Current Desktop capabilities - take 3.
Dave Cridland [Home]
dave at cridland.net
Fri Jun 4 22:50:37 EEST 2004
On Fri Jun 4 15:47:25 2004, Michael Meeks wrote:
> Hi Dave,
> On Thu, 2004-06-03 at 20:05 +0100, Dave Cridland [Home] wrote:
> > You're right that it needs clarification, but ASCII is always
> Of course :-) I'm such a dofus to ignore the '-' character I
> being banned by the spec.
Not really. It's just very much easier to make those mistakes when
you're writing prose rather than ABNF. And it's always easier to spot
them if you didn't write it.
> > Michael, would you like me to write up some ABNF to describe the
> > syntax fully?
> Well - thanks for sending me that; I guess I'm just sceptical that
> makes life any easier for an implementer / reader; we can add it as
> appendix though I guess.
Well... when I'm reading RFCs, I read the text to get the general
idea, the semantics, and a feel for the syntax. I read the formal
syntax - the ABNF - when I write the parser or whatever. It's usually
kept in a section by itself. Feel free to mark it as informative
rather than normative, or just ditch it if you don't want it - as it
stands it's so short it doesn't matter either way.
If you want to namespace the identifiers, add options to the end,
etc, then it becomes more useful - when people ask "Can I do X with
the syntax?", then the answer's in the formal syntax.
For the others, I'll write the basic syntax out at the end of the
mail again. I'll probably write it better this time, I didn't bother
keeping a copy.
> However - before complicating it beyond belief - it'd be nice to
> more input from the ravingly-anti brigade to see if they're still
> so, or
> if this is more acceptable.
[FWIW, I didn't find it offensive that Michael implied my suggestions
were 'complicating it beyond belief', but thanks to Jeff who was
apparently sticking up for me, hence the CC to me, I assume.]
Well, I think it'll never be exactly 'acceptable' to those who want
to be on the purist end of the scale.
It's now couched in terms that make it more politically acceptable,
though I can't actually see a likely circumstance where it might
read, semantically, any differently from a list of preferred
desktops, but anyway.
Either way, it's considerably better than nothing, which is our
current state of play.
Proposed formal syntax in ABNF:
desktop-value = hint *( ":" hint )
lowercase = %x61-7A ; ASCII characters "a" to "z"
hint = lowercase *( "-" / lowercase / DIGIT ) ; DIGIT defined in
RFC2234, section 6.1.
More information about the xdg