I noticed that DFKI posted job openings on comp.ai today, and thought
I'd had a look at what they're doing these days (Until 10 years ago I
worked at the Comp. Linguistics department in Saarbr�cken, which
cooperated a lot with DFKI)
> * 200431 - Three researchers and software engineers in the area of
> semantic web and question answering, starting around 07/04
> (05.05.2004, closing date 15.06.2004)
> http://www.dfki.de/lt/job_details.php?id=200431
I quote:
* Information extraction
* Linguistic analysis
* Software technologies based on XML, RDF, OWL (e.g., Prot?g?, Jena)
* Software development in Java
* Machine learning
* Ontology-based QA
* Evaluation of QA
Good grief! And this used to be a thriving lisp community!
Is this typical for AI/Computational Linguistics research centers these
days?
(In case you think RDF means Reality Distortion Field: It doesn't.
It means Resource Description Framework. But from what I've seen
of RDF, I think you need an RDF to work with it :-()
--
(espen)
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> I noticed that DFKI posted job openings on comp.ai today, and thought
> I'd had a look at what they're doing these days (Until 10 years ago I
> worked at the Comp. Linguistics department in Saarbrücken, which
> cooperated a lot with DFKI)
>
> > * 200431 - Three researchers and software engineers in the area of
> > semantic web and question answering, starting around 07/04
> > (05.05.2004, closing date 15.06.2004)
> > http://www.dfki.de/lt/job_details.php?id=200431
>
> I quote:
>
> * Information extraction
> * Linguistic analysis
> * Software technologies based on XML, RDF, OWL (e.g., Prot?g?, Jena)
> * Software development in Java
> * Machine learning
> * Ontology-based QA
> * Evaluation of QA
>
> Good grief! And this used to be a thriving lisp community!
> Is this typical for AI/Computational Linguistics research centers these
> days?
>
> (In case you think RDF means Reality Distortion Field: It doesn't.
> It means Resource Description Framework. But from what I've seen
> of RDF, I think you need an RDF to work with it :-()
This article finally prompted me to see just what the heck this
semantic web thing I've been hearing about is.
http://www.google.com/search?hl=en&lr=&ie=UTF-8&oi=defmore&q=define:semantic+web
If this isn't a job for Lisp, I don't know what is. Why is XML, RDF,
and whatever else they are contriving better suited to this task than
Lisp expressions?
Wouldn't even a Java application spend less time parsing a Lisp
expression than XML with an available library? What I'm asking here
is that Lisp looks a lot easier to parse.
And to think that XML was supposed to just be a lite weight substitute
for SGML.
--
I wouldn't mind the rat race so much if it wasn't for all the damn cats.
David Steuber <·····@david-steuber.com> wrote in message news:<··············@david-steuber.com>...
> http://www.google.com/search?hl=en&lr=&ie=UTF-8&oi=defmore&q=define:semantic+web
>
> If this isn't a job for Lisp, I don't know what is. Why is XML, RDF,
> and whatever else they are contriving better suited to this task than
> Lisp expressions?
>
> Wouldn't even a Java application spend less time parsing a Lisp
> expression than XML with an available library? What I'm asking here
> is that Lisp looks a lot easier to parse.
RDF and for example xmlrpc is a lot easier to use with foreign
languages, because they simply have to use their expat or sablot
library and don't have to invent lisp parsing.
RDF is de-facto standard and used a lot by this kind of semantic web
thing.
I use it also everyday, and it doesn't hurt. It's just the exchange
format, as HTML is another de-facto standard.
You can easily translate both to simplier expressions without
redundant end-tags and guaranteed proper nesting and identation.
And of course nobody likes to work in sourcecode with this kind of XML
expressions, internally we all use better formats. (Even in PHP, which
I do now)
······@x-ray.at (Reini Urban) writes:
> RDF and for example xmlrpc is a lot easier to use with foreign
> languages, because they simply have to use their expat or sablot
> library and don't have to invent lisp parsing.
For the things that would need a Lisp parser, they need
an RDF parser.
> RDF is de-facto standard and used a lot by this kind of semantic web
> thing.
> I use it also everyday, and it doesn't hurt. It's just the exchange
> format, as HTML is another de-facto standard.
I use it most days, and it does hurt.
-- jd
David Steuber wrote:
> This article finally prompted me to see just what the heck this
> semantic web thing I've been hearing about is.
>
> http://www.google.com/search?hl=en&lr=&ie=UTF-8&oi=defmore&q=define:semantic+web
>
> If this isn't a job for Lisp, I don't know what is. Why is XML, RDF,
> and whatever else they are contriving better suited to this task than
> Lisp expressions?
For the same (lack of) reason that HTML 1.0 was better than all the
other SGML based hypertext systems that went before it.
