Results 1 to 21 of 21

Thread: A new idea

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts

    A new idea

    Good day, I'm asking if my work can be taken seriously on this forum. All of my previous efforts to get ahead thus far have been met only with blank emails, rejections and redicule. On the brighter side my work has not yet been disclosed. Ironically I've formulated a mathematical algorithm that's potentially capable of data compression far beyond the present paradigm. My compression algorithm is a bit unorthodox and may be quite revolutionay. Mathematically
    it can compress large quantities of data down to extremely small sizes with a minimal quantity of processes.( E.g 1Gb to less than 1 Mb) Albeit the fact that this may sound absurd and hence the reason why Iv thus far been met only with ridicule, its absolutely quite possible. My math somewhat finds its way around kolgomorov complexity. I can prove this, but I would like to do a little more than proofs. I would like to see this great work made a reality. I am however more of a mathematician than a software writer, and unfortunately I cant patent a mathematical algorithm. I'm looking for support/partnership to convert my algorithm to a working software. Is that possible on this forum? Is there anyone here who's willing to jump on board to be a part of this great endeavour? I will greatly appreciate any and all feedback and comments. I graciously look forward to a reply...

    Sebastian Phillip Sankar

  2. #2
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    742
    Thanks
    424
    Thanked 486 Times in 260 Posts
    Welcome Sebastianpsankar!

    New algorithms are always welcome!

    Reading your post, I get, that you have an idea, and it is not tested, yet. So how do you know that "it can compress large quantities of data down to extremely small sizes with a minimal quantity of processes"?

  3. #3
    Member
    Join Date
    Oct 2015
    Location
    Belgium
    Posts
    88
    Thanks
    15
    Thanked 64 Times in 34 Posts
    For what kind of data would you get such impressive compression ratios? Obviously it cannot be for arbitrary data.

  4. #4
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Quote Originally Posted by Jon Sneyers View Post
    For what kind of data would you get such impressive compression ratios? Obviously it cannot be for arbitrary data.
    Good day gotty, thanks for the reply, appreciate the response.

    The two concerns you pointed out there are : 1 how do I know my system can perform as described and 2 what kinds of data can it compress?

    My system functions like a mathematical formular, the outcome can be calculated. My system deal with binary sequences, so, with a particular description of binary sequences the outcome can be calculated/predicted/foreseen considering the nature of the mathematics involved. How I know it works is due to mathematical logic. (Considering the nature of my system and the nature of binary systems as it is)

    The second concern is what type of data can my system achieve the compression ratio described:

    This is the technical part, and this where I mostly lose attention, so I will be a little bit more transparent than I usually am. You pointed out before that it obviously cannot be arbitrary data. Now, most if not all data compression system's functions by removing redundancies/repetitions. After all repetitions and redundancies are removed, the data at hand is now random bit sequences and no further compression can take place. Further more if a quantity of data is random at The very beginning, there would be Little or no compression to be done at all.

    The 'thing' is, my system does not compress data by taking advantage of redundancies. This is the tricky part, my system takes advantage of chaos, and randomness, the more random and chaotic the binary sequence is, the more compression can be achieved by my algorithm. If there are alot of redundancies in the data to be compressed, the usual huffman system or any other will be used to achieve all the compression that possibly can until the bit sequence has nothing more to compress and then my system can take advantage of The randomness in The bit sequence that remains to compress further. However, using only my system alone, even if the data at hand contains large amounts of repititions, those repititions can removed by a scrambling formular. The chaotic sequence that remains will then be comfortably compressed by my chaotic compression system.

    What this implies is a bit controversial off course, it implies that even arbitrary data can be reduced down to incredible sizes. What does this mean for kolgomorov complexity, and the pigeon hole principle? The problem comes from the usage of the term compression. Compression in simple is about redundancies and repititions. My algorithm deal with chaos, and random data. For this reason, I'm uncomfortable calling my system a compression algorithm. Compression is limited. If we are dealing with binary as a closed system then 10 binary digits e.g 1001011001 can only represent 2^10 possibilities - 1024. A bit sequence of 11 digits ( 2^11 - 2048 possibilities) can represent more possibilities than 10 binary digits can. This is off course the reason why only certain types of data can be compressed, and furthermore only certain types of data can be compressed to high ratios. So why isn't this the case with my algorithm. In short:

    My algorithm can represent large quantities of arbitrary data using extremely Small amounts of binary sequences. Its not so much about compression, is more about how information is stored and represented. I really do not like this example, but for the sake of clarity. Dna does something like it. My algorithm mutates data into a sophisticated pattern of storage that allows for large sums of data to be stored in extremely small spaces. This is new. I'm a mathematician though. Not a programmer. What are you thoughts and opinions, what about support I'd really love some programmer's to work with. Benefactors as well?

