Perhaps you had an implementation in mind when you wrote the spec. Eas=
e of
implementation depends on your internal representation for attributes, =
I
suppose, and our representation has a legacy of more than a year of IPP=
evolution behind it. We can change it, but there's a large impact since=
attributes are so central to the implementation.
> Only field c) is redundant
I wasn't addressing redundancy in the data stream, although that's a go=
od
point, too. What I meant is that it's redundant to specify a unique co=
mpound
attribute syntax for these two specific attributes (nameWithLanguage an=
d
textWithLanguage]) when we already have a general 'compound-attribute' =
syntax.
-Carl
('Course none of this will matter when we switch to XML.)
Robert.Herriot@Eng.Sun.COM on 01/12/98 09:55:35 AM
Please respond to Robert.Herriot@Eng.Sun.COM @ internet
To: Carl Kugler/Boulder/IBM@ibmus, ipp@pwg.org @ internet
cc:
Subject: Re: IPP Mail Archive: Re: IPP>MOD add another issue [encodin
I found it rather easy to implement when I did it.
Only field c) is redundant, but I found the implementation simpler
when each value was preceded by a length because I already had code
that picked up a length n and then the value as the next n bytes.
It is then easy to check later that the sum of these bytes equals the
value length of the whole compound value.
Bob Herriot
=