Why term for lemma? - Wikidata-tech - lists.wikimedia.org

List overview All Threads
Download

Why term for lemma?

Re: [Wikidata-tech] Fwd: Two...

Wikidata deving and vagrant

Denny Vrandečić

2 Nov 2016 2 Nov '16

9:53 p.m.

Hi, I am not questioning or criticizing, just curious - why was it decided to implement lemmas as terms? I guess it is for code reuse purposes, but just wanted to ask. Cheers, Denny

Attachments:

attachment.htm (text/html — 278 bytes)

Reply

Show replies by date

Daniel Kinzler

2 Nov 2 Nov

10:42 p.m.

Am 02.11.2016 um 21:53 schrieb Denny Vrandečić:

Hi, I am not questioning or criticizing, just curious - why was it decided to implement lemmas as terms? I guess it is for code reuse purposes, but just wanted to ask.

Yes, ideed. We have code for rendering, serializing, indexing, and searching Terms. We do not have any infrastructure for plain strings. We could also handle it as a monolingual-text StringValue, but that offers less re-use, in particular no search, and no batch lookup for rendering. Also, conceptually, the lemma is rather similar to a label. And it's always *in* a language. The only question is whether we only have one, or multiple (for variants/scripts). But one will do for now. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V.

Reply

Denny Vrandečić

3 Nov 3 Nov

1:11 a.m.

Ah, great. Thanks for the answer! I'm glad to hear that it makes development faster :) Will the forms be terms too? That *might* even be a solution for the multiscript issue. On Wed, Nov 2, 2016, 14:42 Daniel Kinzler <daniel.kinzler(a)wikimedia.de> wrote:

Am 02.11.2016 um 21:53 schrieb Denny Vrandečić:

Hi, I am not questioning or criticizing, just curious - why was it decided to implement lemmas as terms? I guess it is for code reuse purposes, but

just

wanted to ask.

Yes, ideed. We have code for rendering, serializing, indexing, and searching Terms. We do not have any infrastructure for plain strings. We could also handle it as a monolingual-text StringValue, but that offers less re-use, in particular no search, and no batch lookup for rendering. Also, conceptually, the lemma is rather similar to a label. And it's always *in* a language. The only question is whether we only have one, or multiple (for variants/scripts). But one will do for now. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. _______________________________________________ Wikidata-tech mailing list Wikidata-tech(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech

Reply

Thomas PT

11:28 a.m.

An other genuine question: why having both the Term and the MonolingualText data structures? Is it just for historical reasons (labels have been introduced before statements and so before all the DataValue system) or is there an architectural reason behind? Cheers, Thomas

Le 3 nov. 2016 à 01:11, Denny Vrandečić <vrandecic(a)gmail.com> a écrit : Ah, great. Thanks for the answer! I'm glad to hear that it makes development faster :) Will the forms be terms too? That *might* even be a solution for the multiscript issue. On Wed, Nov 2, 2016, 14:42 Daniel Kinzler <daniel.kinzler(a)wikimedia.de> wrote: Am 02.11.2016 um 21:53 schrieb Denny Vrandečić:

Hi, I am not questioning or criticizing, just curious - why was it decided to implement lemmas as terms? I guess it is for code reuse purposes, but just wanted to ask.

Yes, ideed. We have code for rendering, serializing, indexing, and searching Terms. We do not have any infrastructure for plain strings. We could also handle it as a monolingual-text StringValue, but that offers less re-use, in particular no search, and no batch lookup for rendering. Also, conceptually, the lemma is rather similar to a label. And it's always *in* a language. The only question is whether we only have one, or multiple (for variants/scripts). But one will do for now. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. _______________________________________________ Wikidata-tech mailing list Wikidata-tech(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech _______________________________________________ Wikidata-tech mailing list Wikidata-tech(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech

Reply

Thiemo Mättig

11 Nov 11 Nov

2:38 p.m.

Tpt asked:

why having both the Term and the MonolingualText data structures? Is it just for historical reasons (labels have been introduced before statements and so before all the DataValue system) or is there an architectural reason behind?

That's not the only reason. First, all data values (including monolingual text) must implement the same DataValue interface. Term must not implement anything (it does implement Comparable for convenience). All DataValues share the same abstract DataValueObject base class. The only reason for this is code sharing. No code should type hint against DataValueObject (I just checked and hurray, we are clean). MonolingualTextValue could indeed share code with Term. But it's not possible to do "class MonolingualTextValue extends DataValueObject, Term" in PHP. We would need to drop the code sharing with DataValueObject and do "class MonolingualTextValue extends Term implements DataValue" instead, which means we would have to copy all the code from DataValueObject over to MonolingualTextValue. This is entirely possible, but what would be the actual advantage of such a change? Which code would benefit from being able to pass MonoLingualValue's to code that accepts Term's? Best Thiemo -- Thiemo Mättig Software-Entwickler Wikimedia Deutschland e.V. | Tempelhofer Ufer 23-24 | 10963 Berlin Tel. (030) 219 158 26-0 http://wikimedia.de Stellen Sie sich eine Welt vor, in der jeder Mensch an der Menge allen Wissens frei teilhaben kann. Helfen Sie uns dabei! http://spenden.wikimedia.de/ Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e.V. Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg unter der Nummer 23855 B. Als gemeinnützig anerkannt durch das Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.

Reply

Daniel Kinzler

5:27 p.m.

Am 11.11.2016 um 14:38 schrieb Thiemo Mättig:

Tpt asked:

why having both the Term and the MonolingualText data structures? Is it just for historical reasons (labels have been introduced before statements and so before all the DataValue system) or is there an architectural reason behind?

That's not the only reason.

Besides the code perspective that Thiemo just explained, there is also the conceptual perspective: Terms are editorial information attached to an entity for search and display. DataValues such as MonolingualText represent a value withing a Statement, citing an external authority. This leads to slight differences in behavior - for instance, the set of languages available for Terms is suptly different from the set of languages available for MonolongualText. Anyway, the fact that the two are totally separate has historical reasons. One viable approach for code sharing would be to have MonolingualText contain a Term object. But that would introduce more coupling between our components. I don't think the little bit of code that could be shared is worth the effort. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V.

Reply

Thomas PT

8:01 p.m.

Ok! Thank you very much for the answers! Cheers, Thomas

Le 11 nov. 2016 à 17:27, Daniel Kinzler <daniel.kinzler(a)wikimedia.de> a écrit : Am 11.11.2016 um 14:38 schrieb Thiemo Mättig:

Tpt asked:

why having both the Term and the MonolingualText data structures? Is it just for historical reasons (labels have been introduced before statements and so before all the DataValue system) or is there an architectural reason behind?

That's not the only reason.

Besides the code perspective that Thiemo just explained, there is also the conceptual perspective: Terms are editorial information attached to an entity for search and display. DataValues such as MonolingualText represent a value withing a Statement, citing an external authority. This leads to slight differences in behavior - for instance, the set of languages available for Terms is suptly different from the set of languages available for MonolongualText. Anyway, the fact that the two are totally separate has historical reasons. One viable approach for code sharing would be to have MonolingualText contain a Term object. But that would introduce more coupling between our components. I don't think the little bit of code that could be shared is worth the effort. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. _______________________________________________ Wikidata-tech mailing list Wikidata-tech(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech

Reply

2720

days inactive

2729

days old

wikidata-tech@lists.wikimedia.org

Manage subscription

6 comments

4 participants

tags (0)

participants (4)

Daniel Kinzler
Denny Vrandečić
Thiemo Mättig
Thomas PT