How mature/stable is design of LZMA algorithm

  • My question is not necessarily about the 7-zip application software or the state of the general LZMA source code....I was wondering if the LZMA algorithm itself was undergoing changes, or was it very stable, with only parameters, optimization or porting changes occuring?

    If it is very mature/stable, has someone thought of presenting it to the IETF as a standard.  A quick review of its performance seems to indicate that it can idealy compress text twice as efficiently (size wise) than that of standard deflate.  If this is true, I think that LZMA (because of its open software compatibility) could quickly suppliment deflate in the apache web servers and mozilla browsers.  It is my understanding that a new version of Apache 2.1/2.2 is getting ready for release, and that this version better supports caching of pre-compressed html content.

    Thank You in advance.

    • Igor Pavlov
      Igor Pavlov

      LZMA algorithm is stable. It means that decompression code is same. Compression code can change, but it will be compatible
      with decompression code.

      In comparing with deflate for small text files (html), LZMA doesn't provide much more better compression. The main goals of LZMA - better compression for big files and files of mixed/complex structures.

    • LZMA as a standard.. ineresting


    • LZMA could still be very usefull for web sites. One of the best rules of thumbs I have every heard was to "design for the worst cases, since the best case takes care of itself".  LZMA would take care of the worst cases... large text sites, such as "slashdot", and other blogs. Increasingly RSS feeds have become a problem... do to their popularity. A recent solution to have permeated the RSS field was to enable the use of the "deflate" on RSS feeds.  These RSS feeds are large amounts of text, and would even do much better with LZMA.

      As I understand it, the current IETF standard for deflate involved the browser telling the server that it supports deflate. The web server then has the option to return either plain text or deflated text. A possible extension to this concept, would be if a browser requested LZMA compression, it would get back either LZMA, deflate or text...depending on the web server.

      The creation of a open source IETF LZMA standard would see a proliferation of the use of LZMA in many
      more applications.