Re: SHF clarifications
Joachim Strömbergson <watchman <at> ludd.ltu.se>
2004-09-14 02:53:36 GMT
Tony Hansen wrote:
> I've been thinking of various dump formats I've used in the past. If I
> had used shf in those applications, I definitely would have had no use
> for the name attributes for either <dump> or <block> in some of them,
> and no use for the blocks or length attributes in some of them. (In
> other words, they would have been useful in some, but not all of those
> applications.)
I can agree that the name attribute could be considered optional. The
reasoning behind the name attribute was to add the ability to tag a dump
with a short and sweet name/description that would aid in the ability to
identify the dump.
Normally, dump files seems to be named short things like "KROM",
"A87DRVC" etc. It would seem that the ability to say "Kernel ROM for AVR
ATmega162L revision 1.0 2004-09-13" or "BIOS update for ASUS 87x boards,
Rev C" would be an aid in clarification - At least a few weeks after you
created the dump or for the user that for some reason don't have the
same ability to grok the smart naming scheme the author of the dump used.
Having the field mandatory might not help the adoption of being
expressive about the contents of a dump file though so I guess it's ok
leaving it as optional.
> I would have had no use for the checksum attribute in any of them. Data
> integrity was not an issue, whereas getting a dump of the data was an
> issue.
The normal usage of the checksum in a dump file is not to verify that
the dump has been transported correctly from A to B, for example via
(Continue reading)
RSS Feed