Breaking down lines?

[Jacob Baltuch]

Let's say you have two lines written in Nagari as:

 pazupatirapitaanyahaanikRcchraadagamayadadrisutaasamaagamootkah
 kamaparamavazaMnaviprakuryurvibhumapitaMyadamiispRzantibhaavaah

In translitteration or in Nagari in certain textbooks for beginners
(abusing the viraama in the process) they would be written as:

 pazupatir api taany ahaani kRcchraad agamayad adrisutaasamaagamootkah
 kam aparam avazam na viprakuryur vibhum api tam yad amii spRzanti bhaavaah

(Btw I find the use of the viraama this way, i.e. writing <r.h a> for <ra>,
to be really an eyesore in that it distorts the general character of this
writing system). In modern editions it would be written as:

 pazupatirapi taanyahaani kRcchraadagamayadadrisutaasamaagamootkah
 kamaparamavazam na viprakuryur vibhumapi tam yadamii spRzanti bhaavaah

that is breaking the line where Nagari allows one to do so (which seems
to be a bit of a half-assed solution). Now why didn't modern editors come up
with something like this instead:

 pazupatira \pi taanya \haani kRcchraada \gamayada \drisutaasamaagamootkah
 kama \parama \vazam na viprakuryur vibhuma \pi tam yada \mii spRzanti bhaavaah

(where \ would be a sign analogous -- but not identical -- to the avagraha,
which would indicate that the last vowel of the previous word in fact belongs
to the following word) To me such a system seems superior to the method of
"splitting where you can" while having all the advantages of splitting the text
completely into words without misusing the viraama. (But it would require the
introduction of a new sign.) But maybe such a system has got problems I don't
see? What would be in your opinion the problems with such a system? Have there
been other creative solutions to resolve the dilemma of splitting a Nagari text
into words (in those cases one would like to) while preserving the general
character of the writing?