Admin please respond (regarding submission question) : General Discussion

Hello, I am in the middle of a programming project which will allow for new functionality to be quickly and easily implemented to bots for any game. (technical details below)

Since I am not a premium user here and it is not for any single game where should I post it? Since it is inherently technical in nature I am concerned that it will get modded down unfairly by uninformed users who perceive it as useless because it is not a ready-to-run bot, or because similar functionality exists in a more limited scope elsewhere.

I have had that experience before - not only that, but my code was ripped off by the same community that denied its usefulness.

- - - - -

And here are the technical details:

I am currently working on an optimized Optical Character Recognition (OCR) library for integration into any arbitrary AutoIt bot. It is designed to run in linear, O(n), time by predetermining relative salience amongst known values.

This allows the bot to recognize complex patterns orders of magnitude more rapidly then more clumsy methods. This approach requires a large, one time investment at the time of the bot's creation in exchange for rapid recognition (ie, only the bot designer must do this, and only once). The result is that searches which would else wise take several seconds are completed almost instantly.

The originally intended application was for identifying inventory items so that a bot could choose what to sell and what to keep, but I am confident that it can be scaled to more ambitious purposes, like optical recognition of a finite library of words (being able to read on-the-fly the specific stats on an item).

This approach benefits hugely from a limited search scope - that is, that my OCR tool does not have to search for the proper offset, as it can be reasonably assumed that the offset is determinable through other means. While this could theoretically be a weakness in certain games I have not found any examples in the wild and doubt that I will. This method of OCR is fully resilient against partially transparent and unpredictable backgrounds as it only looks at the foreground.