[01:47:29] @vrandecic there should also be a point on celestial body type [01:48:09] usually for Earth, but I think we should keep it broad to support, say, Mars, or such. [03:32:31] [[WF:Type proposals/Point on earth]] (re @Feeglgeef: @vrandecic there should also be a point on celestial body type) [14:05:50] Can Wikifunctions search lexemes? Like, if I give it a string and a language, can it find whether there is a lemma or a form that is the same as this string? [14:07:15] we can't even manually search by language https://phabricator.wikimedia.org/T271776 [14:26:37] Not yet, no. There will be a UI selector component for Wikidata lexeme forms but that won’t be usable within a function. [14:26:37] https://www.wikifunctions.org/wiki/Wikifunctions:Support_for_Wikidata_content#Selector_components (draft) (re @amire80: Can Wikifunctions search lexemes? Like, if I give it a string and a language, can it find whether there is a lemma or a form tha...) [16:28:30] …in any event, we should specify the oblate ellipsoid of revolution, as in Q11902211 (re @Feeglgeef: usually for Earth, but I think we should keep it broad to support, say, Mars, or such.) [16:56:55] What is that? (re @Al: …in any event, we should specify the oblate ellipsoid of revolution, as in Q11902211) [16:58:03] The current proposal is just for earth, as I think that would keep it simpler [17:01:39] As I said… “in any event”. If the type is implicitly WGS84, we need to know that. (re @Feeglgeef: The current proposal is just for earth, as I think that would keep it simpler) [17:02:33] This needs a height, or is it not really a point on earth, but a line (covering all heights). (re @Feeglgeef: [[WF:Type proposals/Point on earth]]) [17:03:03] I still don't know what you are talking about (re @Al: As I said… “in any event”. If the type is implicitly WGS84, we need to know that.) [17:03:06] please ELI5 [17:03:16] It's on earth (re @Jan_ainali: This needs a height, or is it not really a point on earth, but a line (covering all heights).) [17:03:19] that's a height [17:11:13] I’m talking about the need to specify the https://en.wikipedia.org/wiki/Spatial_reference_system. (re @Feeglgeef: I still don't know what you are talking about) [17:12:09] Is ISO 6709 a valid answer? [17:15:30] No. ISO 6809 requires that the CRS be specified (item 4). (re @Feeglgeef: Is ISO 6709 a valid answer?) [17:18:54] Wait huh? [17:19:05] When did we start talking about rubber compounding? [17:19:10] I'm so confused [17:25:30] “A geographical point is specified by the following four items: [17:25:31] • First horizontal positioncoordinate (https://en.m.wikipedia.org/wiki/Horizontal_position) (_ϕ_ or _y_), such as latitude (negative number south of equator and positive north of equator) [17:25:33] • Second horizontal coordinate (_λ_or _x_), such as longitude (negative values west of Prime Meridian and positive values east of Prime Meridian) [17:25:34] • Vertical coordinate, (https://en.m.wikipedia.org/wiki/Vertical_coordinate) i.e. height or depth (optional) [17:25:36] • Identification of coordinate reference system (https://en.m.wikipedia.org/wiki/Coordinate_reference_system) (CRS) (optional) [17:25:37] The first three items are numerical values called coordinates. The CRS gives the relationship between the coordinates and a point on the earth. The identification of CRS could be a full description of properties defined in ISO 19111; (https://en.m.wikipedia.org/wiki/ISO_19111) only an identifier given by some registry (such as EPSG) (https://en.m.wikipedia.org/wiki/EPSG) [17:25:37] is used [17:25:39] in most cases, since only such identification is enough for most information exchange purposes.” [17:25:40] https://en.wikipedia.org/w/index.php?title=ISO_6709&oldid=1242977174 [17:26:55] Are we talking about ISO 6709 or 6809 [17:26:57] ? [17:27:06] oh (re @Al: “A geographical point is specified by the following four items: [17:27:07] • First horizontal positioncoordinate (ϕ or y), such as latitud...) [17:27:11] it's optional [17:27:14] we don [17:27:17] 't need it [17:27:19] :) [17:28:39] Yeah, I said we should, not we must. (But we must.) [17:28:55] Why should we [17:29:03] What even are the options [17:29:09] what's the difference [17:31:38] We don't need to store height or system, we can have them or not [17:31:39] but we absolutely need to say explicitly what we choose (and ideally why) [17:32:07] height is the height of the point that (re @Nicolas: We don't need to store height or system, we can have them or not [17:32:07] but we absolutely need to say explicitly what we choose (and id...) [17:32:10] 's on earth [17:32:54] For a same longitude/latitude there can be multiple point on Earth [17:33:09] In buildings for instance [17:33:14] ex. (27.986065, 86.922623) is 8,848.86m above sea level [17:33:25] its the highest point that's on earth [17:33:54] This is so oversimplified on so many levels... (re @Feeglgeef: its the highest point that's on earth) [17:34:07] how so [17:34:16] look the height is just ignored [17:34:22] pretend it doesn't exist [17:34:28] most functions don't require it [17:34:38] For starter, saying that the Everest is the highest point is subjective and relative to the system you chose [17:35:00] no like (re @Nicolas: For starter, saying that the Everest is the highest point is subjective and relative to the system you chose) [17:35:31] The peak of Everest is the heightest point at the line that makes up the coords of Everest [17:35:44] -something is the height of somewhere in the Netherlands [17:39:55] Only in areas untouched by humans. In a city where you may have multiple levels of "ground", which height is "on" the earth? (re @Feeglgeef: that's a height) [17:40:18] Not relevant (re @Jan_ainali: Only in areas untouched by humans. In a city where you may have multiple levels of "ground", which height is "on" the earth?) [17:40:24] It represents a line [17:40:28] And that's OK [17:40:43] Because that's what most function would use [17:42:06] So we should rename the type to "line on Earth" to be clear (and have the name available for an actual point in the future). (re @Feeglgeef: It represents a line) [17:42:22] No [17:42:48] Ok, I'll write my formal oppose then. [17:42:49] We have to pretend the earth is a 2D non-euclidian geometry [17:43:00] Just pretend that [17:43:15] Kk (re @Jan_ainali: Ok, I'll write my formal oppose then.) [17:43:27] On which ellipsoid? (re @Feeglgeef: We have to pretend the earth is a 2D non-euclidian geometry) [17:43:45] The Earth (re @Al: On which ellipsoid?) [17:44:16] Which abstract definition? (re @Feeglgeef: The Earth) [17:44:22] Why are you guys overcomplicated this [17:44:30] It's normal coordinates [17:44:53] If I tell you (27.986065, 86.922623) [17:44:58] You can tell where it is [17:45:00] Google it [17:45:05] It even shows you a map [17:46:51] I would like to remind everyone to remain polite, refrain from name calling, assume good faith, minimize the use of coarse language, and to work towards consensus. Thanks! [17:47:09] Soz (re @vrandecic: I would like to remind everyone to remain polite, refrain from name calling, assume good faith, minimize the use of coarse langu...) [17:47:40] "normal" coordinates in Sweden look like this: `6166984 1322751` (re @Feeglgeef: It's normal coordinates) [17:47:55] Well [17:48:01] Nothing comes up when I Google that [17:48:27] We can make the renderer language specific? [17:49:23] If I search with Google dot se the results seem relevant [17:49:24] (because we are not using Q11902211) (re @Jan_ainali: "normal" coordinates in Sweden look like this: 6166984 1322751) [17:50:28] @Jan_ainali That's a nice system, by the way -- I like the implicit precision [17:50:31] I provided a link earlier… [17:50:31] “A *spatial reference system* (*SRS*) or *coordinate reference system* (*CRS*) is a framework used to precisely measure locations on the surface of Earth as coordinates. It is thus the application of the abstract mathematics of coordinate systems (https://en.wikipedia.org/wiki/Coordinate_system) and analytic geometry (https://en.wikipedia.org/wiki/Analytic_geometry) to [17:50:31] geograph [17:50:33] ic space. A particular SRS specification (for example, "Universal Transverse Mercator (https://en.wikipedia.org/wiki/Universal_Transverse_Mercator) WGS 84 (https://en.wikipedia.org/wiki/WGS_84) Zone 16N") comprises a choice of Earth ellipsoid, (https://en.wikipedia.org/wiki/Earth_ellipsoid) horizontal datum, (https://en.wikipedia.org/wiki/Geodetic_datum) map projection(htt [17:50:33] ps://e [17:50:34] n.wikipedia.org/wiki/Map_projection) (except in the geographic coordinate system), (https://en.wikipedia.org/wiki/Geographic_coordinate_system) origin point, and unit of measure. Thousands of coordinate systems have been specified for use around the world or in specific regions and for various purposes, necessitating transformations (https [17:50:34] //en.wikipedia.org/wiki/Geographic_coord [17:50:36] inate_conversion) between different SRS.” [17:51:24] I would recommend that for coordinates we also try to stay compatible with Wikidata, as we would like use that data frequently [17:51:59] Why don't we