We should take into account that we never will have precise and mechanically applicable rules

• proper names of individual persons, places, buildings, organizations etc. and other stable specific names of local phenomena should be in original language (transliterated to Latin letters if needed), in the official or most common form; however the English exonym or endonym should be used as far as it is widespread or (officially) established or commonly found in English-language documents. Ad hoc made English translations of proper names are undesirable.
• terms which don’t exist in English or that their English form is strange should be used in original language or in the most widespread form

We should take into account that we never will have precise and mechanically applicable rules

• “transliterated in Latin letters”: so words such as user:SJu would be forbidden as they are not Latin letters. On one hand, that would increase my comfort as it is a pain the ass to have always to find a copy of it for pasting (autofill in search does not work on it), on the other hand it seems too restrictive to me. On the other end of the scale with names like Category:Hinh truy?n len b?i thanh vien Luu Ly seems a bridge too far. In the past, I tried to propose a compromise but no one seemded interested.
• With the current imprecise definition, any non-English name can be used for anything, it can be abused as the “proper name” term has been abused many times. The terms exonyms and endonyms are confused with proper and local names. The term “most widespread” use is the perfect way to open endless debates.
• This potential Commons rule precision is indeed needed but it will be very difficult to find an acceptable compromise. This will be even more difficult as many people are only interested in using their own local names and have very little consideration for the overall Commons needs.

My reply to your comment from , is included in my comment from , (UTC). (Even English-speaking user can find by Google or in a map or in a train timetable rather “Karlovy Vary dolni nadrazi” than “Carl’s Wells Lower Station” or some hybrid name). The question what form of the local name can be found by an English-speaking user in most accessible maps (and what type of map is adequate to such type of subject) is a reasonable way to choice of a optimal name form. The second criterion should be consistency: when two or three Prague squares (bridges, streets. ) have established English forms of names but hundred others haven’t, we should decide whether a benefit from translated forms is greater than a deviantion from the official forms. This is regular dilemma. But it’s out of doubt that we shouldn’t use some “creative” ad-hoc translations and semi-translations of toponyma, especially of officially stated names. It is abolutely evident that some local names should be used in English and other un-English names should be used in original form. We have to understand benefits and disavantages of both possibilities and on the basis of comparison and counterbalance of them we have to find optimal solutions for various types of names and for individual boundary cases. Sensible consideration is more feasible than assertion of some unconsidered fundamentalist thesis. Latin-based alphabets of many languages are purposely build as compatible upgrade of basic Latin letter set. Latin letters with diacritical marks don’t need transliteration as far as software facilitates them almost fully. An international project like Commons should be able to not mutilate names like Sor-Trondelag, Resita or Kardasova Recice. We should have respect to the current practice and policy. A constructive discussion aimed at searching of optimal evenness is welcome. Stubborn negation without any constructive and realistic proposal would be waste of time. Current “de facto policy” of Commons is broader, more multifarious and more proven than only Foroa’s current “de facto policy”. Let’s go out from efficient experiences, not only from an extremist and unrealistic thesis. The general rule can describe essential reasons and prevent from most extremist bias only. Many specialized problems remain to solution. E. g., Umgebindehaus is a special German type of house. As far as this term exists in German, Czech and Polish language only, the category cannot have an English name. But when we use the German term, should we use the German form of plural (Umgebindehauser)? Singular is infringement of general naming rule, German plural can be unintelligible for English-speaking users, mixing of German word and English grammar is problematic as well. (We could find an English term for “Umgebindehaus” possibly but there exist many other similar cases.) –SJu ( talk ) , (UTC)