Open Stack

On Tue, 2012-07-31 at 11:21 -0700, Vishvananda Ishaya wrote:
> I've been thinking about a simple way to solve this issue, and I've
> come up with a spec[2] that i think is workable. I'd like to get some
> decisions on the following at the meeting:
> 
> a) is this approach acceptable?
>  * does it need to be modified slightly?
>  * is renaming the badly named extensions ok?
>  * is there a more general approach that extensions could use?
> b) is there an alternative to this that is easier/better?
>  * alternative: bump minor version of the api?
> c) who wants to help on produce the required code in the next week?
> d) who wants to help update the documentation?
> e) who wants to investigate xml?

I unfortunately won't be able to make it to the meeting, however, I
wanted to contribute a few comments on this.  The primary comment I have
is that it should be possible, with a few tweaks, to extend the XML
templates stuff such that it can decode from XML, as well as encode to
XML; this would make it possible to create an extensible XML decoder for
something like servers.  This would likely go a long way to addressing
the problem you've brought up.

The second comment I have would be far more oriented to OS-APIv3: Let's
plan on just removing the XML support altogether.  My argument for this
is simple: if you examine the extensions that have been contributed, you
will discover that several of the more recently contributed ones don't
have XML templates, which means that the developers that have
contributed them probably aren't concerned with supporting XML, and that
the reviewers who approved them aren't either; under those conditions,
our XML support is going to suffer bitrot over time, and I feel we
should try to nip the problem in the bud for v3.
-- 
Kevin L. Mitchell <kevin.mitchell at rackspace.com>