> Wouldn't even a Java application spend less time parsing a Lisp
> expression than XML with an available library? What I'm asking here
> is that Lisp looks a lot easier to parse.
No, parsing Lisp in Java isn't that much easier than parsing XML.
Parsing a subset of s-expression syntax that doesn't require symbol
packages, a read table and reader macros would be (eg KIF), but Lisp
itself isn't any simpler than XML (or even XML+namespaces). Lisp's more
compressed, and parses to a tree where the children aren't partitioned
to attributes and elements. XML has more redundancy, you have one extra
check that the closing tag matches, but no typing, so 0123456701234567
is always a string. Lisp you'd have to put it in quotes if it's a credit
card number or other numeric string id. (With ASN.1 you declare it as
numeric string and have the metadata available, and it would be even
faster to parse)
> And to think that XML was supposed to just be a lite weight substitute
> for SGML.
In what way is it not?
The point of using XML in the semantic web is that many web sites
already have X(HT)?ML based content so you need XML to hook in to them.
If they all had s-expression based content, then there'd be a call for
s-expressions in semantic web.
For most web content, we are talking text, not data. Direct-to-source
authors seem much happier escaping '<' and '>' when they have to than
'(' and ')'. Most the time angles don't get used in text, but
parentheses are far more common.
But instead, the preferred web formats are SGML derived, so anything
that extracts semantic information from them has to work with the
available infrastructure.
The problems with the semantic web are beyond syntax, and no-one is
expecting to write the RDF stuff by hand, so it doesn't matter too much
what the interchange format is. But Lisp has its place as an
implementation language, even though the interchange forms try to be
implementation independent. Do you care that the messages between
telecoms switches or Internet routers aren't s-expressions? So why care
that the machine-machine messages for web agents aren't?
Machine interoperability systems that work already (ISO 10303 for
interchange between tools by shared semantic models, ASN.1 for telecoms
message syntax) separate the abstract and concrete syntaxes; hopefully
the semantic web will evolve enough to do the same.
Pete
Pete Kirkham <·················@cafemosaic.co.uk> writes:
> No, parsing Lisp in Java isn't that much easier than parsing
> XML.
I disagree. XML parsers tend to be alarmingly large,
and when RDF enters the mix, it's even worse.
> > And to think that XML was supposed to just be a lite weight substitute
> > for SGML.
>
> In what way is it not?
>
> The point of using XML in the semantic web is that many web sites
> already have X(HT)?ML based content so you need XML to hook in to
> them.
That doesn't mean you have to use XML syntax for the ontologies,
just for instance.
Web sites also have a not on non-X(HT)?ML based context,
and the semantic stuff is supposed to apply to that as well.
> For most web content, we are talking text, not data.
There's a lot of non-text out there. Pictures, say.
> The problems with the semantic web are beyond syntax, and no-one is
> expecting to write the RDF stuff by hand, so it doesn't matter too
> much what the interchange format is.
That sounds plausible, but now that I've had to do a lot with XML and
RDF, I don't agree that the format doesn't matter too much.
> So why
> care that the machine-machine messages for web agents aren't?
Well, for one thing, it's not only machine-machine messages.
> Machine interoperability systems that work already (ISO 10303 for
> interchange between tools by shared semantic models, ASN.1 for
> telecoms message syntax) separate the abstract and concrete syntaxes;
> hopefully the semantic web will evolve enough to do the same.
See Relax NG's compact syntax for a good example of how
much better a non-XML syntax might be.
-- jd
David Steuber <·····@david-steuber.com> wrote in message news:<··············@david-steuber.com>...
> If this isn't a job for Lisp, I don't know what is. Why is XML, RDF,
> and whatever else they are contriving better suited to this task than
> Lisp expressions?
>
> Wouldn't even a Java application spend less time parsing a Lisp
> expression than XML with an available library? What I'm asking here
> is that Lisp looks a lot easier to parse.
Do you really think that the ease of vparsing XML or SEXPs is anywhere
other than in the noise when trying to deal with semantics?
--tim
Espen Vestre wrote:
> Good grief! And this used to be a thriving lisp community!
> Is this typical for AI/Computational Linguistics research centers these
> days?
that's quite normal. today, most students never heard of lisp[1], but
nevertheless they might be competent ai researchers. and, I'd say, all
of them know java, some better, some not so good but better than
anything else (if they know any other language at all). generally, you
move to the lantern that is circled by flies if you don't want to stay
hungry. here contracts are for 2 years, generally, and budgets are
extremely tight. you can't waste that time teaching the people another
language first.
--
Matthias Buelow; ···@{mukappabeta,informatik.uni-wuerzburg}.de
[1] ok, they might have heard it being mentioned maybe once or twice
during some undergrad lecture, but not have memorized the term.
Matthias Buelow <···@mukappabeta.de> writes:
> you can't waste that time teaching the people another language first.
