margolin, bar...@bbnplanet.com
GTE Internetworking, Powered by BBN, Burlington, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Don't bother cc'ing followups to me.
Kent M Pitman
> In article <sfwg18n...@world.std.com>,
> Kent M Pitman <pit...@world.std.com> wrote:
> >Nevertheless, you're right that there are a wealth of very usable online
> >documents for Lisp, and it's surprising that the other languages don't
> >have such things.
>
> I think it's mostly because you took the initiative to work out the legal
> issues with ANSI that allowed you to create the Hyperspec, and folks in
> those other language communities have not.
Well, yes, and in fairness it's worth noting that this was an
extraordinarily difficult task. My private research at the time led
me to believe that a great many committees have undertaken the same
task, with varying ad hoc degress of success (usually leading to a
commercial hardcopy document, the legal issues surrounding were no
less complex than those we confronted in getting the HyperSpec out).
It's perhaps only an accident of history that we were the first to
do an extensive online document.
All that said, though, the fact is that technical preparation of the
HyperSpec was (relatively speaking) a trivial matter given Lisp as a
starting point. It took about 3 months of my time to do, but mostly
only because I had to implement a special-purpose tool that could
parse and "execute" TeX (not LaTeX) code. Also included in that time
was that I implementd the translator to output something completely
unrelated to HTML (a Harlequin document format) because HTML was not
yet popular; I retargeted the tool at the last minute to output HTML
instead. But it all began as a hack because it seemed like it would
be fun.
My guess is that it would have, at the time certainly and perhaps even
still today, taken someone longer than three months to do that task.
And that in most cases they would have just not bothered because it
wouldn't have seemed either tractable or fun.
A lot of what makes Lisp so cool is that it inspires people to do
things for sport.
Erik Naggum
| Nevertheless, you're right that there are a wealth of very usable online
| documents for Lisp, and it's surprising that the other languages don't
| have such things.
|
| ;-)
it could also be that people using other languages don't care quite so
much about the language as such. for instance, we have seen a lot of
heat about Common Lisp not having a standard socket library and such, but
nobody in their right mind would accuse C of not having the same, despite
availability of socket libraries in all Common Lisp and C implementations
alike. obviously, these people aren't just interested in having some
piece of code "work" or "do the job" in the Perl sense, and are therefore
more likely to want to know what the real thing says.
I'm not so certain about the size of the book market for Lisp books, and
whether the relatively small market contributes to the above, or whether
the above contributes to the small market. however, the need for books
that explain _all_ parts of Common Lisp is not at all satisfied. I have
myself spent countless hours merely studying how things were supposed to
be used from the HyperSpec and reading up on old literature and sometimes
even talking with other Lispers.
for instance, a friend of mine who has used Common Lisp since CLtL1 came
out didn't know about PPRINT-DISPATCH. I had stumbled upon it when being
intensely dissatisfied with the same data having four slightly different
shapes depending on the recipient and that got translated into a lot of
hairy code. there had to be a better way, and that turned out to be the
pretty printer. it was great fun to tell him about it.
#:Erik
--
SIGTHTBABW: a signal sent from Unix to its programmers at random
intervals to make them remember that There Has To Be A Better Way.
Christopher Browne
>In article <sfwg18n...@world.std.com>,
>Kent M Pitman <pit...@world.std.com> wrote:
>>documents for Lisp, and it's surprising that the other languages don't
>>have such things.
>
>issues with ANSI that allowed you to create the Hyperspec, and folks in
>those other language communities have not.
I'd strongly agree. It takes having someone who *wants* to push for it
in order for it to happen.
Add to that that:
a) Some of the language standards were hammered out before people
thought of "non-dead-tree-editions" as being really valuable. ANSI C's
establishment predates this, for instance.
b) ANSI FORTH did get released in electronic form; I don't have a URL
handy, but I think it got released as an MS Word document.
c) There are publishers that would be just as happy to have a "monopoly"
on the sale of the standard guide.
--
Where do you *not* want to go today? "Confutatis maledictis, flammis
acribus addictis" (<http://www.hex.net/~cbbrowne/msprobs.html>
cbbr...@hex.net- <http://www.hex.net/~cbbrowne/languages.html>
Erik Naggum
| b) ANSI FORTH did get released in electronic form; I don't have a URL
| handy, but I think it got released as an MS Word document.
|
| c) There are publishers that would be just as happy to have a "monopoly"
| on the sale of the standard guide.
I find the latter two points to be quite amusing taken together. there
is now a company that has monopoly, not just on the software product that
is necessary to read the document, but also on the ability of anyone to
read that document in the future. Microsoft has gotten a little softer
on the "upgrade or die" choice that its victims were facing, but if there
were only _one_ company in the world I wouldn't trust with a single bit
of information valuable to me, it'd be Microsoft.
Stephen J Bevan
> b) ANSI FORTH did get released in electronic form; I don't have a URL
> handy, but I think it got released as an MS Word document.
The standard did not get released in electronic form, you have to buy
it like most other standards. However, there is a copy of the final
draft available online (see comp.lang.forth FAQ for details). Whether
this is sufficient depends on exactly what you want the standard for.
Howard R. Stearns
I like it. Thank you for doing this!
Now, can we get lambda-ir working so that we don't have to use
cgi/thunderstone.... (I searched for lisp within the thunderstone site.
Nada.)