[moby] RE: [MOBY-dev] vote on the RFC discussing Error Reporting

[David González Pisano]

Hello Martin,

I've just sent you the latest version (2.02, 3rd November 2005) version 
which includes some of the changes that were discussed during October. 
The document is also in the bugzilla 
(http://bugzilla.open-bio.org/attachment.cgi?id=246)

My vote is YES, just in case somebody is wondering ;-)

Martin Senger escribió:

>Could anybody sent me the latest document we are voting on? (But do not
>sent me a link to bugzilla please).
>
Some comments about this latest version (the author's name and the date 
the comment was sent to core developers included, plus the request and 
the action taken and reflected in the 2.02 version). No major changes, 
the proposal is still the same than 1 month ago with some rewordings and 
typos corrected:

- Mark (29/09/2005) - Request for change the wording of Change 1 to "In 
the case of an error, failure should raise an exception and an empty 
mobyData block with the appropriate queryID will be returned" - INCLUDED

- Paul (17/10/2005) - Request to rename the proposal to something more 
meaningful - DONE, now the proposal is called "Exception Reporting in 
MOBY-S"

- Paul (17/10/2005) - Doubts about the error 701 - DONE, the ambiguous 
example has been removed

- Paul (17/10/2005) - Errors 70* suggestions - DONE partially, added the 
"error" 700 as a generic service intrinsic exception placeholder for any 
generic situation where the execution was ok and we want to use 
mobyException. Other codes not added, see next comment

- Martin (18/10/2005) - Type "information" in exception reporting - I 
think it not hurts to leave it there. It could be used to report 
informative (non erroneous or warning) exceptions about a specific 
element (Simple or Collection) or mobyData entry in a structured 
(parseable) way, because mobyException allows us to map the information 
to that specific element. We still maintain the Notes in serviceNotes to 
report general unstructured information, and for functional 
compatibility with old moby versions

- Martin (18/10/2005) - Errors 70* suggestions - DONE partially, added 
the "error" 700 (see above) and removed all the codes but 701 
(SERVICE_INTERNAL_ERROR). As Martin suggests, created a 227 
INPUT_INCORRECT_NAMESPACE

- Pieter (3/11/2005) - typos in the document - DONE, the tag is called 
refQueryID, reviewed and changed

- Pieter (3/11/2005) - refArticlename - NOT INCLUDED. I know this can 
open a new can of worms about whether rticlenames should be mandatory or 
not, and we have talked about this in the past. The problem is not with 
the exception reporting proposal, but with the MOBY specification 
itself: if Articlenames are not mandatory, and I can send several 
elements to the same type to my service, how can I differenciate them? 
If there is no need to differenciate them, or that does not make sense 
from the point of view of the service execution, then it makes no sense 
from the point of view of the exception reporting. If I cannot identify 
them individually, then I cannot report exceptions about them 
individually and thus I have to say "There is a problem with one or more 
or the inputs". If I can differenciate them because they have an 
Articlename, then I use the refElement tag to refer to the problematic 
element... Does this make sense?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dgpisano.vcf
Type: text/x-vcard
Size: 338 bytes
Desc: not available
URL: <http://lists.open-bio.org/pipermail/moby-dev/attachments/20051103/44cbafb8/attachment.vcf>