¿What about adapting MsWord into ID?

Question

This post is related to how to confront or (un)manage large portions of text, coming from Word.
First scenario
[Almost perfect... although it happens 5% in the daily life].
Procedure: Simply you apply your main body style to the incoming word galley and immediately the whole text is converted, preserving italics and bold characters.
Pros: You get an instant translation
Cons: You may lose styles previously tagged originally in word (headings, quoted text, etc). To avoid this you have to check these styles and step by step convert them to the new style in InDesign.
Second scenario
[The waste time and it covers 95% of importing procedures]
Situation: You apply the body text (or else) and the text does not change at all.
Procedure (devastating): First, you have to «save» italics (may be bold, also necessary) and tag them with a temporary character style. Immediately you apply your main style (body text) to the whole text.
As text has not changed (sometimes perhaps small and irregular portions) now you are in scenario 2 and the style appears with the sign (+). With keyboard option you reapply the style and obviously all the text is converted. But here happen many accidents: lost of formatting, superscripts, footnotes, etc.
Pros: You finally obtain a reasonable text translation.
Cons: A lot of time has been invested creating unnecessary character styles. Or styles coming from word may disappear when this procedure is applied. But is the only method to really govern (tag) a text that is so reluctant to import in perfect conditions.
Why? Why this situation appears so fickle, almost to the wrong side? Convert .doc to rtf is not a clue, neither. No script until this moment appears to be an elixir. Sometimes you have to implement absurd temporary methods (like changing each font in the coming test from word, like Verdanas, Arials Times, Times New Roman, Calibris +regular+italics+bold, etc).
Conclusion: the daily tragedy repeated day after day.

Answer 1

That would be applicable to
Procedure (devastating): First, you have to «save» italics (may be bold, also necessary) and tag them with a temporary character style. Immediately you apply your main style (body text) to the whole text.
I deal with two kinds of customers: the ones that use our in-house way of marking up text for paragraphs -- local formatting doesn't need special markup --, and people that barge in, throw stuff on my desk, and run away.
For the former, we have a few scripts that first save local formatting (that's "preptext"), and then another that does all paragraph stuff (relying on their own markup), and another that does table stuff. Plug in pictures, book is ready. Less than a day work. Formatting errors are their own.
For the latter, I run the same save-local formatting script, but before that, I check all fonts and most of the local formatting (you'd be surprised how many people use bold italic underlined spaces when there should be, er, just a space; or underline a + to get ±; or superscript an o to get a degree). Then I go through the document, paragraph by paragraph, applying the correct styles one by one. None of that "apply main to all"! You might loose important visual cues that do get imported, such as left-and-right indenting text, which ought to be formatted as a quote. Since I have assigned hotkeys to all major styles years ago, I can go through an entire book in, well, about a day. It then takes some more time to finish the regular stuff, and it also happens that I make formatting errors myself. On the other hand, these clients usually have tons of rewrites, so my errors sort of disappear in the overall calculation.
(preptext can be d/l'ed here: http://www.jongware.com/binaries/preptext.zip -- there have been a few discussions on this forum on proper usage; esp. http://forums.adobe.com/message/2159410 mentions a few important gotchas)