• files restructure code

    From mark lewis@1:3634/12 to Ozz Nixon on Mon Jul 15 12:08:26 2019
    Re: files restructure code
    By: Ozz Nixon to All on Sun Mar 17 2019 19:07:37

    Here is what I have in Xenophobe BBS as a files.bbs sturcture:

    FILENAME:12
    SPACE
    SIZE:9
    SPACE
    SPACE
    DATE:8 MM/DD/YY
    SPACE
    SPACE
    DESCRIPTION:46

    i was rereading some old mail and came across this... initially i like it but then my gut kicks and say why? why keep restricting file names to the old DOS 8.3 format? granted, we're not yet moving LFNs around but why restrict ourselves from being able to do so?

    just a thought or two...

    )\/(ark
    --- SBBSecho 3.07-Linux
    * Origin: SouthEast Star Mail HUB - SESTAR (1:3634/12)
  • From Ozz Nixon@1:1/123 to mark lewis on Tue Jul 16 02:03:30 2019
    On 2019-07-15 14:08:26 +0000, mark lewis -> Ozz Nixon said:

    Re: files restructure code
    By: Ozz Nixon to All on Sun Mar 17 2019 19:07:37

    [snipped]
    i was rereading some old mail and came across this... initially i like it
    but
    then my gut kicks and say why? why keep restricting file names to the old
    DOS
    8.3 format? granted, we're not yet moving LFNs around but why restrict ourselves from being able to do so?

    just a thought or two...

    Not a problem... in PCB 16, I have actually opened all to LFN, Virtual Directories, and using #tags and descriptions, as elements to navigate
    the file system. It is a lot more effort for a sysop to merge in a CD,
    or folders - however, it allows for:

    LFN
    SIZE (64bit)
    DATE (Epoch 64bit)
    DESC (45x10)
    TAGS: Unlimited Link List

    The "Library" section of the BBS supports:
    Library 1: Night Owl CDs (which is a tag #nightowl, applied to all the
    files from the night own collection), then I can support #vol1, #vol2,
    etc. Then I address the file itself, #game, #win32, #256color,
    #freeware, etc. then the search engine incorporated the description (1
    liner from old FILES.BBS, unzips and looks for DESC, FILE_ID, etc. and
    lastly I am working on logic to pull from README.* files). So the more
    words available, the tigher/more accurate a search could be.

    Like: Search %lewis #pascal #source

    %(anything) incorporates: SoundEx, Metaphone, Levenshtein, Similarity comparison of (anything) against each word in the DESC field. That
    result list is then filtered and ordered by LFN or DATE (users option),
    where TAGS contain #pascal and #source.

    I am developing all this without SQL, and experimenting with different indexing trees to try and make this lightning fast. Cool part, is I
    name flag my result as a Virtual User Directory - and share it w/ other
    users via "Users Library" or, store it for later downloads (incase I
    exceed my ratios). ;-)

    Been developing this for 2 months straight... going bald pulling hair ;-)

    Ozz

    --- ExchangeBBS NNTP Server v3.1/Linux64
    * Origin: (1:1/123)