I'd say that knowing _at_least_ 2 fundamentally different programming
languages should be a minimum qualification for doing any serious
programming... (but if one of them is Common Lisp, I'd accept FORMAT
as the other ;-P)
--
(espen)
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> I'd say that knowing _at_least_ 2 fundamentally different programming
> languages should be a minimum qualification for doing any serious
> programming...
Indeed.
> (but if one of them is Common Lisp, I'd accept FORMAT
> as the other ;-P)
Or even LOOP?
--
Arvid
Espen Vestre wrote:
> I'd say that knowing _at_least_ 2 fundamentally different programming
> languages should be a minimum qualification for doing any serious
> programming... (but if one of them is Common Lisp, I'd accept FORMAT
but what if the candidates are not qualified in this regard? after all,
ai and linguistics research is not mainly about programming.
--
Matthias Buelow; ···@{mukappabeta,informatik.uni-wuerzburg}.de
Matthias Buelow <···@mukappabeta.de> writes:
> but what if the candidates are not qualified in this regard? after
> all, ai and linguistics research is not mainly about programming.
Well, then they should mainly do explorational programming, not produce
end-user software, which again means that it's even less sensible to
use java instead of prolog or lisp.
--
(espen)
Espen Vestre wrote:
> Good grief! And this used to be a thriving lisp community!
> Is this typical for AI/Computational Linguistics research centers these
> days?
One could think that universities still use a lot CL and especially in
AI it is heavily used. But well, many projects are getting rewritten in
Java, even AI projects.
Andr�
--
André Thieme <······························@justmail.de> writes:
> Espen Vestre wrote:
> > Good grief! And this used to be a thriving lisp community! Is this
> > typical for AI/Computational Linguistics research centers these days?
>
> One could think that universities still use a lot CL and especially in
> AI it is heavily used. But well, many projects are getting rewritten
> in Java, even AI projects.
But why the rewrite? What advantage does Java offer over Lisp?
--
I wouldn't mind the rat race so much if it wasn't for all the damn cats.
David Steuber <·····@david-steuber.com> writes:
> André Thieme <······························@justmail.de> writes:
>
>> Espen Vestre wrote:
>> > Good grief! And this used to be a thriving lisp community! Is this
>> > typical for AI/Computational Linguistics research centers these days?
>>
>> One could think that universities still use a lot CL and especially in
>> AI it is heavily used. But well, many projects are getting rewritten
>> in Java, even AI projects.
>
> But why the rewrite? What advantage does Java offer over Lisp?
Curly braces. They make the program run faster.
David Steuber wrote:
> But why the rewrite? What advantage does Java offer over Lisp?
people who can work with and/or extend the system (or ai shell). from
my experience, students today know java, and nothing else. (ok, maybe
some php or stuff from their home programming or part-time job.) a
pity, of course. in the last 5 years java has replaced C as the lingua
franca in computer science, and has eaten into the shares of higher
level languages like lisp/scheme, sml, etc. because it is also deemed
"high level" (because it's "safe", byte-compiled, object oriented,
garbage collected, etc.) and people arguing for ever shortening study
times also put pressure on faculty to teach quickly learned, "industry
compatible" concepts.
Matthias Buelow <···@mukappabeta.de> writes:
> David Steuber wrote:
>
> > But why the rewrite? What advantage does Java offer over Lisp?
>
> people who can work with and/or extend the system (or ai shell). from
> my experience, students today know java, and nothing else. (ok, maybe
> some php or stuff from their home programming or part-time job.) a
> pity, of course. in the last 5 years java has replaced C as the
> lingua franca in computer science, and has eaten into the shares of
> higher level languages like lisp/scheme, sml, etc. because it is also
> deemed "high level" (because it's "safe", byte-compiled, object
> oriented, garbage collected, etc.) and people arguing for ever
> shortening study times also put pressure on faculty to teach quickly
> learned, "industry compatible" concepts.
It sounds like Java, with the help of the attitudes you describe, is
going to throw computer science into a dark age.
Java doesn't seem to fake being high level to me. It looks an awful
lot like C. Or should I say C++ lite?
One nice thing I'm beginning to find in Lisp is that I can think more
about the problem I am trying to solve than the low level details of
how to add and subtract. What I mean is Lisp has a much higher level
of abstraction than Java. It's no contest. Java may mean not having
to make sure your integers are word aligned, but a C compiler can do
that too.
--
I wouldn't mind the rat race so much if it wasn't for all the damn cats.
David Steuber wrote:
> Matthias Buelow <···@mukappabeta.de> writes:
>
>
>>David Steuber wrote:
>>
>>
>>>But why the rewrite? What advantage does Java offer over Lisp?
