[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: RE: Highly Declarative Designs

• From: John Cowan <cowan@mercury.ccil.org>
• To: "Costello, Roger L." <costello@mitre.org>
• Date: Tue, 8 Feb 2011 10:31:44 -0500

Costello, Roger L. scripsit:

> Consider this question:
> 
>    What is the mime type of the audio files?

Well, of course you can't know until you try.  Perhaps the server is
configured to label .ogg resources as audio/mp3 and .mp3 resources as audio/ogg.
Of course that won't work well unless the .ogg resources really are in
MP3 format etc.

But neglecting that for a moment...

> With the second form the question is readily answered simply by
> navigating to the type attribute. Using XPath this is accomplished
> as follows:
> 
>     audio/source[1]/@type
>     audio/source[2]/@type
> 
> With the first form the question is not so readily answered. It
> requires an application (e.g., browser) to parse "witchitalineman.ogg"
> and infer the mime type from the file name suffix.

I don't see why this isn't declarative.  Regular expressions, in particular,
are just as declarative as XPaths.  The mapping isn't made overt here,
but that's also true of some things like SQL statements that nobody dreams
of denying declarative status to: there's a lot of SQL you can't interpret
correctly until you know at least something about the tables being referred
to, like "select * from table1".

-- 
Don't be so humble.  You're not that great.             John Cowan
        --Golda Meir                                    cowan@ccil.org

• References:
  • Highly Declarative Designs
    • From: "Costello, Roger L." <costello@mitre.org>
  • RE: Highly Declarative Designs
    • From: "Costello, Roger L." <costello@mitre.org>