I previously asked this

Where can I ask for algorithm recommendations?

And was asked to migrate my question to softrec. here


The thing is, there are only two algorithms, at least that I could find, I don't know whether both are bad, at least for me or whether the dictionaries or implementations are bad.

In any case, I feel like only 2 kind of people can answer my question, either a software engineer working on a language related project or a computer scientist working on research papers related to that subject.

No programming related SE site is willing to recommend algorithms, so I was forced to move it to soft rec, but I feel like it belongs here, I'm afraid to post it though. I'm not looking for an implementation, just an algorithm name.

  • $\begingroup$ Keep in mind that people have lives and the planet has timezones. After 15h, neither a question on meta nor on a main site has necessarily received authorative and/or good answers. $\endgroup$ – Raphael Jan 12 '19 at 8:53


  1. Are you looking for an algorithm -- an abstract thing usually expressed using pseudo code or mathematics -- for you to study or implement including further work?
    --> probably ontopic

    Or are you looking for a program, that is a ready-made implementation of an algorithm, for you to use in a larger program?
    --> definitely offtopic

  2. Is your question: "How can I solve this problem at all?"
    --> suited; any correct algorithm is a good answer.

    Or is it: "Which of the N algorithms is the best?"
    --> usually unsuited for SE since "best" is not well-defined. However, if you can describe well what your requirements are, it can work.

    Or: "What are all the algorithms for this problem?"
    --> unsuited as list question and too broad.

You seem to be looking for a program. While you have two options already, you are looking for a list of others that are "better" in an unspecified sense. That makes your question unsuited for cs.SE, and indeed most SE sites.

  • $\begingroup$ I'm not looking for a program, there are only 3 algorithms for transliteration, two of them work for Arabic, none are good, at least not to my case. I'm looking to see if there are other algorithms that fit my requirements $\endgroup$ – Lynob Jan 12 '19 at 9:08
  • $\begingroup$ Forgot the part where i talk about finding a python package $\endgroup$ – Lynob Jan 12 '19 at 9:09
  • $\begingroup$ My answer stands; try to answer the questions honestly for yourself. Keep in mind that even if the question is technically ontopic and suited, we may not have the experts for your niche. Linguistics or Data Science may have, but I don't know whether your question would be ontopic there. Maybe visit their respective chatrooms. $\endgroup$ – Raphael Jan 12 '19 at 10:16

In your specific situation, I would suggest that you avoid phrasing your question as a request for a recommendation. Recommendations are subjective, and subjective questions don't work well here.

Instead, if you want it to fit well on this site, state it as a problem that you want solved. Tell us what your requirements are, or what criteria you will use for evaluating solutions. Give us some background on what approaches you currently know of, and why they are unsuitable. Then your question might be suitable.

As a caveat, after all that, it's also possible that it might be unsuitable. If your question amounts to "solve this open research problem", it's probably not reasonable to expect that to happen in response to a question here. But it's the best shot I can see for taking your question in a direction that is suitable here.

For instance, your criteria might be "minimize error rate", or might be "error rate below 5%". If so, say so, and say how you plan to measure error rate. Then, tell us about what research you've already done. Summarize each of the approaches you are aware of, and what their error rate is, and why they are unsuitable.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .