Wikidata talk:WikiProject Protected areas/Properties/Ontology

This is a discussion about the proposal for change ontology in order to homogenize the management of "protected area" of the item and its IDs.

To facilitate the discussion, make your comments in the topic section.

Properties of "official protected area" vs just an identifier

• .

• .

P1435 as a backbone of protections

Yes, I think it makes sense to use P1435 more widely for natural heritage. While for nature reserves and alike alone it's plausible to provide designation as P31 value, then it's less plausible for other types of protected objects (individual lakes, hills, trees, boulders etc.) that essentially aren't its designation (see #Scope below). So I'd see consistency between different kinds of natural heritage as the the main benefit of using P1435. 2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8 11:05, 17 February 2019 (UTC)

Changes proposed by each property

• .

• .

P31 considerations

The example of Montseny confuses me a little. Is the object that has the protection supposed to be Q937042, not Q1401508? I'd not add any designations (including biosphere reserve (Q158454) of Q1401508) under P31 if P1435 is used for designations. Then for protected areas P31 value could be a broad category like protected area (Q473972). This might be similar to classification for persons that are all instances of broad category "human" while its useful to provide some more specific classes as values of separate properties (e.g. occupation or position held) comparable to heritage designation. 2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8 11:05, 17 February 2019 (UTC)

@2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8: Yes, probably this is not the best example. But I presume you understood what I mean. In my responses to #Scope are some more comments. Amadalvarez (talk) 07:37, 21 February 2019 (UTC)

P814 considerations

I'm not sure I understand why it would a be a good idea to move value of this property under P1435. Designations as certain protection measures or certain recognition, and IUCN categories as borad protection levels can be considered distinct. Protected object is sorted into an IUCN category due to being an object of some designation in the first place. WDPA database also provides IUCN category apart from designation. 2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8 11:05, 17 February 2019 (UTC)

@2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8: Sorry for the delay in my answer. If the P814 would always exist with another more specific protection, we could leave it as it is (as you propose) and collect in the P1435 only the specific protection. However, there are 6,000 items where there is the P814 without any other protection ID property and about 20,000 cases with P814+P809 and no other protection ID property. These cases would be without any value at P1435, if we don't use P814 in P1435. Thanks,Amadalvarez (talk) 07:27, 21 February 2019 (UTC)

But all objects with a WDPA entry do have some designation in additon to IUCN category, don't they? Currenlty, for protected areas this designation seems to be often given as P31 value. Presenting IUCN category as a designation, instead of an actual (or original, primary) designation doesn't seem suitable to me. I assume items with P814, but no P809 nor P1435, also have some actual designation. If this designation already isn't given as P31 value, then it should be found out what it is. If the idea is to have P1435 value for all designated objects, then we could use P1435 consistently for all designations, instead of providing some designations as P31 value, as described above under P1435 and P31 considerations. PS Pinging an IP address doesn't work. (To keep track, I just bookmark those a couple of discussions that I've participated lately.) 2001:7D0:81F7:B580:64AE:30E7:C870:F961 09:16, 21 February 2019 (UTC)

P809 considerations

Maybe for some cases, like Natura 2000 areas, P809 as a separate statemetent could use statement is subject of (P805) as a separator in order to work around single value constraint if there are multiple values, instead of omitting P809 as a separate statement. However it's unlclear to me which designations are designations for the same entity and which are designations for different but coextensive entities. It might be reasonable to have one item for Natura 2000 area and its closely related multiple designations (though, there already are different items along with Wikipedia articles like Q25444739 and Q25444718 for a single Natura 2000 area). For other cases like the example of Yellowstone given here it's less clear if national designation and two international designations for a coextensive area should be given in the same item. 2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8 11:05, 17 February 2019 (UTC)

