accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Christopher Tubbs (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-1933) Make unit on memory parameters case-insensitive
Date Thu, 02 Jan 2014 17:11:52 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-1933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13860344#comment-13860344
] 

Christopher Tubbs commented on ACCUMULO-1933:
---------------------------------------------

I'm actually somewhat opposed to this ticket. The SI-prefixes (and "B" for bytes) I used for
this are case-sensitive for a reason. "B", for instance means "bytes", vs. "b", which means
"bits". "M" means "mega", vs. "m" which means "milli". "P" for "peta", vs. "p" for "pico".

Now, in the context of memory, I realize it doesn't make sense to have a "millibyte" or a
"picobyte" ("B" for byte rather than "b" still applies, though), but I'd strongly prefer to
stick to the SI case-sensitive precedent.

That said, the proper prefix for "kilo" is "k" (lower-case), not "K". However, there is no
SI unit for upper-case "K" in conflict, and in computer networking, it is not unusual to see
both "kbps" and "Kbps" for "kilobits per second". We could make the "k" case-insensitive.

Coming from a physics background, I think it's quite important to recognize that case matters
in units, and I'd rather encourage people to use SI-prefixes when possible, for maximum clarity,
than to introduce ambiguity that requires context (such as asking the question, "does milli-
make sense here?") to resolve.

> Make unit on memory parameters case-insensitive
> -----------------------------------------------
>
>                 Key: ACCUMULO-1933
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1933
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: shell
>    Affects Versions: 1.5.0
>            Reporter: Josh Elser
>            Priority: Minor
>              Labels: newbie
>
> {{config -t my_table -s table.split.threshold=2g}} fails with a non-intuitive error.
> {{config -t my_table -s table.split.threshold=2G}} succeeds
> It would be nice for both to work.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Mime
View raw message