You don't say, but I'm assuming the PROJCS string in your question is the string being read from your dataset. There are several errors in the string. Here's what I found on SpatialReference.org:
PROJCS["WGS 84 / UTM zone 41N",
GEOGCS["WGS 84",
DATUM["WGS_1984",
SPHEROID["WGS 84",6378137,298.257223563,
AUTHORITY["EPSG","7030"]],
AUTHORITY["EPSG","6326"]],
PRIMEM["Greenwich",0,
AUTHORITY["EPSG","8901"]],
UNIT["degree",0.01745329251994328,
AUTHORITY["EPSG","9122"]],
AUTHORITY["EPSG","4326"]],
UNIT["metre",1,
AUTHORITY["EPSG","9001"]],
PROJECTION["Transverse_Mercator"],
PARAMETER["latitude_of_origin",0],
PARAMETER["central_meridian",63],
PARAMETER["scale_factor",0.9996],
PARAMETER["false_easting",500000],
PARAMETER["false_northing",0],
AUTHORITY["EPSG","32641"],
AXIS["Easting",EAST],
AXIS["Northing",NORTH]]
Your central meridian is wrong, as is the false easting.
Have a look under the "Project Properties" dialog. QGIS will use the project's settings for ellipsoid when calculating distances in a degree based coordinate system.
If you're not using OTF reprojection or have no ellipsoid set, then measurements will usually be cartesian (ie in degrees or degrees squared).
The one exception to this is if you change the units in the measure dialog to something like metres while OTF reprojection is off and you're using a degree based CRS. In this case QGIS will use the very rough approximation of 1 degree = 111300 metres. All this is described in the notes at the bottom of the measure dialog.
I'd suggest enabling OTF reprojection and selecting a suitable ellipsoid in the project properties dialog to get accurate distance measurements.
Best Answer
You should look at what part of the system is calculating the length. If the layer / shapefile is producing the data, then it should be a UTM metres projection. If you are using the measure tool, the canvas CRS needs to be UTM metres projection.
EDIT
Doing some work with 2.01 now, it seems there is a problem with field calculator and $length. There is not the same problem in 2.10
But I can't reproduce it with a blank shapefile....