Le samedi 9 novembre 2002, à 10:26 , Ken Shmidheiser a écrit :

> When creating a large array from raw data, say, a large word list, what 
> are some techniques for detecting elements with errors.

this surely depends on your 'raw' data.

the most common method would be to include an additional byte as
a  checksum here and there, this would then be checked against the
data and an error flagged if they don't match.

if your data is truely 'raw' then you may know other information
about it, for example it is text, then if non-alphanumeric [and
punctuation chars] like cntrl characters or zeros are found... it
is likely that they are errors.

if it is tabular data then is the marker missing? is an entry
too long?

i think that we need more info about the rawness and your expectations
of the data...

:-j