Re: A92: Zip program for Ti-92


[Prev][Next][Index][Thread]

Re: A92: Zip program for Ti-92



Are there any places online that talk about all these different compression
algorithms? I've never heard of any of them. I have some ideas for compression, but
I would like to make sure they aren't the same thing as and those aren't better
than mine.

Bryan Christopher Turner wrote:

> >  I'm just writing one, which uses a combined Lempel-Ziv - Huffman algorithm.
> It
> > will work in this way, that you could create an archive at your PC, and send
> it to
> > your TI-92. The unpacker at the TI-92 will work like FBrowser, and will
> regard
> > archives as directories. To run a program, it will be uncompressed to a
> temporary
> > memory block, and executed by the core. If you have some infos about Huffman
> or
> > Lempel-Ziv or any other suggestions, mail me! (I haven't any texts about
> Lempel
> > Ziv or Huffman, so I had to develope my own algorithms)
> >
> >                                             David Kühling
>
>    The problem I found in creating a compressor w/i the calculator was two
> fold: the memory needed for decompression, and the fact that small data sets do
> not compress well with LZ77, LZ78 OR LZW.  Huffman is good, but requires a lot
> of useless overhead for the bitstrings.  I was examining a project which would
> compress and decompress programs using a huffman tree but arithmatic encoding,
> also the data would be compressed on top of the existing file, thus not
> requiring any additional memory.  Decoding would require some extra room
> though.
> I recommend a book called 'The Data Compression Bood, 2nd Ed.'.  Its very good.
>    I have not examined the run-length encoders (pplib and skittleslib both have
> one).  I doubt they work as well as Huffman or LZ on machine code though.
>
> --Bryan
> bcturner@eos.ncsu.edu




References: