Hi All, I am talking about unix col(1) command used for some reverse line filtering etc. And I notice that the stdout of this command is line buffered i.e. the stdout will flush the data in its buffer line by line. So the number of writes performed by stdout are more. So now if I make stdout as fully buffered it will reduce the write hits. :) But if I do these changes I think there will some problems -> when slow terminals use col(1) as command , then the turn around time will get reduce for it self. eg: cat file | col -b | lp Here lp will receive data in terms of buffer size but before it was receving in terms of each lines. So can I change change col(1) command's stdout to fully buffered? So do any of you know some more problems related with this change , and can I see some link about col(1) command design stating that it should be either line buffered or fully buffered etc. Waiting for positive replay, Sunil kumar. -- Thanks & Regards, Sunil Kumar S

Linux: col [-bfpx] [-l num] -lnum Buffer at least num lines in memory. By default, 128 lines are buffered. Luke Chi

Linux is an open source. Download the source code and read "col" command. Luke Chi