[moritzwarhier]

I can't follow. If you'd disregard the size of the decompressor, you could encode as much data as you want in there, e.g. just assign a number to each article and include the corresponding text in the decompressor, or assign a number to each sentence occuring in Wikipedia, just dump the 1GB of data in there verbatim, etc...

That would make the size of the compressed file meaningless.

Disregarding the time complexity of the compression is interesting, even ignoring the space complexity during compression or the size of the compressor.

But the size of the decompressor can't be ignored when trying to "measure" Kolgomorov complexity of the source data.

[frutiger]

> I can't follow. If you'd disregard the size of the decompressor

GP is talking about excluding the size of the compressor not excluding the size of the decompressor which obviously has to be included.