Talk:Append a record to the end of a text file: Difference between revisions

Content added Content deleted
(→‎Changed: change the task?)
Line 94: Line 94:


: I didn't like the task from the very begining because of the lack of focus. If it's about maintaining a database, then it should include all basic operations: adding record, removing record, replacing existing record, and record lookup. If it's about concurrent writing to a file, the passwd-like record requirement is unnecessary complication (colon-separated strings are pretty uninteresting anyway). I think it's worthwhile to have a task that demonstrates concurrent file access, and the simpliest thing is ask for implementations of whole file or file-region locking with fixed length, random-access records. --[[User:Ledrug|Ledrug]] 00:11, 27 September 2011 (UTC)
: I didn't like the task from the very begining because of the lack of focus. If it's about maintaining a database, then it should include all basic operations: adding record, removing record, replacing existing record, and record lookup. If it's about concurrent writing to a file, the passwd-like record requirement is unnecessary complication (colon-separated strings are pretty uninteresting anyway). I think it's worthwhile to have a task that demonstrates concurrent file access, and the simpliest thing is ask for implementations of whole file or file-region locking with fixed length, random-access records. --[[User:Ledrug|Ledrug]] 00:11, 27 September 2011 (UTC)

My mistake... It seems that appending a structure/record to the end of a text file in a concurrent environment is both unusually difficult and unusual, especially if the file is a simple (but mythical) CSV "passwd" file. I speculate that this is because it is not really a "language" function, more an OS and FS function. Certainly your feedback has been interesting. So, unless someone else objects, I'm am not strongly against having this task deleted. How do we do this?

[[User:NevilleDNZ|NevilleDNZ]] 01:05, 27 September 2011 (UTC)