Home > Archive > Compression > June 2007 > Re: compressing short XML messages without including dictionary or
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Re: compressing short XML messages without including dictionary or
|
|
| Josiah Carlson 2007-06-16, 3:56 am |
| collection60@googlemail.com wrote:
> On Jun 15, 10:48 pm, Josiah Carlson <josiah.carl...@sbcglobal.net>
> wrote:
>
> I have implemented the decoder ;) Just not the encoder, which will be
> many times harder than the decoder of course.
>
> And it's not navel gazing. It's exactly what I said before. I've been
> out of time. But thanks for the statements about someone you know
> nothing about.
If you don't have time, you don't have time. But I don't need to know
you (or anyone) to say that a compressor without code is worthless.
(Funny how you earlier posted about mathematicians getting famous for
equations and falling short on implementation ;)
However, when you talk about encoding data based on bytes, it makes me
think that you haven't come to the understanding that encoding/decoding
is the least of your worries. Put symbols and probabilities in a
huffman or arithmetic coder and move on.
If you need a huffman/arithmetic coder, speak up, I'm sure someone here
will be able to offer a link or an implementation in whatever language
you happen to be using.
The real trick is *what symbols* with *what probabilities* to use.
LZ77, LZ78, LZW, LZ-whatever, BWT-based, xml-preprocessors, PAQ, etc.,
are all about assuming that data has a particular set of features, which
are modeled in a particular way. Once you have your model, then the
algorithms are all about encoding the input (or decoding the output)
based on those models (or really assigning probabilities to the
huffman/arithmetic coders and sending/receiving a symbol).
Anyways. Take your time. But I stand by my original statement:
Implement it.
- Josiah
| |
|
|
|
|
|
|
|