[Peerpress-main] XML schemas.

[Joakim Ziegler]

On Mon, Jan 08, 2001 at 06:14:22AM +0000, nloth at senet.com.au wrote:
> Nice work, Joakim. My comments are inline.

> > <sitelist>

> >   <site id="kuro5hin.org" system="scoop" version="pp0.1">

> Where do we get the id from? Is it the domain name? What about systems hosted 
> in a directory off a domain, or multiple systems on the same domain? Maybe the 
> id should be the baseuri, in which case it might be redundant?

Well, since Peer Press will be operating with the concept of a "member site"
(important for licensing stuff), then this should be a string assigned when
the site becomes a member. Using the domain name sounds like a good idea to
me, although perhaps not always. The main idea is for it to be unique.


>>     <name>Kuro5hin</name>

> Can people choose anything, or does this need to be registered somehow?

Yes, this is free-form. Basically, it's the name of the site, for use in
stuff like summary screens for choosing sites, attribution lines, lists of
Peer Press member sites, etc. It's the human-readable site name.


>>     <languages>
>>     EN_us
>>     </languages>
>> 

> Is this comma separated? Why not a list of languages:
> <languages>
>   <language>EN_us</language>
>   <language>Klingon</language>
> </languages>

White space separated, was my basic idea. I was just trying to keep the
verbosity down a little, since the sitelist can get long.


>>     <licensetypes>pp-exclusive, fee-based</licensetypes>
 
> Again, why not a list? It makes it easier to parse and use XPath filters on.

Same here.


>>     <techcontact>tech at kuro5hin.org</frontpageuri>
>>     <editorialcontact>rusty at kuro5hin.org</frontpageuri>
>>     <licensingcontact>licensing at kuro5hin.org</licensingcontact>

> How about:
> <contacts>
>   <contact type="tech">tech at kuro5hin.org</contact>
>   <contact type="editorial">rusty at kuro5hin.org</contact>
>   <contact type="licencing">licensing at kuro5hin.org</contact>
> </contacts>
 
> Isn't that potentially less trouble to extend? Also, does this contact have to 
> be an email address?

It could be. It's more work to get a validating XML parser to check that all
required ones are included, though, for instance. As for being an email
address, what else do you envision that it could be?


-- 
Joakim Ziegler - Helix Code web monkey - joakim at helixcode.com - Radagast at IRC
      FIX sysop - free software coder - FIDEL & Conglomerate developer
            http://www.avmaria.com/ - http://www.helixcode.com/