CS

.....project ideas?

Requirements:

[list]
Something useful that can be later sold. Games are allowed but discouraged
Something that shows a lot of OOp implementation
Something hard to program. (I'm the best in class. I need to win the honors)
[/list]

I was originally going to do a OCR reader for hard written documents. Upon finishing the first research session I was aware the project had too high a chance of failure and too small a time frame to be completed successfully. I downgraded it to a simple captcha reader but again it was probably bound to fail.
I am now stuck without any ideas.

Can please /prog/ provide some ideas?

Now you have two problems. No exceptions.

Oh, you're the guy who made those threads about captchas.

Fibs
Regex-based sexpcode parser
Printing "Sussman" to the standard output

Something useful that can be later sold
You don't seriously expect to sell it, do you?

>>4
>Regex-based sexpcode parser

Do that 'un

Sounds like a web service of some sort. Maybe something non-obvious but useful either academically or for industries in your area. For the CS angle you could do some intelligent data mining/processing, maybe integrate OCR by using an existing engine.

Ask someone if they need something complex doing for them. You don't actually have to present them the finished product. It's standard A-level procedure.

Ask someone what they want written in C, write it in Java, tell them it's an ENTERPRISE solution.

>>6
Now you have ASTs full of problems.

(I'm the best in class. I need to win the honors)
[/list]

Community college.

Get the bug list of some open source project. Fix as many bugs as you can and contribute your fixes to that project.

An implementation of Scheme (R6RS) with the front end built with Antlr ( http://antlr.org ) and the back-end targeting LLVM ( http://llvm.org ) allowing for both JITing and generation of native machine code binaries.

>>4
Regex-based sexpcode parser
I thought that every 'languages and automaton' course teaches that matching patterns do not form a regular language.

Of course, as long as by 'regex' you mean 'regular expressions'. If, by chance, you mean 'perl regular expressions', you have two problems.

>>13
I'll accept R5RS.

>>15
R5RS is shit. R4RS or GTFO.

>>16
As far as I've been able to tell from the R4RS, if he provided the R4RS call/cc with a different name to the R5RS call/cc (say call-with-less-useful-current-continuation)it would be a strict superset.

>>1
you will suck dick for money.

>>17
It would be a strict superset in the same way that a cake covered with shit is a strict superset of a cake not covered with shit.

>>11

I'm perfectly confident that you appear more of a retard than me, given you are trying to judge my supposed intellect and decent problem solving skills on a single syntax mistake in BBCODE.

>>19
You have failed to explain which features have "covered it in shit". So lets go through the changes then
The empty list is now required to count as true.
Already the case in most implementations, compatible with IEEE Scheme and it quite frankly makes a lot of sense.
The classification of features as essential or inessential has been removed. There are now three classes of built-in procedures: primitive, library, and optional.
This is really a report change rather than a scheme change.
Programs are allowed to redefine built-in procedures. Doing so will not change the behavior of other built-in procedures.
Admittedly R4RS did not forbid this, it left such behaviour unspecified.
Port has been added to the list of disjoint types.
Did any implementation actually make use of the facility to have ports be another type, say string? I don't think they did. A disjoint type makes sense.
The macro appendix has been removed. High-level macros are now part of the main body of the report. The rewrite rules for derived expressions have been replaced with macro definitions. There are no reserved identifiers.
I could conceivably see why you would disagree with this, if you are in favour of define-macro, but it was already an appendix and moving into the main body of the report improves portability.
Multiple-value returns, eval, and dynamic-wind have been added.
OK, I'm sure this is where your beef lies, that is assuming you actually know what you are talking about. Multiple value returns are very useful, but IMO could be better expressed if you made all continuations take n arguments (where n >= 0). Dynamic wind is also useful, and like r5rs continuations, completely expressible with r4rs continuations, but as was pointed out previously, redefining call/cc would result in unspecified behaviour, and not necessarily what you would want.
Regarding the change in identifier syntax allowing '@' characters: this was done specifically in order to allow internet email addresses (a new concept since R4) to be represented as symbols.
This one was a mistake in light of the use of ,@ for unquote and I believe it is forbidden by R6RS.
R5 dropped the requirement for 'load', 'with-input-from-file', 'with-output-to-file', 'transcript-on', 'transcript-off'
If it were me, I'd drop 'transcript-on' and 'transcript-off' altogether. Maybe also load, it's really not necessary. I'd keep 'with-input-from-file' and 'with-output-to-file', because they are actually useful.

So there are all the changes, courtesy of the schemepunks wiki and the r5rs report. If actually have the ability to justify your statement, please do so.