@2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8: It's a good idea use P805, but it is a "element type of data", and WDPA is a text (the ID) or even could be a link. Really, I don't mind if we keep P809 as a separate statement, while its role were just a ID, but no a "protection mark" for the item. The Yellowstone example is just to show that P809 is a container and not a protection authority, so we must use the green/yellow colored properties to protect and P809 to get information. Regarding the multiple and/or overlapping protections, I figure out that "cultural heritage" vs "natural heritage" have a key difference. Usually, cultural heritage protects specifics objects (buildings, old structures, etc.) that have been defined with on WD item. So is a "point in the space" with its own properties that, in addition, are granted a protections. In these cases, it's difficult to find a WD item for this singular protection (except in some UNESCO World Heritage). Otherwise, the natural protections are oriented to an area that may contains one or more "geographical spots traditionals" but the area itself didn't exist before the protection. So, we have (in WD) an item for the protection before to have an item for the protected phisical block. Then, If WD item is a traditional place with several protections over, all of them must be in P1435 (same as cultural heritage). But, if WD item is about the protected area "by definition" which covers spaces, the P1435 should contain its own protection(s); the other protected areas overlapping space should show with territory overlaps (P3179); and the "geographical spots traditionals" inside the protected area should by in has part(s) (P527). Excuse me for large response, but we are openning new (and interesting) topics. Thanks Amadalvarez (talk) 08:48, 18 February 2019 (UTC)

As for P805, it can be used as a qualifer under P809 statement. For instance, if Natura 2000 area has multiple P809 values then respective P805 qualifer values can be Q796174 and Q2463705 (similar to use of P805 in Q12364484 for another identifer). Again, I'm not saying that multiple designations like these have to be together in one item necessarily, just saying that it's possible to work with multiple values like this without dropping the single value constraint.

I don't know, P809 or any other identifier as a separate statement seems like merely an identifer to me. I'd rather think of a particular designation as something like a "protection mark".

Cultural heritage designations indeed tend to be for indiviual objects (there are also area designations like Q25504552, though) and natural heritage designations might be more often for areas. As several issues arise from this area orientation (e.g. also property constraints issues like this), then that's why I started a separate section (#Scope) on natural protections and other natural heritage designations not being oriented to areas only. 2001:7D0:81F7:B580:E4BF:736D:653F:804 11:20, 18 February 2019 (UTC)

P1848 considerations

For P1848, I feel that the problem is more a lack of protected area type. Like for Isthme De Miquelon-Langlade (Q1795991) or you exemple of Pointe au Sel (Q3393570) are in french Terrain acquis par le Conservatoire du Littoral ("Land acquired by the Conservatoire du Littoral") who don't seem to have item. Also these place have also IUCN protected areas category (P814). --Fralambert (talk) 16:12, 16 February 2019 (UTC)

@Fralambert: Yes. In some other protections the "list of types" are well-defined by the own protection, as Natura2000 which have 3 specific types (that can be simultaneously in the same item). In P1848, the list is not clear.

On the one hand we have the list that appears in the "Types d'espaces" menu in the search for "espaces protégés". This list shows some values that are not specific to the INPN (World Heritage of UNESCO, Biosphere Reserve, RAMSAR area) and we would already have them via their own protection. The rest of the values, could be valid as subclasses of protections of the P1848 and, if they do not exist, we could create them.

On the other hand, the list of current values of P31 can be assimilated by 80% with the mentioned list of the INPN website, but some others have to be verified or matched.

See the new table matching both, the list of INPN web and the present P31. I hope it may help us to review and decide which INPN subtypes should have. Thanks Amadalvarez (talk) 21:46, 16 February 2019 (UTC)

Scope

Perhaps it would be a good idea to define scope of this ontology or wikiproject more explicitly as "natural heritage" (similar to "WikiProject Cultural heritage"), rather than "protected areas". There are designations like Q25499506, Q21573182 or Q422211 for objects that essentially or by definition are something else than a protected area or its designation, i.e. lakes, hills or non-areas (individual trees, boulders) etc. Designations for that sort of natural heritage along with objects that essentially are protected areas (designations for nature reserves, nature parks etc) are at least in some countries granted under same acts and objects are listed in same registries. WDPA database also includes all sorts of natural heritage, not only protcted areas. So I think it'd make sense to have these both kinds natural heritage less apart and their ontology as consistent as possible. 2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8 11:05, 17 February 2019 (UTC)

@2001:7D0:81F7:B580:55AF:BB3D:3B1A:D1F8:. In my opinion, it is already within the scope. In fact, everything that can be included within the UICN categories, are protected. Of the 3 examples you mentioned, the 2nd and 3th are already a subclass of protection area. The Q25499506 is not protected, because the subclass assigned to protected natural object (Q12366109) doesn't contain a "protected". If something is protected, some of its upper classes must be "protected" as well. I mean, not all the "natural objects" are protected, but all the "protected natural object of Estonia" should be. So, instead of open the proposal to all "natural heritage", I think that we must created "protected natural object/element" at same level of protected areas and both depending from natural heritage. The changes I propose in subclasses, are:

• natural heritage (Q386426)
  • protected natural object
    • individual protected natural object
  • protected area (Q473972)
    • La Albufera Natural Park (Q2367959)
    • nature reserve (Q179049)
    • etc....

