[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [search-ws] recordPacking - A Proposal
Thank you Tony for this clear explanation of the problem, I now understand it. I certainly have to admit that I didn't understand it until now. I am in favor of your proposed solution A, EXCEPT we cannot, ABSOLUTELY CANNOT, call the new parameter recordPacking. That is, -we could rename the old recordPacking parameter "recordXMLEscaping". We would take some heat for that, but I think it is defensible. - we could introduce a new parameter for packed and unpacked. - but we CANNOT call that new parameter recordPacking We would simply need to find an entirely new name for it. How about dataPacking? Ralph, I am not offended by the seemingly boolean nature of packed and unpacked as I don't see that it implies only one way to pack. I think the semantics of this are as stated earlier in the discussion: packed means strict adherence to the schema; unpacked means that the data can float around at will. I would still prefer "strict' and 'loose' to 'packed' and 'unpacked' but I am not going to hold out for that. --Ray ----- Original Message ----- From: "Hammond, Tony" <t.hammond@nature.com> To: "OASIS SWS TC" <search-ws@lists.oasis-open.org> Sent: Friday, April 16, 2010 7:31 AM Subject: [search-ws] recordPacking - A Proposal > Hi: > > (Sorry for the length but needs to be said.) > >> - Finally, the orthogonality seems to me to make this parameter/element >> sematics overloaded and akward. > > Well, I think we can all agree with that. But I think we need to > understand > better the reasons why this has arisen. And then figure out what we're > going > to do about it. > > It seems to me that there *are* two distinct usages under the microscope: > > 1. transport > > 2. data access > > The first usage ("transport") is what the original SRU 1.* use of > recordPacking sought to address. In effect it offered a means to escape > the > XML data for transport purposes so that an XML data record could be snuck > by > an XML processor. (Not clear why a CDATA section couldn't have been used > for > this purpose.) Note that this is *not* data packing as is normally > understood. There is no actual packing involved - just XML escaping. > > The second usage ("data access") is more in keeping with the named > operation: record packing. This relates to packing of record data into > (and > out of) a schema - in this case the SRU/XML schema. This is a direct > consequence of SRU 2.0 taking on board alternate media types which are > less > constrained than the canonical SRU/XML response. While we can define a > reasonable reference mappings of SRU/XML to these media types we have a > real > world requirement to make the data more readily available to consumer > apps. > (And if we allow the media type representations to vary too much from > standard SRU/XML we introduce destabilization of the protocol.) > > So this would seem to suggest an impasse. But we need to make progress. > > So let's do that. > > 1. SRU 2.0 != SRU 1.* > > 2. A "recordPacking" parameter suggests a value set of "packed" and > "unpacked" (or alterantely boolean equivalents) > > 3. The "data access" issue is a data packing issue > > 4. The "transport" issue is not a data packing issue > > 5. SRU extension formats (RSS, etc) should align with the native format > (SRU) - or there should be reference registered versions that do align > > 6. SRU extension formats are an integral part of the SRU 2.0 deal > > On that basis I volunteer two alternate ways forward: > > == > A. Rename SRU 1.* "recordPacking" parameter to SRU 2.0 > "recordXMLEscaping" with "xml", "string" values, and introduce new SRU 2.0 > "recordPacking" parameter with "packed", "unpacked" values > > e.g. > > recordPacking = packed > recordXMLEscaping = string > > Bottom line: We break the orthognality. > > == > B. Retain SRU 1.* "recordPacking" parameter but overload it so that it > provides hints to the server as to what actions to take viz-a-vis > "transport" issues and "data access" issues. We would not overspecify (way > too complex) but would leave server free to act. > > e.g. > > recordPacking - packed, string > recordPacking - unpacked, xml > > Bottom line: We find a compromise solution. > > == > > Notes: > > * In my mind there is no contest but that SRU 2.0 should go with > Solution A: introduce new "recordXMLEscaping" parameter to handle the > restricted case of tunnelling XML record data. > > [[ And honestly, if this escaping was established for sole purpose of > passing back bad XML then the world will be a better place for not > promoting > that activity! ]] > > * recordXMLEscaping could be generalized to recordEscaping but that > could affect the value set - e.g. do we add "json", etc in? Suggest we > keep > this parameter purely for the SRU/XML case. > > * This is SRU 2.0 - a MAJOR revision. Taking on board alternate media > types comes at a price. > > * If we wimped out and went with Solution B there would still be > confusion as to what would happen with non-XML formats (e.g. JSON). Or > would > we also throw a "json" value into the mix. And then what about other media > types? (Note that nature.com is next week deploying a "text/turtle" media > type. Would we also allow a "turtle" value?) > > * And if, horror of horrors, we go with the status quo then we are > propagating a badly named parameter for a limited use case which denies a > more general usability (and/or conformability) of open media types and > undermines credibility of the whole SRU effort. (Or something like that. > ;) > > Done wrong we risk ending up with another niche digital library protocol. > Done right we still have same background but we are at least facing > forward > with respect to the Web and living up to the promise of SRU as a true ZING > thing. > > Solution A is a bold step but IMHO is the correct way ahead. > > Cheers, > > Tony > > > > > ******************************************************************************** > DISCLAIMER: This e-mail is confidential and should not be used by anyone > who is > not the original intended recipient. If you have received this e-mail in > error > please inform the sender and delete it from your mailbox or any other > storage > mechanism. Neither Macmillan Publishers Limited nor any of its agents > accept > liability for any statements made which are clearly the sender's own and > not > expressly made on behalf of Macmillan Publishers Limited or one of its > agents. > Please note that neither Macmillan Publishers Limited nor any of its > agents > accept any responsibility for viruses that may be contained in this e-mail > or > its attachments and it is your responsibility to scan the e-mail and > attachments (if any). No contracts may be concluded on behalf of Macmillan > Publishers Limited or its agents by means of e-mail communication. > Macmillan > Publishers Limited Registered in England and Wales with registered number > 785998 > Registered Office Brunel Road, Houndmills, Basingstoke RG21 6XS > ******************************************************************************** > > > --------------------------------------------------------------------- > To unsubscribe from this mail list, you must leave the OASIS TC that > generates this mail. Follow this link to all your TCs in OASIS at: > https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]