Re: Re: [SQL] Re: [GENERAL] lztext and compression ratios...

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
>> Unless the differential is huge we might want to just skip the
>> patent worries.

> If we found later that there was a patent problem, could we just issue a
> new release with a new compression algorithm?

Yeah, we could, and it could presumably even be a fully compatible
dot-release with no change to the on-disk representation. That
representation and consequently the decompression algorithm are safe
enough, it's the details of the compressor's search for matching
patterns that are a patent minefield.

However, changing the code after-the-fact might not be enough to keep us
from being sued :-(. I'd rather use something that's pretty well
established as being in the clear...

regards, tom lane