Template talk:IPA

Slashes and brackets

Could the template detect if the text is surrounded by / / or [ ] to not add them itself? Currently we can have ugly things like /[rɪd]/ and French pronunciation: [/fʁɑ̃sɛ/]. -- Error (talk) 17:10, 22 May 2024 (UTC)

As long as it's in the guide-linking mode (i.e. the first parameter is a language code), it should use brackets except for English, per MOS:PRON. Otherwise it could link a phonemic transcription to a phonetic key. Nardog (talk) 07:40, 23 May 2024 (UTC)

Would a div tag option be useful?

Recently, an editor used this template to wrap a block of content that was already wrapped in <poem>...</poem> tags, which caused a Linter "span tag wrapping a div tag" error. I reverted because I didn't see a way for this template to emit div tags, and I didn't know if that would be useful.

I added a test case, since removed because I don't want to clutter up the Linter error reports, showing how the template was used in that case. If it would be useful and in conformity with the purpose of this template, a |div=yes option could presumably be added to the module code; that option would use div tags instead of span tags in this template's output. We inserted a similar option into {{Collapsed infobox section begin}} to fix Linter errors that cropped up in some situations. – Jonesey95 (talk) 22:09, 28 May 2024 (UTC)

You can just use {{IPA}} inside <poem>...</poem> instead of the other way around. Nardog (talk) 00:57, 29 May 2024 (UTC)

IPA for Brazilian Portuguese

inputting {{IPA|pt-BR|...}} generates a text in incompatible fonts (seemingly with HTML sans-serif when I inspected it), when the same template with {{IPA|pt|...}} uses default fonts. Juwan (talk) 14:08, 1 August 2024 (UTC)

bumping and updating that this happens with every language with a country or region tag added. Juwan (talk) 23:51, 11 September 2024 (UTC)

As far as I can tell from reading IETF language tag, this is a bug on the browser's part. Nardog (talk) 00:52, 12 September 2024 (UTC)

Apparently it needs to be pt-Latn-BR-fonipa instead of pt-BR-Latn-fonipa. I'll deploy the fix once other pending bugs are ironed out. Nardog (talk) 05:34, 26 December 2024 (UTC)

@Nardog oh that's annoying. just to note that this is not actually a bug for just Brazilian Portuguese after all but for all [language code] + [country code] codes. good luck fixing that! Juwan (talk) 14:00, 26 December 2024 (UTC)

I've already fixed it in the sandbox, I'm just holding off on deploying it to reduce the number of times the job queue is flooded. Nardog (talk) 17:34, 26 December 2024 (UTC)

Deployed. Nardog (talk) 10:24, 27 December 2024 (UTC)

Category:Pages with Mari IPA is suddenly empty and it displays an error message: "Error: language code for "Mari" not found. Please see Template talk:IPA for assistance.". Should its history be merged into Category:Pages with Mari (Russia) IPA or is this a separate issue? ✗plicit 11:12, 27 December 2024 (UTC)

It's one of the things fixed. Just deleting it is fine I think, as the whole page is generated by the template. I've tagged it with C4. Nardog (talk) 11:36, 27 December 2024 (UTC)

[und] should generate an error-tracking category

ISO = [und] should probably never be used on WP, as it's unlikely we're going to ever have IPA transcriptions where our sources are unable to identify the language. [mis] should be used if there is no ISO code for the language, but if we just haven't worked out what the ISO code is, [und] is wrong and should be corrected when we figure it out. For that there should be an automatic error-tracking category that's generated whenever [und] is used. [In the unlikely event that [und] is actually correct, there could be an override.] Or perhaps we could use a bot to change all instances of [und] to a non-ISO substitute code like [xxx], and have that generate the tracking category for future cleanup. — kwami (talk) 07:48, 28 January 2025 (UTC)

So what is the correct IETF tag for a phonetic sound regardless of the language in which it occurs? Nardog (talk) 09:56, 28 January 2025 (UTC)

No idea. I only know ISO 639-3, and there's nothing for it there. — kwami (talk) 09:59, 28 January 2025 (UTC)

AFAICT from our articles, IETF is based on ISO, and none of the other ISO sets have a code for 'other coded language that we don't bother to specify'. — kwami (talk) 10:11, 28 January 2025 (UTC)

That strikes me as odd. So there's no way to semantically notate e.g. "the letter ⟨A⟩, regardless of the language"? Nardog (talk) 10:15, 28 January 2025 (UTC)

You would identify that as 'Latn' for Latin script in ISO 15924. But AFAICT that's intended as a suffix to ISO 639, which is only intended to identify actual languages. Maybe you could use its Unicode number, under ISO/IEC 10646, as it appears that's also part of IETF. But it would be rather redundant to ID the letter 'A' with the Unicode point for the letter 'A'. — kwami (talk) 10:24, 28 January 2025 (UTC)

{{IPA}} has used und for two and a half years now, you're just waking up to it. You may regard Category:Pages with undetermined IPA as an error-tracking category if you say so.

What is your source for the claim that "undetermined" means "the language this is a representation of has yet to be determined" rather than "this representation has no determined language"? Nardog (talk) 10:07, 28 January 2025 (UTC)

I looked it up last time something like this came up, and verified that our description is correct, but we don't seem to have a ref in our article. I can ask the ISO Registrar. But ISO is intended to identify languages. It wouldn't make sense for it to have a code for 'I know what the language is, but I'm not going to tell you'.

We could choose one of the [Qxx] codes that are reserved for private use. — kwami (talk) 10:16, 28 January 2025 (UTC)

Then please find a published reliable source. Nardog (talk) 10:32, 28 January 2025 (UTC)

As I said, I'm asking ISO. But ISO codes are all identification codes. They all identify in some fasion. [mis] means 'no ISO code exists for this language.' [mul] means 'multiple languages are involved; it's not possible to identify with a single code.' [zzx] means 'this is not human language'. No ISO code means 'we can't be bothered to provide the ISO code'. That would just be silly. If you haven't looked up the ISO code for a language, then you don't identify it with an ISO code until you do. If we need something for that meaning on WP, then we need to use something other than ISO.

As far as just picking up on this, I remember saying something to this effect some years ago. This debate is oddly familiar. — kwami (talk) 10:58, 28 January 2025 (UTC)

That's not all of what are using und. A lot of them are "this sound/sequence of sounds in no particular language". Nardog (talk) 11:10, 28 January 2025 (UTC)

But that's not an undetermined language, but no language at all. If there's no language, we shouldn't have a tag that 'maybe someday we'll discover what the language is'. I'll ask about that too. — kwami (talk) 17:45, 28 January 2025 (UTC)

She just copied section 5.7.5 of the standard, where it describes 4 codes 'for special concepts' that 'do not designate identified individual languages or language groups' -

— “mis”: miscellaneous language (no language code element is assigned for this individual language or language group or included in the subset of the language code used by a given application);

— “mul”: multiple languages in the content;

— “und”: language undetermined;

— “zxx”: no linguistic content.

In most cases where we use 'und' the language is known, we just haven't looked up the ISO code.

I asked about using 'und' for recordings of pronunciations of IPA letters, and she just rolled her eyes. — kwami (talk) 01:36, 29 January 2025 (UTC)