Actually I missed out two

Syntax-rules now allows vector patterns.
Not really controversial, if you already accept syntax-rules

The calls that are required to be implemented in a properly tail-recursive fashion are defined explicitly.
I'm not sure what your beef with this one would be.

:|

How boring!

I want something fun to program. Something that takes more time writing the pseudo code rather than implementing it in a language.

I was going to go for a chess engine but that's already implemented a billion times and given most engines will be superior, my project won't make any good impressions.

Anything else?

>>23
maybe join a opensource project, find one an start hacking.
maybe even join the team if you are dedicated or just submit patches.

>>23
I could use an "election" module which serves to simulate the collection and aggregation of votes for various voting methods. Please use a decent language like C, Lua, or Scheme, though. Thx.

How about something simple, that can be as complex as you want?
Like a webcrawler, that's a nice and fun project.

You could do an orc for printed documents quite easily and load in a few fonts/neural network it. Captures are just less impressive imo. You do a simple web browser and have it interpret basic html. A web server, an IRC server, a bot for some game or website, could do a bot that grabs stock information and does some shitty prediction/analysis of the data. Um.... ya that's all I got that's not really a game.

How about a game where you have to perform task x that is otherwise not that complex, like a web server a la Game and Watch or something.

decent language like C, Lua, or Scheme
IHBT

Write a gopher server with mod_python support.

http://fare.tunes.org/computing/term-project-proposal.html

2.11 Building a Reflective Lisp Environment
In this project, the intern will develop a Reflective Lisp Environment, with introspective and intercessive capabilities such that it is possible to efficiently build on top of them such features as garbage collection, regions management, partial continuations, concurrency, distribution, code migration, orthogonal persistence of data and code, version control, configuration management, etc. All these features will be demonstrated. Both a Common Lisp layer and a Scheme layer will be built atop this environment.

The executive will be built on top of Ian Piumarta's COLA, or using similar techniques.

Community Benefits: This proposal, if completed, will provide the Lisp community at large with an environment capable of supporting live programming, and enjoying the benefits of an integrated environment capable of coherently automating meta-level tasks that programmers have to do manually in existing systems.

Intern Benefits: The successful intern will gain key insight on many aspects of the implementation of computing system.

Difficulty of the Task: This is a project that requires a lot of patience and debugging, especially when it comes to interfacing to existing systems, and overcoming the various quirks of hundreds of variants of the underlying hardware and other software that has to be interfaced with.

Do all the problems on Project Euler.

http://projecteuler.net/index.php?section=problems&page=6

A program which extracts youtube download urls from pages and downloads HD clisp to user specified directory with full titles.
A stable and fast SecondLife clone.
An alternative to Shiichan and/or 4chan which trumps both in features and speed.

>>33
An alternative to Shiichan and/or 4chan which trumps both in features and speed.
Textboards aren't improved by having more features than Shii_tchan (unless one of them is SexpCode support), and the amount of time it takes to execute the PHP scripts is negligible compared to the network latency.

Tetris

>>32

OP here. I already did quite a fuck ton of those.

Also thanks to anybody posting ideas.

I can't yet decide if anyone of the proposals are valid with preliminary research.

Thanks anyway!

Contribute to Goblin Camp and make it less boring

Create a blowjob robot.

>>38
This, it must be at least as good as the penistron, if not better

Create a fractal explorer that allows you to enter in the interation to run and change the colouring rules. Make it have smooth zoom then smoke weed and watch it zoom into the mandelbrot set for hours.

>>40
U MEAN DROP ACID

>>41
No, that would be redundant.

>>33
downloads HD clisp
Nice idea, but I'd prefer HD sbcl personally.

design and implement a language and a language runtime with a bytecode vm.

Make a web browser that doesn't suck.

>>45
Wahey!

Write an ANSI C compiler age 12.

and what is this

>>47
Fuck off.

Regex-based sexpcode parser
#!/usr/bin/perl

use feature (say);
use strict;
use JSON;

sub parse_sexp(;$);

sub parse_sexp(;$)
{ @_ = $_ unless @_;
  my @ret = ();
  if(shift =~ /^(\(((?:(?>[^()]*)|(?1))*)\))$/)
  { my $inner = $2;
    push @ret, $1 ? split ' ', $1 : parse_sexp $2
      while $inner =~ /([^()]+)|(\((?:(?>[^()]*)|(?2))*\))/g; }
  return \@ret; }

say to_json parse_sexp while <>;

>>45
http://surf.suckless.org/

52 Name: Anonymous 2011-02-04 17:25

Don't change these.

Name:		Email:
Entire Thread Thread List

All trademarks and copyrights on this site are owned by their respective parties. All comments and posts are the responsibility of their own posters.
- Tinychan + 2ch Mode -