Imagine a Latin text—e.g. the first verse of the Latin Vulgate Bible:
In principio creavit Deus caelum et terram Terra autem erat inanis et vacua et tenebrae super faciem abyssi et spiritus Dei ferebatur super aquas.
Most of us don’t know Latin, but at least we can read it, and guess at the meaning of some words—or look them up.
After further help from Rob Dunsmuir of Custom Coding, version 0.99 of the Square Hebrew-Simulator is now up and running and available for public use.
I don’t know if Israel counts as a “top country” in this regard, but it does have a problem with transliteration of Hebrew into English (or in Roman characters in general).
Essentially, the problem is one of what is known in information theory as information loss: in the process of converting from Hebrew to English, information about the spelling of words (including the distinctions between letters that are essentially homophonic, such as ḥet and khaf, tet and tav, kaf and quf, samekh and sin, etc.) is lost, so if you were to try and convert back to Hebrew, you wouldn’t be able to reconstitute the original spelling—even if you were a native Hebrew speaker and recognised the words, as in the following example, from a poem of homophones titled Modern-day Ecclesiastes:
Ask the average Westerner what Hebrew looks like, and they would probably imagine something like this:
(My answer to this question at Quora.com)
The biggest misconception, by far, is that it is written in the “Hebrew script”.