>>
>>people who can work with and/or extend the system (or ai shell). from
>>my experience, students today know java, and nothing else. (ok, maybe
>>some php or stuff from their home programming or part-time job.) a
>>pity, of course. in the last 5 years java has replaced C as the
>>lingua franca in computer science, and has eaten into the shares of
>>higher level languages like lisp/scheme, sml, etc. because it is also
>>deemed "high level" (because it's "safe", byte-compiled, object
>>oriented, garbage collected, etc.) and people arguing for ever
>>shortening study times also put pressure on faculty to teach quickly
>>learned, "industry compatible" concepts.
>
>
> It sounds like Java, with the help of the attitudes you describe, is
> going to throw computer science into a dark age.
No, they are moving towards the light: OO, GC, and now they are throwing
out a lot of orthodoxy and doing Groovy, adding more dynamism, less
static typing.
kenny
--
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application
David Steuber wrote:
> It sounds like Java, with the help of the attitudes you describe, is
> going to throw computer science into a dark age.
Well.. in computer science it probably doesn't matter that much. In
university research, programming is only a means of more or less trying
out newly developed algorithms and concepts and providing reference
implementations, and not producing industrial strength software.
The problem is more that the graduated workforce, so to say, know only
Java, and the whole industry is embracing this (simple language, no
shortage in applicants, cheap labour, so to say.)
But maybe we should also look at the third factor, namely the open
source community, which doesn't follow budget policies and overly
rational decision making. Here a lot more of experimenting is going on,
_and_ people want to produce useful software, and have fun doing so.
Useful scripting languages like Perl, Python, PHP etc. weren't industry
designed. However, they're now major players in the industry. In the
early 90ies, the death of Unix has (once more) been proclaimed, and
Windows NT has been dubbed the "Unix killer". Nothing like it has
happened. With the advent of the free Unix clone Gnu/Linux, a
renaissance of Unix and Unixoid operating principles has emerged, and
not only has it dragged along commercial SVR4 systems and BSD but it has
even managed to greatly expand into commercial and end-consumer markets.
Don't forget what, for example, MacOS X is, after all. So the free
software community is something that shouldn't be underestimated. And I
see rising interest in "alternative" languages in the community. Scheme
is on the rise (and that for what was meant to be a purely educational
language), exotic languages like ML are being used for real-world
consumer utilities (like the OCaML dialect for mldonkey), and I think
the Common Lisp community is growing aswell (I haven't seen sites like
www.common-lisp.net 5 years ago.)
Maybe the open source community can produce the momentum where academics
has "failed", or where the academically based industry has bungled
majorly, producing lasting caution towards the Lisp community (like with
the special-purpose lisp machines.)
--
Matthias Buelow; ···@{mukappabeta,informatik.uni-wuerzburg}.de
Matthias Buelow <···@mukappabeta.de> writes:
> Maybe the open source community can produce the momentum where
> academics has "failed", or where the academically based industry has
> bungled majorly, producing lasting caution towards the Lisp community
> (like with the special-purpose lisp machines.)
There's a hope. So long as the FOSS people don't all decide to go
with Java so that it is easier to get a job.
For my own use, I'm interested in Lisp because it offers the
possibility of even faster development than high level scripting
languages such as Perl and Python along with faster execution speed.
The issue of libraries seems to be sorting itself out. FOSS is also
easier to distribute.
Naturally it would be very nice to be able to get paid for Lisp
development. Right now, my only shot of that seems to be starting my
own company.
--
I wouldn't mind the rat race so much if it wasn't for all the damn cats.
David Steuber <·····@david-steuber.com> writes:
> Naturally it would be very nice to be able to get paid for Lisp
> development. Right now, my only shot of that seems to be starting my
> own company.
I suggest that you start asking to your friends and relatives. Some
of them may need a simple application, and they may not care which
language it is written in.
Paolo
--
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (Google for info on each):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
"David Steuber" <·····@david-steuber.com> wrote
>
> Naturally it would be very nice to be able to get paid for Lisp
> development. Right now, my only shot of that seems to be starting my
> own company.
Yes, you should do that. ;-)
I've done that and it's the best way of doing what you like
(Lisp+electronics design for me). The bigger drawback is that you will be
very busy after that!
Marc
According to David Steuber <·····@david-steuber.com>:
> It sounds like Java, with the help of the attitudes you describe, is
> going to throw computer science into a dark age.
It sounds like OOP has survived Java.
Several (Smalltalk) people have mentioned that the last several ECOOPs had
been all redoing-old-stuff-in-Java-calling-it-new but that phase has passed
and now new stuff are being done, not necessarily in Java.
