Web lists-archives.com

Re: How to react on a factually wrong Debian wiki change ?


Curt proposed:
>  Wodim has proved to fail under certain conditions and as a general rule
>  should probably be avoided when burning to DVD or BD media.

I like this better than what Brian (i assume) committed now:


  "Wodim has proved much less reliable when used with DVD or BD media
   and is best avoided when burning to those media."

In that statement i object the term "reliable". The failures are
reproducible and due to inappropriate preparation commands before
writing begins or inappropriate finalization commands after writing.

Let me give you an example:

With CD-R, the drive expects to get a SCSI command MODE SELECT with
a "mode page 05" as payload. That mode page is a form with several fields.
Among them is the field "Write Type". We normally use 2="SAO" or 1="TAO".
Then there is the "Multi-session" field. 0=finalize, 3=keep writable.

With DVD-R, which is very similar to CD-R, a mode page 05 is sent too.
The Write Type can be 2="DAO" or 0="Incremental". Only with "Incremental"
the Multi-session field can be used to keep the medium writable.

With DVD+R, no mode page 05 must be sent. The decision whether the medium
stays writable is postponed until writing is done. Each track is finished
by a command CLOSE TRACK SESSION. An additional command CLOSE TRACK SESSION
may close the whole medium.

Now imagine wodim applying the rules for CD-R to the other two.
With DVD-R, chances for success are quite good. A CD-R SAO run with
Multi-session = 0 matches a DVD-R DAO run. But DVD-R Multi-session = 1
demands a different Write Type value than CD SAO or TAO.
With DVD+R, the need for the finalizing commands will not be fulfilled.
Also the drive will complain about the mode page 05 being sent.

This is not "unreliable" it is "clueless".
Insofar Curt's proposal is technically more correct.
But actually i see no improvement over my shorter statement.
(Maybe it's better english, but it's not better message.)

Have a nice day :)