What are you doing in Common Lisp lately?

I'm updating CL-Selenium, messing around with Lispy's version storage approach and hacking on a Space Invaders clone.

I'm writing a Go (http://en.wikipedia.org/wiki/Go_(board_game)) AI for a competition with some friends.
(http://ai.mindstab.net)
I'm also trying to get it installed on a hardened PaX linux server so I can play around with it on the web.

I've just implemented a compiler that reads a specification file and outputs code that can encode/decode a communication protocol for an on-line game. Sure beats manually making code to for hundreds of packet types. Another advantage is that I can easily add back ends for many languages. Not to mention that using Lisp made the compiler coding fun

I have been working on a Linux shell written in Common Lisp. It's designed to be extended in interesting ways and Lisp provides the flexibility to do that easily.

findinglisp: Is there a way to test this CL shell?

pperez wrote:findinglisp: Is there a way to test this CL shell?

No, not yet. It's still a work in process and nowhere near ready for release.

Libraries! I just finished modifying a slightly aged bit of code in the CMU AI repository to make Gambol (basic prolog-like functionality).

My current big project is to make my SVG library lively enough to privide basic GUI behaviors. That means I have to learn Javascript and the whole DOM API first. Fortunately, parenscript has made this go more nicely, so far at least.

wm.annis wrote:Libraries! I just finished modifying a slightly aged bit of code in the CMU AI repository to make Gambol (basic prolog-like functionality).

Very cool. Lisp needs some more Prolog-like libraries. I started learning a bit of Prolog the other day, and while I really couldn't see writing a large program only in Prolog, I can see that it would be very handy for solving certain smaller problems within a larger Lisp program. One of the things I love about Lisp is that sexprs are so convenient for expressing things like Prolog within Lisp.

Learning it, essentially. I'm sick of reading articles about how it's awesome, will find me women and stitch up my socks etc etc.

Instead I want to be in a position to nod knowingly, and stroke a deservedly-earned beard of wisdom.

Also, I've expressed thoughts about writing the Last And Bestest Blog Engine Ever and I reckon I'd like to try my hand at doing it in Lisp. Mostly because I have come to detest Wordpress.

In these few hours, I'm writing a program that will help me manage my tea collection (I keep tea in my own boxes, and I always forget how long I should brew them, what to sweeten them with and so on). Yea, I could just make a list, but where's the fun? I'm also making it modular, so that I can write various frontends to it. I'm not sure why something that tells you how to make your tea should be this big, but what the heck...

I'm currently thinking about doing some experiments with retargetable assemblers, but I'm caught up with a bit of work right now.

findinglisp wrote:Very cool. Lisp needs some more Prolog-like libraries.

Several commercial Lisps have really good offerings for this, but the open choices seem to be mostly associated with books (PAIP, On Lisp) and so require a lot of dancing to disentangle from everything else that comes with the book — to say nothing of licensing confusions.

I started learning a bit of Prolog the other day, and while I really couldn't see writing a large program only in Prolog, I can see that it would be very handy for solving certain smaller problems within a larger Lisp program.

Yep. It's underrated as a database, at least by people who don't know it.

Writing video games with it http://nightschool.near-time.net

Writing a semi-commercial browser game (http://beta.thanandar.de/)

I'm working on two dynamic web pages, a simple one that's mainly an image voting system, and on a bigger one that manages flashcards, as a shared learning tool.

hi,
i'm currently working on a widget based web framework for Common Lisp: http://groups.google.com/group/symbolicweb

• not using continuations .. designing it as a simple "state machine" on the server .. history is maintained on the client instead; rationale is that this needs to happen anyway since one need to be able to reference to (link to) state in the application in the first place
• comet support .. doing things like (setf (background-color-of some-widget) "red") in the REPL (or wherever) propagates to the client instantly
• focus on getting the low level stuff right; dom, attributes, css, events, url/history/back-button .. then building widgets etc. on this
• some unique, i think, features like ability to share a single widget between multiple clients/users .. a container widget is particularly interesting for instance (see the chat example for a shared container widget in use)

.. at the moment i'm porting the "navigation, history and referenceability" support ( http://ajaxpatterns.org/Unique_URLs , i guess it's also called REST) to IE .. *sigh* .. we should all be billing extra for MS and IE support and mention that "this and that" is for adding support for MS's products

.. next up is adding support for "views" as mentioned in this thread; http://groups.google.com/group/symbolic ... 6c002ec6f1 .. i'm actually looking forward to working on this after this little IE nightmare ..

..anyway, i'm having a blast with lisp..

A Synchronous Event Framework. It will store closures in a hash. It will also be multi-broadcast, Each event can have more than 1 listener.

I've started a project called Ply, a 3D mesh editor. Well, actually my first milestone goal is to make it a basic 3D mesh editor. Currently though it does little more than load and display triangle meshes from an OBJ file (meaning OBJ spec not fully implemented - yet). If the design at the M1 stage is good, I might like to take it further. Ply is the first project I've ever really done in CL, so I'm of course looking for any one willing to critique my code.