make a type for all coordinate systems [17:52:11] It has 4 keys [17:52:18] X, Y, and Z [17:52:21] And system [17:58:32] My understanding is that different coordinate systems have a different understanding of their values. Some values go from -180 to +180, some go from 0 to 360, some go from -90 to +90, and some are not degrees at all, such as the one Jan just mentioned, or the what3words system. [18:00:07] I'll pretend you didn't say anything then 😉 (re @Feeglgeef: Just pretend that) [18:01:51] Exactly, I would miss the height but I could accept it if the reason is "be consistent with Wikidata" ( and we could also mention that Wikidata had height but never used it) (re @vrandecic: I would recommend that for coordinates we also try to stay compatible with Wikidata, as we would like use that data frequently) [18:02:44] Wikidata’s P625 supports only WGS84. Conversion between co-ordinate systems is a useful role for Wikifunctions. [18:02:55] France also has a similar system (but almost forgotten now) (re @Jan_ainali: "normal" coordinates in Sweden look like this: 6166984 1322751) [18:03:20] I'm pretty sure France doesn't even exist (re @Nicolas: France also has a similar system (but almost forgotten now)) [18:03:25] I've seen similar things for the uk and switzerland [18:04:42] Yes, we have P613 (re @Nikki: I've seen similar things for the uk and switzerland) [18:04:59] My understanding is that in California there are several systems in use. A "global" one, I think WGS84, and more local ones, which are tied to a certain continental plate, so that in case an earthquakte moves the plate by a meter or two, the reference points are still valid, since they are tied to the local plate, even though the WGS84 coordinate would have [18:04:59] changed. [18:05:19] But it's a while since I looked into it. [18:06:03] Sounds like fun! (re @vrandecic: My understanding is that in California there are several systems in use. A "global" one, I think WGS84, and more local ones, whi...) [18:07:17] yes, WGS 84 is the most common CRS for global coordinates. the US government uses NAD 83, which uses an older spheroid and is constant relative to the the North American Plate, which is fine unless you're on the wrong side of the San Andreas Fault [18:07:47] there's a new US datum coming soon, which will be even more fun [18:09:33] thanks, I couldn't remember the name 😅 I got as far as finding https://www.gov.uk/guidance/uk-geospatial-data-standards-coordinate-reference-systems and then was still thinking about whether I know the name "british national grid" (coordinates are definitely not the first thing that come to mind, at least) when you sent that (re @Al: Yes, we have P613) [18:10:28] it's not very unusual for a coordinate implementation to not store height, since you need a vertical datum for that instead of the horizontal datums used for the N/S E/W coordinates [18:11:10] Yes, the “national grid” is our electricity network. (re @Nikki: thanks, I couldn't remember the name 😅 I got as far as finding https://www.gov.uk/guidance/uk-geospatial-data-standards-coordina...) [18:12:00] Wikidata's coordinate data model is a bit weird in that it supports multiple globes but only one (usually unspecified) CRS per globe [18:26:10] Exactly, I want to avoid this messy situation (re @wmtelegram_bot: Wikidata's coordinate data model is a bit weird in that it supports multiple globes but only one (usually unspec...) [18:26:59] Everyone was "it's obvious what a coordinate is" and then the problem came as people had slightly different understanding [18:31:16] 🤔 the “globe” (reference ellipsoid) is usually part of the CRS specification. I think WGS84 is the reference ellipsoid in many CRS specifications. (re @wmtelegram_bot: Wikidata's coordinate data model is a bit weird in that it supports multiple globes but only one (usually unspec...) [18:32:56] that is incorrect actually, the value for globe for most coordinates on Wikidata is Q2, which is a planet and not an ellipsoid [18:35:29] Oh, that is truly weird! (re @wmtelegram_bot: that is incorrect actually, the value for globe for most coordinates on Wikidata is Q2, which is a planet and no...) [18:36:38] Plus the reference is something subjective, not everyone shares what others have in mind [18:36:42] the documentation defines WGS 84 as the coordinate reference system used for Earth coordinates, but only in documentation [18:37:10] other planets are completely undefined, though most only have one CRS in common use [18:40:48] I think for other globes it's also unclear whether the direction should