- From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
- Date: Wed, 4 Sep 2002 09:30:07 -0700
- To: <noah_mendelsohn@us.ibm.com>
- Cc: "Nilo Mitra" <EUSNILM@am1.ericsson.se>, "Marc Hadley" <marc.hadley@sun.com>, "Martin Gudgin" <mgudgin@microsoft.com>, "Jean-Jacques Moreau" <moreau@crf.canon.fr>, <www-archive@w3.org>
I think that's fine--I would remove the "rigorously" though unless we feel we are particularly sloppy elsewhere and want to call it out explicitly here that we are not ;) Henrik Frystyk Nielsen mailto:henrikn@microsoft.com >-----Original Message----- >From: noah_mendelsohn@us.ibm.com [mailto:noah_mendelsohn@us.ibm.com] >Sent: Tuesday, September 03, 2002 17:54 >To: Henrik Frystyk Nielsen >Cc: Nilo Mitra; Marc Hadley; Martin Gudgin; Jean-Jacques >Moreau; www-archive@w3.org >Subject: Re: Added resolution to issue 320 > > >I think this is close, but have a few quibbles both in terms >of editorial >style and content. > ><latest> >SOAP fault codes are intended for use by software to provide an >algorithmic mechanism for identifying the fault. SOAP fault codes are >organized as a linked list of XML qualified names allowing a >SOAP node to >identify the fault category at an increasing level of detail >of the SOAP >fault. > >...2 paras follow... > ></latest> > ><proposed> >SOAP fault codes are XML qualified names, and are intended to >provide a >means by which faults are rigorously classified. A >hierarchical list of >SOAP codes and associated supporting information is included >in every SOAP >fault message, with each such code identifying the fault >category at an >increasing level of detail. > >(..remaining 2 paras unchanged...) > ></proposed> > >Reasons for above suggestions: > >* I don't think that what's going on is really 'algorithmic', >and it's not >clear to me that it's only software that gets to do the identifying. > >* I'm not sure it's better, but I used the word "classified" >rather than >"identified" in the first para. I think "identified" could be >taken in >the sense of identifying one soap fault message vs. another, >and that's >not what we mean here. The codes exist even before they are >used, and >the same code is applied to many separate instances of faults (two >separate messages, each using in illegal encoding.) So, I went with >classified. Note that where I retained "identifying" it clearly says >identifying a category, which I think is correct. > >* I don't think the lists are linked, in the traditional data >structures >sense. I'm used to seeing the term "linked" list applied to >structures >connected by pointers, as distinct from array-based, etc. lists. > >* I'm not 100% set on the word "rigorously", but I think it's OK, and >closer to the mark than algorithmic. > >What do you all think? Worth changing? Further refinements? > >------------------------------------------------------------------ >Noah Mendelsohn Voice: 1-617-693-4036 >IBM Corporation Fax: 1-617-693-8676 >One Rogers Street >Cambridge, MA 02142 >------------------------------------------------------------------ > > > >
Received on Wednesday, 4 September 2002 12:30:42 UTC