I just wrote an RPC server that acts as a background synchronization mechanism between our Django-powered website and our back-end media system.

I'm maintaining a backend mobile search server, and adding new and exciting features to my SPARQL implementation. Too... busy...

I'm writing a hangman game in CLISP with cl-opengl on windows (2d). Next project I'll start to learn cl-sdl though. One sticking point for me was adapting my programming to the event based OOP that opengl uses. Once I did, however, some things simplified , and some got harder . I know some people look down on it, but I love it when I run the program, switch across and modify the code, go back to the running program and reload it. It's slick!

I'm writing a card-scanner configuration tool for my company. We have about 20 retirement communities, and are using three different types of proximity ID cards. Each type has its own scanner model, and requires its own custom (non-factory) configuration.

One of the sweet things about Lisp (some part of which I could have done in Ruby, my other language) ist that it prints structures in a format Lisp can read back. Thus, rather than custom-code a configuration format, I can take the printed output from a working card reader and include it as a list of configuration settings; Lisp will read it into the correct C-flavored structure (thanks to Swig and clisp).

Then, I can mapcar through the list, applying a generic function set-config to each entry: the GF's methods call the correct configuration function based on the structure that Lisp re-created from the original config output.

This gives me bit-level access to the configurations, without having to hand-code a config-to-struct transform, etc. In addition, the configuration entries are readable, since the Lisp structure I/O is readable: instead of

BIT_SETTINGS = [[0,0,16,26], [1,1,16,26], [1,1,25,32]]

as in my original Ruby implementation, I can represent (one part of this) like

(defvar mt-config
  (list #S(PCPROX-API::ID-BIT-COUNTS :LEAD-PARITY 1
                                     :TRAILING-PARITY 1
                                     :ID 16
                                     :TOTAL 26
                                     :IPAD4 0
                                     :IPAD5 0
                                     :IPAD6 0
                                     :IPAD7 0)
  ...))

I also have used Lisp and Parenscript with Hunchentoot to generate the first cut of a new employees display page, which provided a slideshow of new employees.

i'm working on a network packet dissector sitting on top of plokami
(http://www.cliki.net/plokami) for realtime network packet analysis.
The aim is to create interactive, lisp programmable packet sniffer/generator
something similar to scapy but much faster in realtime processing
(sbcl vs python = no contest).

I also started writing an opengl livecoding environment similar to fluxus entirely in CL.
Its at a very early stage but a sample video of the opengl repl is here:
http://tinyurl.com/6yh7zy (quicktime .mov)

I just did and sold a call shop management program. It routes phone calls to different providers based on the prefix, shows real time cost of calls in each phone booth, prints receipts, shows real time call information on lcd displays in each booth, does the usual statistics and so on. The program runs in each call shop, it is web based but the server sits on the same computer as the browser. I used hunchentoot and did all the ajax stuff on my own. This was done using lispworks.

This was my first lisp commercial project. Overall a pretty good tool for the job. I had some problems with application delivery, and had to fall back to some delphi code in a dll in order to do the receipt printing. I'm not too pleased with the process (as in windows process) handling in lispworks.

Most of my problems were coming from the compatibility layer with other OSes, I'd rather have a tool which is only directed toward windows (where are my customers). For instance there are too many layers to the GDI stuff, it's all undocumented, and the source code is not available. I couldn't build my own grid CAPI component because of this. I have pretty high standards and the provided example was way off. That's why i went for a web interface.

The lispworks people were very helpfull, when i found a bug they fixed it within a week.

And now i'm doing ..... the very same program for another company ! This time they'll want a centralized version as well, so i'm trying to use immutable persitent data structures all the way, which should allow for better concurrency. They're talking about 6000 callshops, that's a lot of requests per second ... we'll see how it goes !

Sacha

I'm creating something like a 3D Emacs that's a media editor instead of just a text editor.

Here's my page: http://cl-beer.blogspot.com/

I have a ton of ideas but right now I am using CFFI to create binding for Cairo [done, as a learning exercise], Pango [Almost Done], cl-OpenGl [Which seems to have broken with the latest cffi], and PixBuf. I imagine that it could display any type of data by abstracting the view from the data itself. Eventually I want to implement multiple users modifying shared data, encryption for communications, intrinsic version control and history, a Lisp sandbox, etc. I know it's a huge project but Lisp is making it possible.

Let me know what you think at my site.

Thanks!

WarWeasle

I am one of the maintainers working on the SDL, SDL_gfx, SDL_mixer, SDL_ttf, SDL_image and OpenRM bindings for the Lispbuilder project.

Building http://toonlet.com, a free make-your-own-webcomic site. It uses UCW and a bunch of other libs. Pretty much everything but the character tool, which is in flash, is lisp-powered.

vi bindings for Lispworks are inching along. (http://common-lisp.net/project/lw-vim-mode/ and http://theclapp.org//blog/tags/vim_mode/) Changing jobs soon, so not a lot of time or energy to work on it. Hoping to pitch Lisp to the new company. (Currently they're using Perl, mostly.)