Subclasses of "natural heritage" are: A) some class of area B) some class of elements. All refered to areas have, in addition, the protected area (Q473972) as a subclasse. Otherwise, the "elements" doesn't have a "protected object" class to add. Concepts as natural monument (Q23790), remarkable tree (Q811534), not always are protected, but as they are not an "area",we need a complementary class to mark which are/not a protected element.

Suggestions will be wellcome. Thanks, Amadalvarez (talk) 21:13, 17 February 2019 (UTC)

Q21573182 set as subclass of Q473972 is probably a mistake. As said, this is an example of a designation that is applied to individual objects like trees or boulders, that are not areas really. As for Q422211 I'm not sure. If all objects with this designation encompass some area and this object (island, hill, meadow etc) is protected then possibly these objects are protected areas in some loose sense. But essentially geographical objects with this designation are something else than a protected area, they exist regardless of their protection status.

I don't quite follow what you mean by Q12366109 being not protected. Similarly, would you say that protected area (Q473972) is not protected because it isn't a subclass of something that is protected?

Surely not all natural heritage is protected, but I don't see a problem here really. There are particular designations to show that objects are protected. Neither are all objects with a designation protected (some designations merely recognize an object as heritage, without applying any protection measures).

(Protected) area can be considered a (natural) object, too. Similar to geographic region (Q82794) being set as subclass of (subclass of) geographical feature (Q618123). Regarding Estonian designations, both protected area and individual protected object are types of protected natural object.

The same way as cultural heritage designations are subclasses of cultural heritage (Q210272), we can use natural heritage (Q386426) as the root class for all sorts of natural heritage designations. (This is what the value type constraint for P1435 already suggests.)

If you say that ontology/wikiproject is already within the scope, then do you mean that this scope is explicitly protected areas, and it should not deal with (designated) natural heritage that are not (protected) areas? To me it still seems that it isn't practical to keep these apart. As said, same legislative acts deal with them, they are listed in same registries, both of them (areas, and non-areas like individual protected objects in Estonia) have IUCN categories. 2001:7D0:81F7:B580:E4BF:736D:653F:804 11:20, 18 February 2019 (UTC)

Excuse me. Probably I did not explain correctly. My english is not optimus to express complex concepts. Don't mind. I try again in another way. As natural heritage (Q386426) has not been considered a "protected class" by everybody, nowadays we have a mix of protected and not protected items under this class. Therefore, if we decide that Q386426 = protected and within the scope, we will have to handle a lot of exceptions. Thats why, in my last message, I talked about to CREATE a new class protected natural object, at the same level of protected area (Q473972), where we can link as subclasses those "elements" / "objects" that are protected, but are not an area". It is, classes as protected natural object (Q12366109) or natural monument in Germany (Q21573182). This will allow us to assign this new class (or even individual protected natural object) to those natural monument (Q23790) that are protected, without include all the natural heritage as protected. Doing this, it will be easy to increase the scope to this new class and its subclasses. Obviously, we will include this new class in the constraints where now Q386426 is alone. Thanks, Amadalvarez (talk) 07:31, 19 February 2019 (UTC)

My idea was that all natural heritage (Q386426) doesn't have to be protected or designated objects to be suitable scope here since for the most part objects considered as natural heritage are designated objects, and generally natural heritage is the subject that conservation designations deal with. Similarly, do you see scope of cultural heritage wikiproject problematic since not everything that might be considered as cultural heritage (Q210272) is necessarily designated?

What sort of exception handling do you have in mind? It's probably worthwhile to check that values for P1435 are actual designations and not loosely defined classes like nature reserve (Q179049) or natural monument (Q23790), but for this purpose it doesn't help that relevant classes are subordinated to other loosely defined classes like protected area (Q473972) or "individual protected natural object". As for cultural heritage, some "none of constraints" have been set for this purpose. (Another option would be to make all designations into instances of heritage designation (Q30634609) and then narrow down P1435 values by that if really necessary.) As for properties like IUCN protected areas category (P814), then "(individual) protected natural object" class doesn't seem to help in relation to current constraint false positives (see here) either. We probably don't want to make individual protected objects (e.g. trees, boulders) into instances of a non-defining characteristic ("protected object", or more specifically, its designation, that would be given as P1435 value instead).

