Re: [MORG] WGLC on draft-ietf-morg-inthread-00

[Alexey Melnikov <alexey.melnikov at isode.com>]

Arnt Gulbrandsen wrote:

> More later. Just now:
>
> Alexey Melnikov writes:
>
> > Alexey Melnikov wrote:
> >
> > > >        search-key   =/ "INTHREAD" SP search-key / "MESSAGEID" SP "<"
> > > >                        id-left "@" id-right ">"
> > >
> > > I am not entirely convinced (but I haven't yet checked that '"<" 
> > > id-left "@" id-right ">"' corresponds
> > > to atom IMAP syntax. I think quoted strings/literals would be needed 
> > > to represent some values.
> >
> > I am now positive about that:
>
> Why?
>
> A client isn't permitted to send INTHREAD unless the server says it 
> understands this syntax, so AFAICT nothing constrains the arguments to 
> atoms.

Creating new data types is bad, as it complicates protocol parsers.
If you remember, we had the same argument in Lemonade about URIs and at 
the end decided not to define a new data type.