[dokuwiki] Re: XMLRPC API
- From: "Harry Fuecks" <hfuecks@xxxxxxxxx>
- To: dokuwiki@xxxxxxxxxxxxx
- Date: Sat, 25 Nov 2006 23:05:34 +0100
improved(?) versions. Harry, one is from you [3] the other one is at
[4].
That's also basically dead - I used it on a project a while back but
it's not really maintained any more.
Some quick notes on the original incutio version though - bugs that
may or may not have been fixed - make sure there's no htmlentities in
there (just replacing with HTML special chars), check if it can cope
with large XML documents (something like fread() chops tags at 8k) and
make sure it's really able to handle UTF-8 (Google: PHP XML Rage -
first result).
- Is a XMLRPC api useful at all?
I'd actually say no. In general the magic word is "REST" - HTTP
already has an RPC interface - XML-RPC tunnels through POST. Few
people are writing desktop tools any more for XML-RPC (and SOAP is
dying thankfully - read this
http://wanderingbarque.com/nonintersecting/2006/11/15/the-s-stands-for-simple/
)
- Should we support some popular Blogging interfaces as well?
I'd expect you'd find those interfaces are past their sell by date...
Just gut feeling.
- What can be said about the library choice?
It's about the best choice for PHP + XML-RPC.
- What else should we think of?
You'd probably make people alot happier with a webdav interface. With
Konquerer there's native support. For Windows users there's NetDrive
(http://www.acs.uwosh.edu/novell/netdrive.htm) - HTTP only (no HTTPS).
Basically extend Dokuwiki like a filesystem over the web
For PHP see
http://talks.php.net/show/webdav/
http://pear.php.net/package/HTTP_WebDAV_Server
- Or should we ditch XMLRPC before we even start and look into ATOM?
Ditch it. Ditch it. Would explore how ATOM publishing works though.
- does anyone know of XMLRPC clients for the Wiki API?
That's the key question. And not just that but also is anyone using them?
I'm assuming the basic mission here is to integrate dokuwiki with the
desktop right, as well as other "machines" for computer to computer
exchange vs. human to computer (via browser)?
Returning to webdav, considering Iannz Open Office to Dokuwiki
converter (http://homepages.paradise.net.nz/hillview/OOo/DokuWiki.sxw),
it takes a OO document and creates an new document filled with
dokuwiki markup plus the original text. Now imagine you could just
save that new document and automatically update the remote server -
with webdav you can...
Assuming you've got some kind of webdav client exposing the remote
filesystem of course - aside from Konquerer and NetDrive there's
davfs2 (http://dav.sourceforge.net/) - uses Fuse (i.e. Linux), client
libraries an pretty much any language (including Perl and Python and
lots more. Plus dokuwiki is arranged around the filesystem anyway -
would surely be a good fit / easyish to implement.
...returning the example, even cooler could be a server side filter
than converts open office documents both ways - to and from dokuwiki
syntax - users feel like they're only working in Open Office. In terms
of webdav that "just" means plugging the logic into the server when
reading / writing a document.
--
DokuWiki mailing list - more info at
http://wiki.splitbrain.org/wiki:mailinglist
Other related posts:
