Re: length-spec vs. length for characteristics

[James Clark <jjc@xxxxxxxxxx>]

At 22:19 16/04/97 -0400, Tony Graham wrote:

>Does this mean that only for characteristics with a length-spec value,
>I can use display-size (also introduced in 12.5.2) to return a
>length-spec that I can use when setting the value of the
>characteristic?

Yes.

>  What other ways are there for arriving at a
>length-spec instead of a plain old length?

table-unit and by using + - * / on length-specs.

>For that matter, what specifies the "value flow object" for
>display-size?

It's the flow object that the characteristic applies to.  See the
definitions of the inherited-C and actual-C procedures.

>I was started on this tack by looking at the following characteristics
>of the box flow object class and their values:
>
>	box-corner-radius	length-spec
>	box-size-before		length
>	box-size-after		length
>	start-indent		length-spec
>	end-indent		length-spec
>
>For example, why is box-corner-radius a length-spec and
>box-size-before a length?

box-size-before/after only applies to inline flow objects, so there's no
real reason why you would need it to be a length-spec.  On the other hand
with a displayed box, it's conceivable you might want the radius of the
corners to be proportional to the display-size.  Sometimes it's a bit random
which characteristics are lengths and which are length-specs.

James