  5. #5
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    742
    Thanks
    424
    Thanked 486 Times in 260 Posts
    Quote Originally Posted by Sebastianpsankar View Post
    This is the tricky part, my system takes advantage of chaos, and randomness, the more random and chaotic the binary sequence is, the more compression can be achieved by my algorithm. [...] The chaotic sequence that remains will then be comfortably compressed by my chaotic compression system.
    Don't be surprised when this thread is moved to Random compression.
    Usually people with random compression ideas have claims with no proof. There are a lot of such cases with 100% failure rate.
    They have no mathematical nor software engineering background. (So they don't have a sense that what they claim does not work.)
    I think they are after responses like "Wow!", "Amazing!", or they think they can get money from investors to work out their new idea.

    Sebastianpsankar, please, read the threads in the Random compression subforum - there you will find similar ideas to yours, and see where it usually goes. It goes nowhere. With a 100% chance.
    If you are interested in data compression, you are very welcome here. But if you insist to be able to do something without proof please don't expect much from us. Most forum members here understand data compression, they have studied and tried many ideas. And they know firsthand what random data is and where are the limits of data compression. What we know from your opening post is that you didn't try your idea yet. Please try, learn, experience. (My suggestion is the same to what JamesB suggested above.) If you think it's OK to disclose your idea, please share it with us, so we can help you to see where is the error in the logic. (As xinix suggested.)

    You may also want to read the head topic in the Random compression subforum: https://encode.su/threads/3135-random-compression-FAQ.

  6. Thanks (3):

    JamesB (2nd January 2021),Shelwien (1st January 2021),xinix (1st January 2021)

  7. #6
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Thanks for the info ....

  8. #7
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hey guys, is it possible to patent an algorithm? There's no source code involved, no programming language. Let say I have an algorithm that mostly tables, maths, and proceedures. Can that be patented or legally protected?

  9. #8
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Can you patent a mathematical algorithm?

  10. #9
    Member lz77's Avatar
    Join Date
    Jan 2016
    Location
    Russia
    Posts
    176
    Thanks
    60
    Thanked 16 Times in 12 Posts
    This is somewhat reminiscent of a email from the widow of a Nigerian dictator for help in obtaining his inheritance... I received this email twice...

  11. #10
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    A new idea

  12. #11
    Member
    Join Date
    Dec 2013
    Location
    Italy
    Posts
    517
    Thanks
    25
    Thanked 45 Times in 37 Posts
    Hello, start giving your CV right here.
    Thank you

  13. #12
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    ...

  14. #13
    Member
    Join Date
    Dec 2013
    Location
    Italy
    Posts
    517
    Thanks
    25
    Thanked 45 Times in 37 Posts
    What is your education?
    Degree?

  15. #14
    Member
    Join Date
    Oct 2015
    Location
    Belgium
    Posts
    88
    Thanks
    15
    Thanked 64 Times in 34 Posts
    You can't just circumvent the pigeon hole principle with 'sophisticated patterns of storage'. You can't represent more than 2^n different uncompressed bitstrings using n compressed bits, and that's not a question of lack of imagination, but just a matter of counting. If decoding a given string produces a unique result, that is.

  16. #15
    Member
    Join Date
    Dec 2011
    Location
    Cambridge, UK
    Posts
    528
    Thanks
    204
    Thanked 187 Times in 128 Posts
    Agreed, the counting argument means no compression method works on all data. None of them. Not even cmix.

    However, most data we care about isn't random. Hence why tools work. Any time someone proposes a method which works recursively then it falls foul of the counting argument again (as by the first pass it's pretty much random data again). Or any time someone proposes a method that offers "too good to be true" ratios without clarifying what the input data is then it's also likely a mistake.

    I get that you *think* you have a cool method that works on all data, but those with experience know you'll have made a mistake somewhere in your thinking. It's unkind to ridicule, but you have to face the undeniable logic and facts. If you insist otherwise, then no one is likely to help you as we know it's a dead end. The best thing you could do is to code it up yourself so you can learn where you went wrong, and take the learning to code bit as a bonus along the way.

  17. Thanks (2):

    hexagone (3rd January 2021),xinix (1st January 2021)

  18. #16
    Member
    Join Date
    Nov 2015
    Location
    -
    Posts
    75
    Thanks
    302
    Thanked 18 Times in 14 Posts
    Sebastianpsankar
    You can tell the essence of the algorithm in the abstract.
    And we can tell you where your error is, you don't have to program to check it.
    But you need some information.
    The "new idea" kind of information is not enough to help you

  19. #17
    Member
    Join Date
    Oct 2015
    Location
    Belgium
    Posts
    88
    Thanks
    15
    Thanked 64 Times in 34 Posts
    Yes, you kind of can. But I would suggest to first try if it actually works, before you're going to pay a patent lawyer and all that.

