Long story short, I designed a compression algorithm that performs well compared to some other compression algorithms.
I was wondering if it's acceptable to write a scientific paper on this algorithm even though there exist other algorithms which are faster and provide better compression ratios compared to mine.
Note: I've never written a scientific paper and am no longer in academia.
Yes. You can't tell before performing the research if the algorithm will be better than old ones. Now that you've done it and found the answer is "no", that's still a result. It won't be the most interesting of results so you won't be able to publish it in top journals, but it's still something.
As an example, see the iterated prisoner's dilemma. There's a simple, robust strategy discovered early on: tit-for-tat. However, people continued to study other strategies - a close parallel to your situation.
I would say that you are fairly unlikely to find a home for such a paper unless your algorithm has some unique or new characteristics. If it is especially good on some kinds of data sets that are interesting then you stand a better chance. Likewise if the algorithm itself is interesting and can lead to deep thinking on some aspects of the problem. Can you adapt the algorithm to some special use? Does it have interesting characteristics of speed or memory usage or ...
Or even is it especially easy to teach or to validate or ...?
If there is anything unique about your algorithm you have the possibility of publishing a paper about it. Otherwise it is likely interesting only to you and a few others. In that case just try to leverage the learning you needed to develop it into the next great thing.