--
Ng Pheng Siong <····@netmemetic.com>
http://firewall.rulemaker.net -+- Firewall Change Management & Version Control
http://sandbox.rulemaker.net/ngps -+- ZServerSSL/Zope Windows Installers
In article <··············@david-steuber.com>,
David Steuber <·····@david-steuber.com> wrote:
>Matthias Buelow <···@mukappabeta.de> writes:
>
>> David Steuber wrote:
>>
>> > But why the rewrite? What advantage does Java offer over Lisp?
>>
>> people who can work with and/or extend the system (or ai shell). from
>> my experience, students today know java, and nothing else. (ok, maybe
>> some php or stuff from their home programming or part-time job.) a
>> pity, of course. in the last 5 years java has replaced C as the
>> lingua franca in computer science, and has eaten into the shares of
>> higher level languages like lisp/scheme, sml, etc. because it is also
>> deemed "high level" (because it's "safe", byte-compiled, object
>> oriented, garbage collected, etc.) and people arguing for ever
>> shortening study times also put pressure on faculty to teach quickly
>> learned, "industry compatible" concepts.
>
>It sounds like Java, with the help of the attitudes you describe, is
>going to throw computer science into a dark age.
Computer science has been in a dark age since the 1970s.
-Lupo
"Democrats are ... the party that says government can make you richer,
smarter, taller and get the chickweed out of your lawn. Republicans are
the party that says government doesn't work, and they get elected to prove
it." -P. J. O'Rourke, Parliament of Whores <··@io.com>
On Wed, 26 May 2004 20:45:12 -0500, Lupo LeBoucher wrote:
>
> Computer science has been in a dark age since the 1970s.
>
From what I've read, I have to agree (I was only born in 1975, so I don't
have first hand knowledge!).
So does anyone think we're going to emerge from this dark age anytime soon?
Cheers,
Bill.
--
Dr. William Bland.
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax. (Ken Anderson).
I'm sure in some respects, that the corporate world will always look a
bit like the dark ages, and drag the Universities into it with them.
Here's one of my favorites from the fortune program on FreeBSD systems:
"Real World, The n.:
1. In programming, those institutions at which programming may
be used in the same sentence as FORTRAN, COBOL, RPG, IBM, etc. 2. To
programmers, the location of non-programmers and activities not related
to programming. 3. A universe in which the standard dress is shirt and
tie and in which a person's working hours are defined as 9 to 5. 4.
The location of the status quo. 5. Anywhere outside a university.
"Poor fellow, he's left MIT and gone into the real world." Used
pejoratively by those not in residence there. In conversation, talking
of someone who has entered the real world is not unlike talking about a
deceased person."
Simply replace "FORTRAN, COBOL, RPG, IBM" with your favorite "love to
hate" programming products and companies.
--Bruce
William Bland wrote:
> On Wed, 26 May 2004 20:45:12 -0500, Lupo LeBoucher wrote:
>
>>Computer science has been in a dark age since the 1970s.
>>
>
>
> From what I've read, I have to agree (I was only born in 1975, so I don't
> have first hand knowledge!).
>
> So does anyone think we're going to emerge from this dark age anytime soon?
>
> Cheers,
> Bill.
William Bland <····@abstractnonsense.com> writes:
> On Wed, 26 May 2004 20:45:12 -0500, Lupo LeBoucher wrote:
> >
> > Computer science has been in a dark age since the 1970s.
> >
>
> From what I've read, I have to agree (I was only born in 1975, so I don't
> have first hand knowledge!).
>
> So does anyone think we're going to emerge from this dark age anytime soon?
I would expect part of the answer to that to depend on the industry
perception of whether we are in a dark age and the willingness to
accept Lisp as a source of light.
I don't have any real answers, but more entrepreneurs making money off
of Lisp at the expense of Java and other "mainstream" languages would
probably go a long way towards that. Lisp is a double edged sword.
On the one edge you can cut down the Java competition quite nicely.
On the other edge, you really do have to know about fundementals of
algorithms and such to make the language dance.
If you don't know the CS 101 stuff, you might as well be programming
in Java.
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1
On Thu, 27 May 2004 17:25:58 +0200, Svein Ove Aas wrote:
> William Bland wrote:
>
>> So does anyone think we're going to emerge from this dark age anytime
>> soon?
>
>
> How do you define "soon"?
Within my working lifetime (next 30-40 years or so I guess) would be fine
by me. Too optimistic?
--
Dr. William Bland.
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax. (Ken Anderson).
William Bland wrote:
> On Thu, 27 May 2004 17:25:58 +0200, Svein Ove Aas wrote:
>
>> William Bland wrote:
>>
>>> So does anyone think we're going to emerge from this dark age anytime
>>> soon?
>>
>>
>> How do you define "soon"?
>
> Within my working lifetime (next 30-40 years or so I guess) would be
> fine
> by me. Too optimistic?
>
Maybe not.