  20. #18
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts
    I wasn't going to a patent lawyer, was just going to a patent office and file a patent. Doesn't cost much to just file it, that way I can disclose it to actual programmers for evaluation and testing. Unless I find someone willing to sign a non disclosure agreement. But my work is apparently nonsense so no one will waste there time to sign any agreement to my work. Only option is to patent, n disclosure. But then my patent can be infringed on, since my patent is a bit abstract, the essence of my concept can be hijacked.

  21. #19
    Member JamesWasil's Avatar
    Join Date
    Dec 2017
    Location
    Arizona
    Posts
    111
    Thanks
    96
    Thanked 19 Times in 18 Posts
    Quote Originally Posted by Sebastianpsankar View Post
    I wasn't going to a patent lawyer, was just going to a patent office and file a patent. Doesn't cost much to just file it, that way I can disclose it to actual programmers for evaluation and testing. Unless I find someone willing to sign a non disclosure agreement. But my work is apparently nonsense so no one will waste there time to sign any agreement to my work. Only option is to patent, n disclosure. But then my patent can be infringed on, since my patent is a bit abstract, the essence of my concept can be hijacked.
    After reading this, I will try to be as respectful as possible and address the above.

    I would consult a professional patent attorney before you pay out a significant amount of money, because there may be ways to do this, but as a utility or design patent for a hardware circuit implementation which is physically tangible original work and as such, can be patented even if you find that your approach and design cannot for whatever reason.

    There are times that a patent-pending status can be more advantageous and offer greater preliminary protections than a completed patent without other protections after it is approved. Find out if that is the case.

    I would say too, as everyone above has suggested, to make sure that your algorithm works before trying to patent it. While yes, you can get a patent for things that are not officially workable or that have errors (many have and continue to from the USPTO as long as it SEEMS to work), make sure that it is not too vague or obvious that it does not, because if the patent office (assuming here that you are doing a US patent but any country's patent office) does not see a reasonable way for it to work, then they are going to reject the application.

    Finding a programmer is not going to be an easy task, because programmers generally do not like to work for free as a trade. Those who are willing to have other devotions to why they are making software, but in each of these reasons they are guaranteed that they will not be wasting their time whether the completed project is paid or free. How do you guarantee them that with any proposal you make? Maybe you can offer something in return, but if it is contingency-based, it is like finding a lawyer who will take a case pro-bono or on contingency, which is equally as rare unless they are sure or guaranteed something if they do.

    Another option you have is to try your algorithm out with MATLAB.

    It is designed more for mathematical expressions and testing than anything else, but if you can prove that your mathematical formulation works with MATLAB, you may be able to get a programmer to bite and agree that way.

    But again, it would have to a) be understandable fully and b) proven to work on that.

    If you can go this route, it may be able to fix the problem you're having and help you find what you're needing.

    No one knows for sure whether your particular algorithm will work or not.

    There have been too many people wanting free programmers who said their algorithm worked but did not, and they have done this from here to Usenet for well over 30 years. Because of this, you will have to find a way to prove that it does.

    If you can do this, then you will have opportunities regardless your traditional credentials.

    Many people have used traditional credentials to scam and undermine others, so credentials mean little to me compared to the authenticity and track record of your successfully completed projects which proves all, and is far better and more valuable than any certificate, degree, or resume could be.

    Proof, evidence, and experience are king.

    Show the proof that works and you will have people asking to write the software for you.

  22. #20
    Member
    Join Date
    Dec 2020
    Location
    Trinidad & Tobago
    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Compression

    This was helpful and encouraging... Thanks...

  23. #21
    Member
    Join Date
    Nov 2015
    Location
    -
    Posts
    75
    Thanks
    302
    Thanked 18 Times in 14 Posts
    ​I don't get it!What do you want from us?

Similar Threads

  1. Idea
    By Rudi in forum Data Compression
    Replies: 9
    Last Post: 2nd February 2017, 00:08
  2. XORpressor. Yet another idea for a compressor
    By Cristo in forum Data Compression
    Replies: 5
    Last Post: 24th August 2016, 17:29
  3. Compr idea
    By Rudi in forum Data Compression
    Replies: 0
    Last Post: 8th May 2016, 02:06
  4. StartMT (an idea)
    By SvenBent in forum Data Compression
    Replies: 4
    Last Post: 13th May 2014, 19:07
  5. Idea: Smoothed probability
    By Piotr Tarsa in forum Data Compression
    Replies: 2
    Last Post: 19th February 2013, 21:30

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •