At 10:19 AM 4/8/98 -0700, Jeff Vance wrote:
>Many thanks for all of the feedback and good suggestions! We have
incorperated
>many of these good ideas into the proposal below.
>Please notice the questions at the end!!
>
>Summary of changes:
>- listf,1 and ,2 will display all asterisks if the data doesn't fit a field
>- 2 new output formats, 10 ("Lsummary") and 20 ("Ldisc")
>- r/b and extent info dropped from ,20 output
>- room for one extra digit for record width
>- no commas in numbers (although I am considering user-defined outputs with
> edits that would support commas)
I find all the changes listed so far are good.
>- k-bytes and M-bytes numbers are base 10 not base 2.
Uh, I think that's a mistake. It will make the numbers come out rounded,
whereas using 1024 bytes as one KB will make the space allocated always be
an integral number of KB.
><<<<<<<<< Existing Output Formats >>>>>>>>>
No problem with what you propose for the old formats.
><<<<<<<<<<< New Output Formats >>>>>>>>>>>>>
>
>:listf @,10
>ACCOUNT= SYS GROUP= OTHER
>
>FILENAME CODE ------------LOGICAL RECORD------------
> SIZE TYP EOF LIMIT
>
>LARGE 80B FA 500 1234567890
>LARGE1 80B FA 12345678902 12345678902
>L2345678* AAAAA 999999W FBxx 999999999999 999999999999
>SMALL 80B FA 327 327
>
>
>:listf @,20
>ACCOUNT= SYS GROUP= OTHER
>
>FILENAME CODE ------------LOGICAL RECORD------------ --SPACE (decimal)--
> SIZE TYP EOF LIMIT K BYTES M BYTES
>
>LARGE 80B FA 500 1234567890 40 0
>LARGE1 80B FA 12345678902 12345678902 987654313 987655
>L2345678* AAAAA 999999W FBxx 999999999999 999999999999 9999999999 99999999
>SMALL 80B FA 327 327 27 0
>
>
>:listfile ./@,10 (or "Lsummary")
> PATH= /SYS/OTHER/
>
> CODE ------------LOGICAL RECORD------------ FILENAME
> SIZE TYP EOF LIMIT
>
> 80B FA 500 1234567890 LARGE
> 80B FA 12345678902 12345678902 LARGE1
> AAAAA 999999W FBxx 999999999999 999999999999 *L2345678
> 80B FA 327 327 SMALL
>
>
>:listfile ./@,20 (or "Ldisc")
> PATH= /SYS/OTHER/
>
> CODE ------------LOGICAL RECORD------------ --SPACE (decimal)-- FILENAME
> SIZE TYP EOF LIMIT K BYTES M BYTES
>
> 80B FA 500 1234567890 40 0 LARGE
> 80B FA 12345678902 12345678902 987654313 987655 LARGE1
> AAAAA 999999W FBxx 999999999999 999999999999 9999999999 9999999 *L2345678
> 80B FA 327 327 27 0 SMALL
Why do we need mode 10? I have never understood the point behind having
mode 1 when mode 2 does everything mode 1 does with no discernable side
effects. What's the point? (The only possible reason I can think of is
that it takes a nontrivial amount of work to add up the extent sizes and so
when the disc space isn't needed, modes 1 and 10 may be more efficient than
modes 2 and 20, respectively. However, I don't know whether this is the
case or not.)
>Questions:
>----------
>1) how should megabytes be rounded?
> - mathematical rounding, ie. <= 4 round down, >=5 round up
> - how should MB be handled for files less than 999 KB?
> o apply same mathematical rounding?
> o always report zero MB?
> o always report 1 MB?
Personally, I don't see why we need to show MB at all. Especially if MB is
defined as 10^6 bytes and KB is defined as 10^3 bytes, the numbers will be
the same (ignoring rounding up) other than dropping the last 3 digits of
the KB. I think we should show only KB for the space, where KB is defined
as 2^10 bytes. This will allow us to show exact disc space used without
any rounding, whereas using 10^6 or 10^3 bytes will always be rounded from
some multiple of the 4096 byte page size.
If there's some reason why we *really* need a 10^6 type MB then make it a
field which may have (or always has?) a floating decimal place, because it
will always be an approximation. Small files will simply show more digits
to the right of the decimal point than large files. Personally I think
this field isn't needed at all.
>2) is it acceptable to exclude R/B and extent info from the new ,20 format?
Absolutely. :)
--
Jeff Woods
[log in to unmask] at Tivoli Systems
[log in to unmask] at home [PGP key available here via finger]
Haiku by Rahul Sonnad:
There is a chasm
of carbon and silicon
the software can't bridge.
|