In an ideal world we'll get <insert current wording for molecular
nanotechnology here> woroking and useful sometime safe. As this world
isn't ideal, we'll probably do it too fast, but anyway:
*Eventually*, be it 20, 30 or 90 years from now, we'll have the required
technology for people to make whatever hardware they need themselves,
extremely cheap. There is also likely to be an extreme shortfall of jobs;
hopefully we'll manage /that/ readjustment well.
In any case, can you say "Lisp Machine"?
Thought you could...
In article <·····························@abstractnonsense.com>,
William Bland <····@abstractnonsense.com> wrote:
>On Wed, 26 May 2004 20:45:12 -0500, Lupo LeBoucher wrote:
>>
>> Computer science has been in a dark age since the 1970s.
>>
>
>From what I've read, I have to agree (I was only born in 1975, so I don't
>have first hand knowledge!).
Hey, I ain't far behind you, though I did code in OS-9 when I was a kid.
It's sort of emblematic; I can still get a pretty good job coding in
OS-9, and it's still a pretty good OS choice for embedded systems.
>So does anyone think we're going to emerge from this dark age anytime soon?
It doesn't look like anyone is doing any real research.
Just about every modern "innovation" I can think of was sitting in
research labs in about the same state of "doneness" in the 50's and 60's.
The only difference is the cheapness of memory and CPU, which was really
an electrical engineering achievement.
There is an annual "Progress in Computing" journal. I sat down and read
the whole thing when I was trying to avoid writing my dissertation. It is
extremely depressing reading.
In 1960 they were trying to predict climate and weather with computers;
present state of the art -about the same, as far as actually predicting
climate and weather goes.
In 1960 they were trying to make reliable spoken word translation;
presently, about the same level of doneness as far as accuracy goes.
In 1962 they were doing multiple processors, and thinking it would
eventually rule the world. It still is about as hard to do as it was then.
Maybee the compilers are a little better.
In 1964 they were using neural nets and Kalman filters to do things, in
2002 or so, Norvig wrote a book introducing the "AI" community to neural
nets and Kalman filters.
In 1968 they were postulating holographic storage as ... the future ...
They still are (though they did also had an article about LSI in ...the
future... and that one was right).
In 1970 they had articles on word processing, CAS and GUI's. In 2004, we
still use the same basic paradigms in all three fields.
In 1972 they noticed that computers were hard to make secure. In 2004 they
are still hard to make secure.
In 1976 they said they'd like to do automated programming. We'd still like
to do that.
Blah blah blah. It does on and on. There were a couple of interesting
things which never seemed to go anywhere which might be worth revisiting;
stochastic computing, for example, but by and large, since 1970 or so,
we've been reinventing the wheel on a faster spinning lathe.
We still can't figure out the two most basic things about software
development: how to make large, bug-free projects work, and how to make
reusable code.
I don't know how all this relates to Lisp, except to note that we had Lisp
in the 1950s, and most languages still substantially suck more than Lisp
does.
-Lupo
"If there's one word that sums up everything that's gone wrong since the
War, it's workshop."-Kingsley Amis <··@io.com>
Lupo LeBoucher wrote:
>>On Wed, 26 May 2004 20:45:12 -0500, Lupo LeBoucher wrote:
>>>
>>> Computer science has been in a dark age since the 1970s.
>
>>So does anyone think we're going to emerge from this dark age anytime
>>soon?
>
> It doesn't look like anyone is doing any real research.
>
> Just about every modern "innovation" I can think of was sitting in
> research labs in about the same state of "doneness" in the 50's and 60's.
> The only difference is the cheapness of memory and CPU, which was really
> an electrical engineering achievement.
>
> There is an annual "Progress in Computing" journal. I sat down and read
> the whole thing when I was trying to avoid writing my dissertation. It is
> extremely depressing reading.
> In 1960 they were trying to predict climate and weather with computers;
> present state of the art -about the same, as far as actually predicting
> climate and weather goes.
(snip)
> but by and large, since 1970 or so,
> we've been reinventing the wheel on a faster spinning lathe.
>
> We still can't figure out the two most basic things about software
> development: how to make large, bug-free projects work, and how to make
> reusable code.
Wow. That's really fascinating stuff. I always had a feeling there has been
an awful lot of wasted effort in the CS field. But then again, I find it
hard to believe it could have been any other way. Even if there hasn't been
much progress in solving computing problems, there's been the growth of IT
that has led to the near ubiquity of computers in large areas of the world.
That sort of rapid growth can easily throw a field of study into a
tailspin.
Besides that, it seems to me that the main problem in software development
is keeping up with all the different hardware that needs to be supported.
Hardware keeps changing (not so much in the past decade, but certainly from
the sixties to the early nineties) and software needs to be rewritten.
Sure, you can port software, but it's going to be rewritten anyway.
