[Cfp-interest] Comments on current documents

[Jim Thomas]

Thanks for the comments, Mike. Please see responses below.

-Jim

On Feb 15, 2013, at 1:44 AM, "Mike Cowlishaw" <mfc at acm.org> wrote:

> Notes from 2013.02.14 telecon, as promised.   These are all
> minor/editorial and refer to the draft documents:
> 
>  cfp1-20121210.pdf
>  cfp2-20121215.pdf
>  cfp3-20130212.pdf
> 
> 
> a)  There are some possible inconsistencies in the form of the
>    Introductions of the three documents:
> 
>      cfp1 mentions all five parts under 'Purpose',
>      cfp2 mentions only parts 1 & 2, under 'The formats', and
>      cfp3 mentions only parts 1 & 2, under 'C Support...'.
> 
> b) in cfp2, page vi line 19, has:
> 
>     "C11 specifies floating-point arithmetic using a two-layer
>     organization. The first layer provides..."
> 
>   but I then got a bit lost as I was expecting the second layer to
>   be mentioned/described, but it never was?

Good points. The introductions need some restructuring and cleanup.

> 
> c) some nits in cfp3 Introduction:
> 
>   Page 1 line 10: ending period missing

This prompted me to look up style suggestions for bullets. I believe it's ok to not punctuate bullets that are phrases or fragments (not sentences).

> 
>   line 17: "Interchange format may or may not be supported as
>     arithmetic." is ill-formed (and I'm not quite sure what the
>     intent is)

Should be "Interchange formats may or may not be supported as arithmetic formats."

> 
>   line 19: "arithmetic formats where computation may be done" sounds
>     a bit awkward; perhaps: "arithmetic formats which can be used
>     for computation"

Agreed.

> 
>   line 38: "C language provides..." -> "The C language provides..."?
>     (Or should it be "The C language specification provides…"?

I think "The C language provides …" would be ok. Do you see a problem with it?

> 
>   line 38: refers to 'three "generic" floating formats" -- should
>     line 29 also refer to the 128 bit format (and use the word
>     'generic')?

Line 38 should be "… three "generic" floating types." The current C standard only mentions a 128 bit floating-point format as a possible format for long double. (I might not understand your point.)

> 
>    Page 2 line 11 'proposed' -> 'specifies'?

Agreed.


> 
> Mike
> 
> _______________________________________________
> Cfp-interest mailing list
> Cfp-interest at oakapple.net
> http://mailman.oakapple.net/mailman/listinfo/cfp-interest