be the value you enter or the value it displays (unless I'm misremembering how it works) [18:43:07] since someone on wikifunctions will probably want to write some coordiante transformation functions, it would probably make most sense to store (X,Y) and a CRS, either as a Wikidata ID or an EPSG ID (the standard database for coordinate reference systems) [19:29:18] yes (re @wmtelegram_bot: since someone on wikifunctions will probably want to write some coordiante transformation functions, it would pr...) [19:33:27] for other planets and celestial bodies, there is two competing system : planetographic and planetocentric [19:33:28] even for an amateur like me who knows them quite a bit, it's not always easy to know what system is used (re @wmtelegram_bot: other planets are completely undefined, though most only have one CRS in common use) [19:49:31] I've created Z19641 as a function that could benefit from a point type [19:50:19] Also why are y'all still yapping about the coordinate system? It's going to be 3 keys of integers and function authors can deal with other stuff [19:50:35] sorry, rational numbers/floats, not integers [19:52:30] Aniother warning. Please stay friendly. If you cannot, feel free to remain more quiet. We don't call others volunteer contributions "useless", we don't call discussing "yapping", we don't use four letter words for complex systems. [19:53:42] When did I call something useless? (re @vrandecic: Aniother warning. Please stay friendly. If you cannot, feel free to remain more quiet. We don't call others volunteer contributi...) [19:55:54] Also what 4 letter words for complex systems are you talking about [19:56:02] please elaborate [20:00:33] e.g. (re @Feeglgeef: This is not complex shit) [20:00:54] e.g. (re @Feeglgeef: the noun has 65 forms lmfao) [20:01:32] e.g. (re @Feeglgeef: lmfao basque has all top 8 spots) [20:02:25] We will treat each other and each other's languages and other cultural systems (e.g. calendars) with respect. [20:05:31] that's a good example (since it's independant of the system as long as both the point and the polygon are in the same system), thanks for creating it (but it would requires some more documentation) (re @Feeglgeef: I've created Z19641 as a function that could benefit from a point type) [20:06:09] Yeah (re @Nicolas: that's a good example (since it's independant of the system as long as both the point and the polygon are in the same system), t...) [20:08:22] how would you make a function that convert from one system to an other? for example from Lambert to WGS (Q541151 to Q11902211) [20:08:24] more specifically, how would you check that the input is indeed in the intended system and deal with errors? [20:10:21] or an other example from planetographic to planetographic? (for example for Mars, Q106948929 to Q106948918) [20:10:51] Find how to convert (re @Nicolas: how would you make a function that convert from one system to an other? for example from Lambert to WGS (Q541151 to Q11902211) [20:10:51] m...) [20:10:58] Implement it in python [20:11:06] Google how to validate them [20:11:12] Implement it in python [20:11:41] first, I don't do Python ;) [20:12:08] You asked me how I would (re @Nicolas: first, I don't do Python ;)) [20:12:26] then, coordinates validation is not always easy (planetographic and planetocentric "looks" the same, a regex can't determine what you meant) [20:15:28] also, does your fonction for enclave? a city in Lesotho is technically inside South Africa but it's not the answer most people would commonly expected [20:16:20] South Africa is, by definition, not a polygon (re @Nicolas: also, does your fonction for enclave? a city in Lesotho is technically inside South Africa but it's not the answer most people w...) [20:16:28] And is thus excluded [20:16:50] If you do want to include South Africa you can make a simple composition [20:17:01] in the end, it reminds my of this citation: "Everything simple is false. Everything which is complex is unusable." by Paul Valéry [20:17:01] we need to find the less worse trade-off [20:17:24] no, *I* can't, it's too complex for me (re @Feeglgeef: If you do want to include South Africa you can make a simple composition) [20:18:14] It's literally isInPolygon(South Africa) and not isInPolygon(Lesotho) [20:20:53] and not in eSwatini ;) (re @Feeglgeef: It's literally isInPolygon(South Africa) and not isInPolygon(Lesotho)) [20:21:15] SA has two enclaves (that I know of) [20:21:55] Eswatini borders Mozambique [20:22:00] IIRC [20:24:24] Also if you count it Kwandebele [23:46:23] 2222 [23:46:53] Hello