Yes, there media issues, but the problems are egg-saturated due to the
issues with the implementations.
--
http://www.standards.com/; See Howard Kaikow's web site.
"Mike Richter" wrote in message
...
Howard Kaikow wrote:
To be more accurate:
There is nothing wrong with packet writing itself, rather the problem
has
been the poorly designed/implemented, and less than adequately tested,
implementations of ISO/IEC 13346.
Some of the problems are NOT necessarily the fault of the implementers.
ISO/IEC 13346 was published with numerous errors.
I've not yet seen an effort to correct the ISO/IEC standard, which would
go
a long way towards helping the implementers.
The problem in fact lies with the error rate of CD media in general and
the specific problems of erasable media. The impact comes in the TOC,
which despite the best efforts of the implementers must at least in part
be scrubbed - erased and rewritten - whenever a file is written or erased.
Because of the limited number of erase cycles the medium will tolerate,
fixed-length packets are implemented by holding the TOC in RAM. (Yes,
matters would be better if it were held on the hard drive instead, but
the advantage would not be as great as one might hope.) Since it is in
RAM, it's subject to scrambling with a power transient or OS crash.
The same limitation on erasures means that the disc is fully fragmented
by design - the sectors of even the first file written are scattered
across the disc so that - except for that critical area of the TOC - no
region will be scrubbed if a file is updated frequently. As a result,
speed is dictated largely by seek and high-speed drives do not deliver
the read/write as expected.
All of that applies to fixed-length packets only. On a write-once disc
with variable-length packets, there is some fragmentation (that's why
Level 2 can't be used to close the session). That costs a bit in speed,
but since there's no scrubbing and none of the spontaneous decay
characteristic of high-speed erasable media.
Mike
--
http://www.mrichter.com/