I'm a Lisp newbie, and having fun learning how things work.  I've been
reading the newsgroups for a couple of months, but haven't gotten a
great impresssion of what people are *actually* using Lisp for.

I'd go first, but have no active project.

Any takers for the question "What are you doing in Lisp, right now?"

Tait

Writing a Telephony Based Appointment Confirmation System for a group
of Ear/Nose/Throat Doctors.  I am working with Dave Bakhash (PortusGroup).

http://www.portusgroup.com

LispWorks, Windows and Linux.  CAPI Interface, Client/Server (Both written in Lisp).

- Writing something that will talk to an imap server and make an archive
of an email account that I have no access to except via imap.  I've got
something working using cl-postoffice, but now I'm thinking I should have
used mel-base... oh well, in any case it's just about working... except it
writes mbox format that other programs don't seem to like reading very
much.  Must get around to fixing that some time.

- Converting an old project of mine
(http://www.abstractnonsense.com/portfolio/ocr.html) from Java to Lisp,
and being constantly amazed at how much smaller it's getting.  I've
converted about 90% of it now.

Writing macro-writing utilities, and trying to figure out if it's possible
to get lazy evalution to work.

(2) I am developing a portable GUI framework involving OpenGL, OpenAL,
arbitrary graphics files and more so I can...

(3) ...resurrect a line of commercial educational software.

For what would you like to use Lisp?

First, i have a macro that automagically generates a clsql view-class
from a table in my postgres database. This includes setting up all the
relationships with the proper :join-class info. The hardest part of this
was figuring out the postgres system tables.

The macro also defines a view-class-metadata method that returns some
info about the table (col-names, types, pkeys, fkeys, etc.)

Using the created view-classes and metadata, i then define a UCW
componenent with a custom sql-view-class parent class that can
autmatically display/query/edit/add/etc by calling a more specific
component (view-class-search say) with itself as a slot.

(UCW is a great web-app development platform, similar to struts or
seaside, only quite a bit better IMO).

What this gets me is a complete basic CRUD screen with roughly 2 lines
of code (connect to database, generate classes). Then it's simply a
matter of specifiying the methods on my new classes to change the
default behavior. I can then whip up a quick data driven web app in a
few short hours of coding.

Eventually, the goal is to create an Access-like package on top of
mozilla and UCW, and allow my clients to create minimal databases with
crud screens, simple by creating tables in the DB (which can be done via
a wizard interface .. no sql needed)

  Since it's all generated from a postgres database, and a lot of the
functionality of the interface derives from a proper specification of
primary and foriegn keys, i think the databases will be  of a higher
quality then Access-msjet type projects, even though they are written by
the same class of user.

I was inspired by a similar package written in Perl (Maypole). I found
that while it created the defaults very well, as soon as you tried to
change it you ran into hackery. With my lisp approach, everything is
seamless, and simple to extend/modify. A lot of it's ease is due to the
incredible UCW package, which makes web-dev simple, as it handles all
the HTTP bits for you. I suggest you check it out if you are a web
developer.

oh, and of course, clsql rox0rs!

1. A code-walker for the commercial product produced by my employer. It
walks the full source code of the product (which is written in 5
programming languages), finds every bit of code that emits any text to
the Linux system logging facility, parses out the messages, fnids
references to variables and constants and substitutes their values, and
writes a nicely-formatted report in which all the messages are sorted
according to logging level.

(Something I'm doing in return for non-monetary consideration:)

2. Figuring out how to interface a certain Lisp to the OSX UI frameworks.

(Something I'm doing to help some friends get some business deals:)

3. Client code to suck very large datasets from publicly-available
databases (mostly genome and proteome records) and parse them into tree
structures that are easy to search, annotate, and update.

(Things I'm doing because they are fun and/or someone would like me to
do them:)

4. A programmable word processor (Alpaca) for Mac OS X.

5. An economic simulation game (Hansa2).

6. A Lisp IDE for Mac OS X.

7. The framework code and tools used to build (4), (5), and (6) (Bosco).

8. A VM suitable for running the output of Lisp compilers.

Yes, this is too many projects. Yes, it means that releases take a long
time.

1. A web application which dynamically generates PDF, RTF, or HTML
   from XSL-FO templates. Data is filled in from a hierarchical
   database which is itself extended and modified on the fly by the
   templates when they're uploaded. We use TBNL/mod_lisp plus CL-WHO
   for the web app and CLSQL/PostgreSQL for backups. For XML parsing
   we prototyped with XMLS but will now switch to either CXML or
   Franz' new SAX parser. The conversion from XSL-FO to the
   afore-mentioned output formats is done with external Java tools.

2. Another web application which gathers data from surveys done by
   pharma companies and produces up-to-date graphical output of the
   current results for the client. The data is uploaded as CSV data
   and the graphical output can be modified by the guy who coordinates
   the surveys with a simple Lispy configuration file - he's not a
   programmer himself. These are static pages created by a cron job
   using CL-GD and CL-WHO.

3. Maintenance of a Windows app (done with LispWorks' CAPI) I wrote
   about a year ago. It is used by my client to assemble data from
   various websites and generates CSV output for MS Excel. It makes
   heavy use of CL-PPCRE to extract the relevant data from the various
   sites it encounters. (No, this is not used for unsolicited
   marketing.)

For fun:

4. A .NET layer for Common Lisp, currently for LispWorks. I've stopped
   working on it after I showed an initial prototype in Amsterdam but
   I'll continue real soon now... :)

Cheers,
Edi.

--

Lisp is not dead, it just smells funny.

When I finish the book I will--finally--get to pop the stack back to
my long delayed project to implement a genetic programming system for
breeding critters that play the game of Go.

-Peter

--
Peter Seibel                                      pe...@javamonkey.com

I'm writing, in Common Lisp, a programming language for writing
models of human cognition.

We have been using it in its evolving state for about 5 years now and
it is giving us about a 10-1 productivity gain in the initial
development. Since the generated code is (now) essentially bug free,
this also saves LOTS of time. And it is MUCH easier to evolve programs
too - more savings.