Eventually, software becomes obsolete. If software had had a stable
hardware platform on which to build for forty years (with the resources of
modern machines), I think you'd see software in a much more advanced state
than it is today.
What's your understanding of this? Or, do you have any good references? I'd
love to read more about this.
Chris Capel
Matthias Buelow <···@mukappabeta.de> writes:
> oriented, garbage collected, etc.) and people arguing for ever
> shortening study times also put pressure on faculty to teach quickly
> learned, "industry compatible" concepts.
Yes, but this is all PHB bullshit I think.
A comp.ling. student can do interesting things with lisp after relatively
short exposure. I know, 'cause I've taught lisp at a linguistics
department.
The result has to be one of two: Either the student is able to do less
real development on his own, reducing the explorative part of the study,
i.e. a return to a more traditional linguistics or philosophy education.
Or, the student spends more time programming (less) software, i.e. either
the study times are extended or the theory part has to suffer.
--
(espen)
Andr� Thieme <······························@justmail.de> writes:
> But well, many projects are getting rewritten in Java, even AI projects.
what a tremendous waste of human resources :-(
--
(espen)
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> Andr� Thieme <······························@justmail.de> writes:
>
> > But well, many projects are getting rewritten in Java, even AI projects.
>
> what a tremendous waste of human resources :-(
"But this time the wheel is _object oriented_!"
Best,
Thomas
--
Thomas Lindgren
"It's becoming popular? It must be in decline." -- Isaiah Berlin
Thomas Lindgren <···········@*****.***> writes:
> Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
>
> > Andr� Thieme <······························@justmail.de> writes:
> >
> > > But well, many projects are getting rewritten in Java, even AI projects.
> >
> > what a tremendous waste of human resources :-(
>
> "But this time the wheel is _object oriented_!"
But that's not the reason, at least not in my experience.
Java's more portable, it's what other people are using,
it avoids anti-Lisp prejudice, and Java has lots of
libraries that do things people want to do nowadays
like XML.
Something similar happened in AI in the 80s when people
were redoing things in Prolog that had already been
done in Lisp.
-- jd
"Jeff Dalton" <····@todday.inf.ed.ac.uk> wrote in message
····················@todday.inf.ed.ac.uk...
> Thomas Lindgren <···········@*****.***> writes:
>
> > Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> >
> > > Andr Thieme <······························@justmail.de> writes:
> > >
> > > > But well, many projects are getting rewritten in Java, even AI
projects.
> > >
> > > what a tremendous waste of human resources :-(
> >
> > "But this time the wheel is _object oriented_!"
>
> But that's not the reason, at least not in my experience.
>
> Java's more portable, it's what other people are using,
> it avoids anti-Lisp prejudice, and Java has lots of
> libraries that do things people want to do nowadays
> like XML.
>
> Something similar happened in AI in the 80s when people
> were redoing things in Prolog that had already been
> done in Lisp.
But in that case some things arguably were easier to do in Prolog.
Goal-directed knowledge-based behavior, for one thing. Sure, you can do it
in Lisp, but you implement Prolog first, or some moral equivalent. I built
several KRSs in Lisp because Prolog was too expressively weak, but I was
still building some kind of programmable logic - just not the kind Prolog
happened to pick.
There's nothing wrong in picking (or creating) a language that makes a
problem easier to solve. There is something wrong in picking a language that
only makes it more palatable to management.
Thomas Lindgren wrote:
> Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
>
>
>>Andr� Thieme <······························@justmail.de> writes:
>>
>>
>>>But well, many projects are getting rewritten in Java, even AI projects.
>>
>>what a tremendous waste of human resources :-(
>
>
> "But this time the wheel is _object oriented_!"
That's in fact also very important for many students.
I talked with some students and said that Lisp also is a nice language.
They had some experience with Scheme so far.
But the problem with Lisp is, they explained me, that one cannot do oop
in it.
Andr�
--
Andr� Thieme wrote:
> Thomas Lindgren wrote:
>
>> Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
>>
>>
>>>Andr� Thieme <······························@justmail.de> writes:
>>>
>>>
>>>>But well, many projects are getting rewritten in Java, even AI
>>>>projects.
>>>
>>>what a tremendous waste of human resources :-(
>>
>>
>> "But this time the wheel is _object oriented_!"
>
>
> That's in fact also very important for many students.
> I talked with some students and said that Lisp also is a nice language.
> They had some experience with Scheme so far.
> But the problem with Lisp is, they explained me, that one cannot do oop
> in it.
>
One wonders who taught them that. It is arguably true, but then you aren't
talking about Common Lisp. Is this a frequent phenomenon?
Andr� Thieme <······························@justmail.de> writes:
> I talked with some students and said that Lisp also is a nice
> language. They had some experience with Scheme so far.
> But the problem with Lisp is, they explained me, that one cannot do
> oop in it.
