st: censored grouped data regression

[itismejp@hushmail.com]

Hello Stata-community,

i`am currently working on a term paper and hope i can find some help here. 
Unfortunately my dependent variable (y) is limited in the way that only the intervals in which the unobservable latent variable falls are observable. Furthermore y is censored from below and above which means the first interval is -infinity to 0 and the last interval is 50 to +infinity.
I'am aware of the STATA command intreg. Unfortunately i want to go a little further. 

1. Is there any comparable command to the qreg-command i could use to get quantile regression results in case of such a dependent variable?
2. To use intreg i would have to make the assumption of normally distributed errors. If i'am not willing to make that assumption i could implement (for example) an error structure described by a mixture of normal distributions. Can anyone with experience in ML programming in STATA tell me how hard it would be to program such an Estimator or does anyone know an implementation in STATA or R?
3. Is there an already implemented way of working with a semiparametric approach in case of such a dependent variable?

Help would be much appreciated!

With best regards
Georg

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/