Also, I don't know if Wikidata currently deals with any natural heritage designation that is not about protection, but probably such designations exists, similar to some cultural heritage designations that are merely about listing and recognition. Dealing with such designations, I assume, would be also in natural heritage scope. 2001:7D0:81F7:B580:4AF:6DEB:D81E:B89F 10:51, 19 February 2019 (UTC)

My concern is that not all items with natural heritage (Q386426) (or its subclasses) are really protected. Sometimes are just listed for their interest, but without no specific protection. Therefore, including all without having any type of filter on what should be included in P1435, there will be a devaluation of the P1435 utility as the "object protection list".

I still do not know what are the criteria to consider an object as protected / unprotected. This exercise is what I have done for the areas (the original proposal), but for the rest of objects, I have not thought about it. Therefore, in the face of your proposal to treat other objects (not areas), I only propose that we have the instrument to mark them. Note that in the case of protected areas the instrument consists in that the item has an ID property with subclass Wikidata property for authority control for protected areas (Q55978235) vs Wikidata property related to protected areas (Q27642681). But in the case of natural heritage, we can find protected items, items listed or in a dictionary (Q3042478, Q1444610, Q2859800), or false positives (the Q9259 building) are a subclass of "natural heritage", because Q9259 is a subclass of natural heritage (Q206194, Q774215), although in these cases it is not a problem because they are protected as cultural heritage.

In summary, the limitation of the scope of the proposal is not definitive, but because the items in protected areas are more refined and better aligned with the concept of protection. In the case of natural heritage, although the dictionary definition is interpreted as "protected", its application in the use has been broad. Therefore, to include them within the scope, we need a previous work to "clean" the missuses. My suggestion to have protected natural object class is one way (not the unique) to be able to incorporate the protected blocks of items. Thanks, Amadalvarez (talk) 07:26, 20 February 2019 (UTC)

As we have a corresponding discussion here that brings another problem to the clear split up of protected areas and natural heritage. In Germany we have natural monument in Germany (Q21573182) witch is a subclass of protected area (Q473972) and natural heritage (Q386426) because this is one law defined protection type for small areas and individual objects too. As example here is a natural monument Glühwürmchengrund und Immenweide (Q59786218) with a law defined area but a single tree like this Q60465490 has the same protection type. --GPSLeo (talk) 20:04, 24 February 2019 (UTC)

Mix up of physical area and protected areas = bad idea

As it is proposed here there seems to be an idea of having the physical area which is protected and every protection type in the same item. This is not a good idea, because it is much more difficult to show so many thing in one item it would need so many qualifiers that makes it very hard to edit and queries also will get really hard to work up correctly. One example: The Grunewald (Q11705528) is a forest the entire forest is in the landscape protection area Grunewald (Q59780177), but not the hole forest. Some areas of the forest are part of the landscape protection area because they are part of seven nature reserves.(And two former nature reserves too) So saying the forest has one protection category would be wrong, and the landscape protection area contains areas they are not part of the forest. Additionally there are two 95% overlapping NATURA2000 sites the SACFFH-Gebiet Grunewald (Q59780260) and the SPAVogelschutzgebiet Grunewald (Q59780203) in the forest containing some of the nature reserves. That makes three sites with different area, WDPA ID (P809) Common Database on Designated Areas ID (P4762) Natura 2000 site ID (P3425) and IUCN protected areas category (P814). We could be lucky that all these areas are in the same IUCN protected areas category (P814). No one wants to describe this with qualifiers. --GPSLeo (talk) 20:43, 24 February 2019 (UTC)

@GPSLeo: If I understand correctly, do you propose that the physical places (like Q11705528) do not have any identification nor relationship with the protection they (maybe partially) grant ?.Amadalvarez (talk) 08:02, 25 February 2019 (UTC)

Of course there should be statements like located in protected area (P3018) and may some more like "partially protected by" but the protected areas them self need separate items. --GPSLeo (talk)