It's easy and fun to play with oop in Scheme (e.g. based on closures).
And for _serious_ OO programming, nothing beats Common Lisp, you get
more powerful OO than in any other language AND you don't have to tie
your hands and feet to the OO paradigm (A win-win situation too
good to be true for the narrow-minded ;-)).
--
(espen)
Andr� Thieme <······························@justmail.de> writes:
> Espen Vestre wrote:
> > Good grief! And this used to be a thriving lisp community! Is this
> > typical for AI/Computational Linguistics research centers these days?
>
> One could think that universities still use a lot CL and especially in
> AI it is heavily used. But well, many projects are getting rewritten
> in Java, even AI projects.
In some fields of AI the techniques used shifted considerably since
the 80s: From search and symbolic computation to more "numeric"
approaches. You can nicely see this if you compare Norvig's
"Paradigms of AI Programming" with Norvig/Russell "AI A Modern
Approach": There is more numerical computation/optimization and less
search going on. (Overly) simply speaking: Where symbols were
manipulated in the past, people now manipulate probability
distributions and use mathematical optimization. (Note that this
holds only for "some fields of AI", like vision, speech, maybe natural
language, maybe planning, but in these fields the trend is visible.)
While, except for the libraries, I don't know why one would want to
use Java instead of Lisp, I can see that using special-purpose
languages for optimization, numerics, and visualization can make a lot
of sense. (Of course, you can build such languages in Lisp, but
nobody has enough resources and incentive to re-invent such a big
wheel.)
For 10 years (1990-2000) I used mostly C++ for everything (including AI)
There are indeed book like "C/C++ for expert systems" explaining how to
program it in C++.
(Exceptionally crappy book by the way.)
My desire for speed was subdued as my machine is now 400 times faster than
the PC I
used in 1990. Now I use different programming languages for different jobs.
I prioritise programming productivity.
I think beginning programmers prefer to stick with one language. One of
the resons is they need to
focus on the tecniques and do not want to caught up in the clutter of
having to learn a new
language as well. After a while when you have most of the things (data
strucures and the
like) down pat this seems to be less of a problem. Altso I think the
primitivnes of java
may help when learning things like data structures. (But I expect opinions
will differ on this
point.) After a while you find struggeling with these details a waste of
time and then
time has come to switch to something like lisp.
On Tue, 25 May 2004 02:10:00 +0200, Andr� Thieme
<······························@justmail.de> wrote:
> Espen Vestre wrote:
>> Good grief! And this used to be a thriving lisp community! Is this
>> typical for AI/Computational Linguistics research centers these days?
>
> One could think that universities still use a lot CL and especially in
> AI it is heavily used. But well, many projects are getting rewritten in
> Java, even AI projects.
>
>
> Andr�
> --
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Andr� Thieme <······························@justmail.de> wrote in message news:<············@ulric.tng.de>...
> Espen Vestre wrote:
> > Good grief! And this used to be a thriving lisp community!
> > Is this typical for AI/Computational Linguistics research centers these
> > days?
>
> One could think that universities still use a lot CL and especially in
> AI it is heavily used. But well, many projects are getting rewritten in
> Java, even AI projects.
Good for them. If they like mucking around with multiple languages
(Java, XML, SQL, Jython, .net) then they're just assembling lisp by
using a bunch of different languages to counter each of their
weaknesses. The only difference is they're relying on some spooky
father figure for their macros.
Anyway, lisp is now viral. No more dependence on assumptions like
special-purpose hardware. Everyone wondered why the AI winter killed
dinosaurs like the Connection Machine and Symbolics, yet no one
noticed the little mammal Franz scurrying around.
From: Julian Stecklina
Subject: Re: ai without lisp
Date:
Message-ID: <86oeo855ze.fsf@web.de>
Andr� Thieme <······························@justmail.de> writes:
> Espen Vestre wrote:
>> Good grief! And this used to be a thriving lisp community! Is this
>> typical for AI/Computational Linguistics research centers these days?
>
> One could think that universities still use a lot CL and especially in
> AI it is heavily used. But well, many projects are getting rewritten
> in Java, even AI projects.
At the Institute for Construction Informatics (Institut f�r
Bauinformatik) at Dresden University of Technology there is
significant use of Common Lisp. But as the professor that introduced
Lisp is retired, its use is rapidly declining. I do not know whether
Java or something else fills the gap.
Regards,
--
Julian Stecklina
Signed and encrypted mail welcome.
Key-Server: pgp.mit.edu Key-ID: 0xD65B2AB5
FA38 DCD3 00EC 97B8 6DD8 D7CC 35D8 8D0E D65B 2AB5
Any sufficiently complicated C or Fortran program
contains an ad hoc informally-specified bug-ridden
slow implementation of half of Common Lisp.
- Greenspun's Tenth Rule of Programming