Map's currently have a "Contact" metadata field which is just a short text field. It occurs to me that this will soon be obsolete as map's have owners with their own contact fields. I'm working on the map metadata workflow lately. Any advice on where to look for info about to hook this up appropriately? We don't have a UI spec for this, but I would imagine the following would make sense (maybe Rollie can comment on them): - We remove the Contact metadata field - When the map is saved, the metadata form does _not_ have a Contact field available. - Maps have a primary author (or point of contact?) which is tied to the user model - The Map Info page shows the user associated with the map (and the contact information associated with that user, in a later iteration) The difficulty with scoping this feature is that it involves tying together some of the very oldest (Map Composition) and very newest (User management and permissions) features. I'm worried about how we can have GeoNode 1.0 adequately bridge between these two aspects, even though it's not been supported by a particular contract and so hasn't been specified thoroughly. Any ideas on how to move forward? My own inclination is to try to put *something* together, and revisit it later once we have some user feedback. -- Sebastian Benthall OpenGeo - http://opengeo.org
On 07/12/2010 12:50 PM, Sebastian Benthall wrote: > Map's currently have a "Contact" metadata field which is just a short > text field. > > It occurs to me that this will soon be obsolete as map's have owners > with their own contact fields. > > I'm working on the map metadata workflow lately. Any advice on where > to look for info about to hook this up appropriately? We don't have a > UI spec for this, but I would imagine the following would make sense > (maybe Rollie can comment on them): > > - We remove the Contact metadata field > - When the map is saved, the metadata form does _not_ have a Contact > field available. > - Maps have a primary author (or point of contact?) which is tied to > the user model > - The Map Info page shows the user associated with the map (and the > contact information associated with that user, in a later iteration) > > The difficulty with scoping this feature is that it involves tying > together some of the very oldest (Map Composition) and very newest > (User management and permissions) features. I'm worried about how we > can have GeoNode 1.0 adequately bridge between these two aspects, even > though it's not been supported by a particular contract and so hasn't > been specified thoroughly. > > Any ideas on how to move forward? > > My own inclination is to try to put *something* together, and revisit > it later once we have some user feedback. > > -- > Sebastian Benthall > OpenGeo - http://opengeo.org It seems there is a front-end and a back-end side to this. Things are a little easier than with layer metadata because we don't have to worry about contacts independent of users (you have to be a user to make a map in the GeoNode). On the front-end: Just drop the field from the form and don't explicitly send any user info to the Django app when saving a map. Django can tell us who's logged anyway. On the back-end: Change the 'contact' field on Maps to a ForeignKey on users. This will be a backwards-incompatible change, so we should work out a migration process for the live site. Probably, we should just dump the data, reset the database, and hand-modify the data dump to use a user id instead of a string for the contact field. Does this answer the question, or am I missing something? -- David Winslow OpenGeo - http://opengeo.org/
On 07/12/2010 01:13 PM, David Winslow wrote: > On 07/12/2010 12:50 PM, Sebastian Benthall wrote: > >> Map's currently have a "Contact" metadata field which is just a short >> text field. >> >> It occurs to me that this will soon be obsolete as map's have owners >> with their own contact fields. >> >> I'm working on the map metadata workflow lately. Any advice on where >> to look for info about to hook this up appropriately? We don't have a >> UI spec for this, but I would imagine the following would make sense >> (maybe Rollie can comment on them): >> >> - We remove the Contact metadata field >> - When the map is saved, the metadata form does _not_ have a Contact >> field available. >> - Maps have a primary author (or point of contact?) which is tied to >> the user model >> - The Map Info page shows the user associated with the map (and the >> contact information associated with that user, in a later iteration) >> >> The difficulty with scoping this feature is that it involves tying >> together some of the very oldest (Map Composition) and very newest >> (User management and permissions) features. I'm worried about how we >> can have GeoNode 1.0 adequately bridge between these two aspects, even >> though it's not been supported by a particular contract and so hasn't >> been specified thoroughly. >> >> Any ideas on how to move forward? >> >> My own inclination is to try to put *something* together, and revisit >> it later once we have some user feedback. >> >> -- >> Sebastian Benthall >> OpenGeo - http://opengeo.org >> > It seems there is a front-end and a back-end side to this. Things are a > little easier than with layer metadata because we don't have to worry > about contacts independent of users (you have to be a user to make a map > in the GeoNode). > > On the front-end: > Just drop the field from the form and don't explicitly send any user > info to the Django app when saving a map. Django can tell us who's > logged anyway. > > On the back-end: > Change the 'contact' field on Maps to a ForeignKey on users. This will > be a backwards-incompatible change, so we should work out a migration > process for the live site. Probably, we should just dump the data, > reset the database, and hand-modify the data dump to use a user id > instead of a string for the contact field. > > Does this answer the question, or am I missing something? > > -- > David Winslow > OpenGeo - http://opengeo.org/ > Oh yeah, I think the ownership stuff that Luke is working on may involve setting ownership in a different way than just as a field on the model. So you might want to talk to him before we go nuts with the migration. (Although, we already have an owner field on Maps as a result of some work Ariel has done.) -- David Winslow OpenGeo - http://opengeo.org/