OK, got it--thanks for the clarification. You know, I don't know how practical this is, but it would make the world a better place if Google could open-source your heuristic parser once it's finished. This is one of those problems that has been around forever, keeps getting solved badly, and could really use a canonical open-source solution. (If you do it, I'll translate it into PHP! :)
Thanks, Charlie On Aug 5, 11:11 am, Gyorgy Abonyi <[email protected]> wrote: > Hi Charlie, > > Yes we currently create the formatted version if the client writes only the > structured flavor, but not in the other way. If a client writes both flavors > we store both flavors as is, without triggering any conversion. So in terms > of conversion from structured flavor to formatted one, there will be no > change compared to the current behavior. > > And yes I ought to write "right enough" heuristic parsing, it's indeed not > trivial. > > regards, > Gyorgy > > > > On Wed, Aug 5, 2009 at 5:47 PM, Charlie Wood <[email protected]> wrote: > > > Gyorgy- > > > Thanks for the detailed information. Is point #3 needed though? > > > > 3. On modification explicitly clear the flavor the application doesn't > > > need. In this case mutation will trigger heuristics to generate the > > missing > > > flavor. > > > The current behavior seems to be that if I insert new structured data, > > Google will create the formatted version for me without my needing to > > clear the formatted field. > > > Also, I'm skeptical that your parsing algorithm can ever be "right". > > Names are ambiguous. If a contact has the formatted name "Kramer", is > > a first or last name? It's impossible to know. The only "right" > > solution IMHO would be to expose only structured fields in the Gmail > > Contact Manager. I hope that happens! > > > Thanks again, > > Charlie > > Spanning Sync > > > On Aug 5, 9:53 am, Gyorgy Abonyi <[email protected]> wrote: > > > In version 3 we introduced structured names and addresses along with the > > > formatted versions, however at the moment we do not provide conversion > > from > > > formatted to structured at our side. Moreover the behavior of Contact > > > Manager in GMail is suboptimal, only the formatted version can be edited, > > > and upon edit of the name or address field the structure is destroyed. We > > > are working hard to improve the current behavior. > > > > We are intend to introduce heuristic name and address parsing in the API, > > > and to help developers to adjust to incoming improvements we recommend to > > > follow a few simple rules while developing structured and/or formatted > > data > > > aware clients: > > > > 1. Read the flavor (structured / formatted) the client is interested > > > about. > > > 2. If the data type (structured / formatted) your client is interested > > in > > > is not available, then format or parse the other available data type > > for the > > > application needs. > > > 3. On modification explicitly clear the flavor the application doesn't > > > need. In this case mutation will trigger heuristics to generate the > > missing > > > flavor. > > > 4. If the application provides both structured and formatted parts no > > > heuristics will be applied and data will be stored as is. > > > > Please note that point 2 is a temporary solution until not all flavors > > are > > > available in the database. As we intend to back-fill all existing > > contacts > > > with the missing flavors using our heuristic parsers, this step will > > become > > > obsolete in the future. Those contacts that already contain both flavors > > > will remain unchanged. > > > > Our heuristic parser will be used by Contact Manager in GMail too, so if > > the > > > user edits the formatted name or address in the contact manager, we will > > > generate the correspondent structured components. > > > > And to answer your obvious question, when will it happen? As soon as we > > are > > > convinced that our parsing algorithm is right. Please expect that the > > > heuristic name parsing probably will come earlier than address parsing. > > > > Hope this answers lots of your questions and valuable comments about > > > formatted vs. structured names and addresses posted to this group. > > > > Sincerely yours, > > > Gyorgy Abonyi, Google Contacts team --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Google Contacts API" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/google-contacts-api?hl=en -~----------~----~----~----~------~----~------~--~---
